Contents

Introduction

Getting started with ALGLIB

FAQ

AP library description

AMP library description

ALGLIB reference manual

Introduction

Sections

ALGLIB license

ALGLIB is a free software which is distributed under a GPL license - version 2 or (at your option) any later version. A copy of the GNU General Public License is available at http://www.fsf.org/licensing/licenses

Documentation license

This reference manual is licensed under BSD-like documentation license:

Copyright 1994-2009 Sergey Bochkanov, ALGLIB Project. All rights reserved.

Redistribution and use of this document (ALGLIB Reference Manual) with or without modification, are permitted provided that such redistributions will retain the above copyright notice, this condition and the following disclaimer as the first (or last) lines of this file.

THIS DOCUMENTATION IS PROVIDED BY THE ALGLIB PROJECT "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE ALGLIB PROJECT BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS DOCUMENTATION, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

Reference Manual and User Guide

ALGLIB Project provides two sources of information: ALGLIB Reference Manual (this document) and ALGLIB User Guide.

ALGLIB Reference Manual contains full description of all publicly accessible ALGLIB units accompanied with examples. Reference Manual is focused on the source code: it documents units, functions, structures and so on. If you want to know what unit YYY can do or what subroutines unit ZZZ contains Reference Manual is a place to go. Free software needs free documentation - that's why ALGLIB Reference Manual is licensed under BSD-like documentation license.

Additionally to the Reference Manual we provide you User Guide. User Guide is focused on more general questions: how fast ALGLIB is? how reliable it is? what are the strong and weak sides of the algorithms used? We aim to make ALGLIB User Guide an important source of information both about ALGLIB and numerical analysis algorithms in general. We want it to be a book about algorithms, not just software documentation. And we want it to be unique - that's why ALGLIB User Guide is distributed under less-permissive personal-use-only license.

Acknowledgements

ALGLIB was not possible without the contribution of next open source projects:

Getting started with ALGLIB

Sections

Compiling ALGLIB

ALGLIB has a script-based compilation system. 'Script-based' means that you don't have to use make to compile ALGLIB. ALGLIB distribution contains Bash scripts and BAT-files (for Windows users) which will compile and test ALGLIB for you. Each script (build/check/...) is provided in two identical versions: as a Bash script and as a Windows batch file. The only difference is a set of compilers script supports.

Selecting/configuring your compiler

If you are *nix user, your environment is ready for the ALGLIB compilation. If you are Windows user, make sure that your compiler is in your PATH. MSVC users have to execute vcvars32.bat, vcvarsx86_amd64.bat or vcvarsx86_ia64.bat (depending on the hardware they use) in a shell window. These scripts are located in the MS SDK or MS Visual Studio directories.

Installing precompiled GMP/MPFR

Multiple precision version of ALGLIB rely on GMP/MPFR for numerical calculations. It needs following static libraries copied in the ./mpfr subdirectory of the ALGLIB root: mpn, gmp, mpfr. If you are *nix user, you must compile/copy GMP/MPFR static libraries yourself. If you are Windows user, you can use precompiled static libraries (they are installed by executing precompiled_mpfr.bat

Compiling ALGLIB

To compile ALGLIB you just need to cd into ALGLIB directory and to execute compilation script ./build (or build.bat for Windows users). You must specify compiler name and you can specify additional compiler parameters. Multiple parameters must be enclosed in double quotes.

./buildreturns full list of compilers supported
./build gcccompilation using GCC
build.bat msvccompilation using MSVC
./build gcc -m32one additonal parameter, no quotes
./build gcc "-m32 -O3"    multiple parameters in double quotes

Successful compilation will be completely silent. Compiler messages will be redirected to the log.txt in the ALGLIB root directory. In the case of the error you will get short message, but the most detailed info will be in a log file. After compilation is done, header files, libmpalglib static library and GMP/MPFR static libraries will be copied to the out folder of the ALGLIB root directory. ALGLIB is ready to use!

Testing ALGLIB

After ALGLIB compilation it can be tested:

check.batshort help
check.bat msvc allall the library is tested
check.bat msvc all_silentsilent mode, only errors are echo'ed
./check gcc fftfft.cpp is tested
./check gcc fft "-m32 -O3"    custom parameters are passed

Executing examples

Some units are accompanied by examples which can be executed by you. Example sources are stored in ./examples directory. They can be modified - feel free to experiment with them.

example.bat list autogklist examples for autogk.cpp unit
example.bat view autogk_smoothview example source
./example gcc autogk_smoothexecute example
./example gcc autogk_smooth "-m32 -O3"    custom parameters are passed

FAQ

Sections

What is the blank file stdafx.h here for?

MSVC and some other compilers require the #include <stdafx.h> directive in the program code to manage precompiled headers, and create the stfafx.h file when generating a new project. However, some compilers (e.g. BCB) use other tools to manage precompiled headers. In this case the #include <stdafx.h> directive as such doesn't hinder their operation, however if the file with this name is absent, a compilation error occurs. The blank file called stfafx.h is created to avoid this error. If your development environment already created the file, leave it unchanged.

Why are round brackets used when addressing arrays, not square brackets?

Imagine you are addressing the a matrix element in the common notation: a[x][y] instead of a(x,y). Actually there are two index operators called here instead of one. The first is indexing of a matrix by x, that returns a reference to a temporary structure that describes the matrix row. The second is indexing that temporary structure by y, which returns the reference to the needed element. Addressing through overloading the round brackets is much more effective, as no temporary structures are required.

What is the AMP library?

The AMP library (ALGLIB Multiple Precision) is an object oriented interface for the MPFR library, which is used by the ALGLIB programs.

What is the AP library?

AP library is a generic name for a set of libraries in several programming languages performing low-level tasks depending on specific programming languages. The AP library carries out tasks such as working with dynamic one- and multidimensional arrays in languages which do not support this data type, contains implementation of basic linear algebra algorithms, etc. The library is distributed as source codes under GPL 2+ license (GPL 2 or later). The library is attached to the ALGLIB package.

Why do some algorithms (for instance, optimization methods) use reverse communication instead of function pointers, delegates and other means of my programming language?

Optimization, integration and other similar methods are united by one common trait. They need to have a way of calculating the meaning of a function defined by the user at a point defined by the method.

The most convenient way of solving this problem is transferring a function pointer into the module. However bear in mind that ALGLIB package is written using pseudocode that is automatically translated into different programming languages. While each language has its own function pointer analog that is often different from other languages. When the ALGLIB pseudocode was developed, at some point is became clear that adding function pointers in it will be very complex as this feature is implemented differently in every language. This is why reverse communication was chosen as a different kind of solution.

What is ALGLIB aimed at?

It is aimed at creating a convenient and efficient multilingual scientific software library.

What is the difference between ALGLIB and other similar projects?

The ALGLIB package:

What is AlgoPascal?

AlgoPascal is a programming language, designed particularly for this project. The programs, written in this language, are processed by an automatic translator and translated into other programming languages. Almost all ALGLIB source is produced by the AlgoPascal translator.

AP library description

Sections

Introduction

The document describes a C++ version of the AP library. The AP library for C++ contains a basic set of mathematical functions and classes needed to compile ALGLIB package. The library includes the only module ap.cpp.

Conditionals

AP_ASSERT
This symbol enables checking of the array boundaries. If it is set by the "define" directive, then at each addressing to the dynamic array elements, the transferred index is verified for its correctness. In case of error the ap::ap_error exception is thrown. Checking the array boundaries makes the program more reliable, but slows down the program operation.

NO_AP_ASSERT
This symbol disables checking of the array boundaries. If it is set by the "define" directive, then the index being outside the array boundaries is not checked when the dynamic array elements are addressed.

Constants

ap::machineepsilon
The constant represents the accuracy of machine operations times some small number r>1.

ap::maxrealnumber
The constant represents the highest value of the positive real number, which could be represented on this machine. The constant may be taken "oversized", that is real boundary can be even higher.

ap::minrealnumber
The constant represents the lowest value of positive real number, which could be represented on this machine. The constant may be taken "oversized", that is real boundary can be even lower.

Functions

int ap::sign(double x)
Returns:
+1, if X>0
-1, if X<0
0, if X=0

double ap::randomreal()
Returns a random real number from half-interval [0,1).

int ap::randominteger(int maxv)
Returns a random integer between 0 and maxv-1.

double ap::round(double x)
Returns the nearest integer to x. If x is right in the middle between two integers, then the function result depends on the implementation.

double ap::trunc(double x)
Truncates the fractional part of x.
trunc(1.3) = 1
trunc(-1.3)= -1

double ap::pi()
Returns the constant π

double ap::sqr(double x)
Returns x2.

double ap::maxreal(double m1, double m2)
Returns the maximum of two real numbers.

double ap::minreal(double m1, double m2)
Returns the minimum of two real numbers.

int ap::maxint(int m1, int m2)
Returns the maximum of two integers.

int ap::minint(int m1, int m2)
Returns the minimum of two integers.

ap::ap_error class

This is a class of exception which is thrown when different errors occur in the AP library, for example - if the array index is found incorrect when the array boundaries check is enabled. Class contains one member - msg field which may contain additional information in textual form.

Array classes: basic concepts

First we will discuss general principles of working with array classes, then describe the classes and their methods.

Classes of the standard library allow operations with matrixes and vectors (one-dimensional and two-dimensional arrays) of variable size and with variable numeration of elements, that is, the array numeration can start at any number, end at any number and change dynamically. Because the array classes are templates, the arrays of the same dimension have the same set of member functions. And as the member functions of arrays with different dimensions differ only by the number of arguments, there is little difference between two-dimensional and one-dimenstional arrays.

Working with an array starts with the array creation. You should distinguish the creation of array class instance and the memory allocation for the array. When creating the class instance, you can use constructor without any parameters, that creates an empty array without any elements, or you can use copy and assignment constructors that copy one array into another. In case the array is created by the default constructor, it contains no elements and an attempt to address them may cause the program failure. If, during the copy operation, the source array has no memory allocated for the array elements, destination array will contain no elements either. If the source array has memory allocated for its elements, destination array will allocate the same amount of memory and copy the elements there. That is, the copy operation yields into two independent arrays with indentical contents.

After an empty array has been created, you should allocate the memory for its elements, using the setlength method. The content of the created array elements is not defined. If the setlength method is called for the array with already allocated memory, then, after changing its parameters, the newly allocated elements also become undefined and the old content is destroyed.

To address the array elements, an overloaded operator() is used. That is, the code addressing the element of array a with indexes a(i,j,k) will look like a(i,j,k). Below is given an example of factorial array calculation, illustrating the work with arrays.

integer_1d_array factarr(int n)
{
    integer_1d_array result;
    result.setbounds(1,n);
    result(1) = 1;
    for(int i=2; i<=n; i++)
        result(i) = result(i-1)*i;
    return result;
}

ap::template_1d_array class

This class is a template of dynamical one-dimensional array with variable upper and lower boundaries. Based on this class, the following classes are constructed:

typedef template_1d_array<int>     integer_1d_array;
typedef template_1d_array<double>  real_1d_array;
typedef template_1d_array<bool>    boolean_1d_array;
typedef template_1d_array<complex> complex_1d_array;

member functions

template_1d_array()
Constructor. Creates an empty array.

~template_1d_array()
Destructor. Frees memory, which had been allocated for the array.

template_1d_array(const template_1d_array &rhs)
Copy constructor. Allocates the separate storage and copies source array content there.

const template_1d_array& operator=(const template_1d_array &rhs)
Assignment constructor. Deletes destination array content, frees allocated memory, then allocates a separate storage and copies source array content there.

T& operator()(int i)
Addressing the i-th array element.

void setbounds(int iLow, int iHigh)
Memory allocation for the array. Deletes the array content, frees allocated memory, then allocates a separate storage for iHigh-iLow+1 elements.
The elements numeration in the new array starts from iLow and ends at iHigh. The content of the new array is not defined.

void setlength(int iLen)
Memory allocation for the array. Deletes the array content, frees allocated memory, then allocates a separate storage for iLen elements.
The elements numeration in the new array starts from zero. The content of the new array is not defined.

void setcontent(int iLow, int iHigh, const T *pContent)
The method is similar to the setbounds() method, but after allocating a memory for a destination array it copies the content of pContent[] there.

T* getcontent()
const T* getcontent() const
Returns pointer to the array. The data pointed by the returned pointer can be changed, and the array content will be changed as well.

int getlowbound()
int gethighbound()

Get lower and upper boundaries.

raw_vector<T> getvector(int iStart, int iEnd)
The method is used by the basic subroutines of linear algebra to get access to the internal memory of the array. The method returns an object, holding the pointer to a vector part (starting from the element with iStart index value and finishing with iEnd index value). If iEnd<iStart, then an empty vector is considered to be set.

const_raw_vector<T> getvector(int iStart, int iEnd) const
The method is used by the basic subroutines of linear algebra to get access to the internal memory of the array in the read only mode. The method returns an object, holding the pointer to a vector part (starting from the element with iStart index value and finishing with iEnd index value). If iEnd<iStart, then an empty vector is considered to be set.

ap::template_2d_array class

This class is a template of dynamical two-dimensional array with variable upper and lower boundaries. Based on this class, the following classes are constructed:

typedef template_2d_array<int>     integer_2d_array;
typedef template_2d_array<double>  real_2d_array;
typedef template_2d_array<bool>    boolean_2d_array;
typedef template_2d_array<complex> complex_2d_array;

Class members

template_2d_array()
Constructor. Creates an empty array.

~template_2d_array()
Destructor. Frees memory, which had been allocated for the array.

template_2d_array(const template_2d_array &rhs)
Copy constructor. Allocates the separate storage and copies source array content there.

const template_2d_array& operator=(const template_2d_array &rhs)
Assignment constructor. Deletes destination array content, frees allocated memory, then allocates a separate storage and copies source array content there.

T& operator()(int i1, int i2)
const T& operator()(int i1, int i2) const
Array element access.

void setbounds(int iLow1, int iHigh1, int iLow2, int iHigh2)
Memory allocation for the array . Deletes the array content, frees allocated memory, then allocates a separate storage for (iHigh1-iLow1+1)*(iHigh2-iLow2+1) elements.
The elements numeration in the new array starts from iLow1 and finishes at iHigh1 for the first dimension, and similarly for the second dimension.
The content of the new array is not defined.

void setlength(int iLen1, int iLen2)
Same as setbounds. but make zero-based array allocation.

void setcontent(int iLow1, int iHigh1, int iLow2, int iHigh2, const T *pContent)
The method is similar to the setbounds() method, but after allocating a memory for a destination array it copies the content of pContent[] there.
The pContent array contains two-dimensional array, written in line, that is, the first element is [iLow1, iLow2], then goes [iLow1, iLow2+1], and so on.

int getlowbound(int iBoundNum)
int gethighbound(int iBoundNum)

Get lower and upper boundaries of one-dimensional array with number iBoundNum.

raw_vector getcolumn(int iColumn, int iRowStart, int iRowEnd)
const_raw_vector getcolumn(int iColumn, int iRowStart, int iRowEnd) const
The methods are used by the basic subroutines of linear algebra to get access to the internal memory of the array. The methods return the object holding the pointer to the part of column iColumn (starting from the line iRowStart and finishing with the line iRowEnd).
The iColumn parameter must be the valid column number (that is be within the boundaries of the array). If iRowEnd<iRowStart, then an empty column is considered to be set.

raw_vector getrow(int iRow, int iColumnStart, int iColumnEnd)
const_raw_vector getrow(int iRow, int iColumnStart, int iColumnEnd) const
The methods are used by the basic subroutines of linear algebra to get access to the internal memory of the array. The methods return the object holding the pointer to the part of line iRow (starting from the column iColumnStart and finishing with the column iColumnEnd).
The iRow parameter must be the valid line number (that is be within the boundaries of the array). If iColumnEnd<iColumnStart, then an empty line is considered to be set.

int getstride() const
Returns stride (in bytes), i.e. span between first elements of adjacent rows.

Basic subroutines of linear algebra

Basic subroutines of linear algebra included into the AP library are close by their functions to the Level 1 BLAS, allowing to perform the simplest operations with vectors and with the matrix lines and columns.

General and slow interface

Subroutines should be used in the following way. First you need to get an object of the raw_vector type or const_raw_vector type, pointing to the part of the matrix or array being processed using the methods getcolumn/getrow (for the matrix), or getvector (for the array). The object holds the pointer for the line (or column) start, the number of elements in the processed line (column), and the interval between the two adjacent elements. When using a standard scheme for matrixes storage in the memory (that is, by lines), the interval between the elements of one line equals 1, and the interval between the adjacent elements of one column equals the number of columns. The received object is transferred as argument to the corresponding subroutine, which performs operations on the matrix part pointed by the internal object pointer.

Below is given the list of basic subroutines of linear algebra, available in the AP library.

template<class T> T vdotproduct(const_raw_vector<T> v1, const_raw_vector<T> v2)
The subroutine calculates the scalar product of transferred vectors.

template<class T> void vmove(raw_vector<T> vdst, const_raw_vector<T> vsrc)
template<class T> void vmoveneg(raw_vector<T> vdst, const_raw_vector<T> vsrc)
template<class T, class T2> void vmove(raw_vector<T> vdst, const_raw_vector<T> vsrc, T2 alpha)
This subroutine set is used to copy one vector content to another vector using different methods: simple copy, copy multiplied by -1, copy multiplied by a number.

template<class T> void vadd(raw_vector<T> vdst, const_raw_vector<T> vsrc)
template<class T, class T2> void vadd(raw_vector<T> vdst, const_raw_vector<T> vsrc, T2 alpha)
This subroutine set is used to add one vector to another using different methods: simple addition or addition multiplied by a number.

template<class T> void vsub(raw_vector<T> vdst, const_raw_vector<T> vsrc)
template<class T, class T2> void vsub(raw_vector<T> vdst, const_raw_vector<T> vsrc, T2 alpha)
This subroutine set is used to subtract one vector from another using different methods: simple subtraction or subtraction multiplied by a number.

template<class T, class T2> void vmul(raw_vector<T> vdst, T2 alpha)
Multiplies vector by a number and stores the result in the same place.

Alternative syntax (faster)

If both operands are vectors/rows with interval between the elements equals 1 and length equals N, alternative syntax can be used.

template<class T> T vdotproduct(const T *v1, const T *v2, int N)
template<class T> void vmove(T *vdst, const T *vsrc, int N)
template<class T> void vmoveneg(T *vdst, const T *vsrc, int N)
template<class T, class T2> void vmove(T *vdst, const T *vsrc, int N, T2 alpha)
template<class T> void vadd(T *vdst, const T *vsrc, int N)
template<class T, class T2> void vadd(T *vdst, const T *vsrc, int N, T2 alpha)
template<class T> void vsub(T *vdst, const T *vsrc, int N)
template<class T, class T2> void vsub(T *vdst, const T *vsrc, int N, T2 alpha)
template<class T, class T2> void vmul(T *vdst, int N, T2 alpha)

ap::complex class

AP library includes the ap::complex class that allows operations with compex numbers. Access to real and imaginary parts of complex number is implemented through the public fields x and y. Arithmetical operations are supported, the same as with embedded data types, by overloading of operations: addition, subtraction, multiplication and division. Addition, subtraction and multiplication are performed by a usual way (i.e., according to their definition which can be found in any textdook in algebra), division is performed using so called "safe" algorithm that could never cause overflow when calculating intermediate results. The library also includes several functions performing elementary operations with complex numbers.

const double abscomplex(const ap::complex &z)
Returns the modulus of complex number z. It should be noted that the modulus calculation is performed using so called "safe" algorithm, that could never cause overflow when calculating intermediate results.

const ap::complex conj(const ap::complex &z)
Returns complex conjugate of z.

const ap::complex csqr(const ap::complex &z)
Returns the square of z.

AMP library description

Sections

Introduction

The AMP library (ALGLIB Multiple Precision) is an object oriented interface for the MPFR library. This interface was created as part of the ALGLIB project. In order to illustrate the reason why it was necessary to create yet another interface for MPFR, we should first give a brief introduction to the ALGLIB project.

The goal of the project is to create a numerical analysis library which provides programs in different programming languages for each algorithm. This goal is attained by using automatic translation. Algorithms are stored as a pseudocode which can then be translated into the required programming language. Pseudocode syntax allows automatic translation into the most popular programming languages (e.g. C++, C#, Pascal etc.). It should be noted that, in a number of cases, programs in pseudocode can be modified to use multiple precision arithmetic. For this purpose, the AMP library is used as an interface for MPFR substituting standard double precision data type.

Programs using AMP were originally implemented to work with a 64-bit accuracy. This fact has an impact on a library design. Because of this, the following features had appeared: compile time precision definition (in AMP accuracy is given by using the template class parameterization) and prohibition to operate with numbers of different precisions (e.g. addition of 256 bit and 512 bit numbers). These restrictions are necessary in order to use the library within the ALGLIB project.

It should be noted that although AMP was designed especially for the ALGLIB project, it could be useful as just an MPFR interface. Particularly, automatic garbage collection could be useful for a number of applications. It is this part of functionality that we are going to develop, because the ALGLIB part has already been developed accurately enough and doesn't need any extension.

AMP basics

Real numbers. MPFR (and, therefore, AMP) is almost compliant with IEE-754 which describes real number arithmetic implementation requirements. The following features are available: signed zeros, finite numbers, positive and negative infinites, NANs. Denormalized numbers are not supported.

Operations and rounding. MPFR supports four types of rounding: round towards plus infinity, round towards minus infinity, round to zero and round to the nearest representable value. AMP always uses the last one. Thus, all the real number operations are performed by using exact rounding. Special values (infinites or NaNs) appearing during the calculation don't cause exception generation and the calculation is continued under normal conditions. Operations with infinites and signed zeros are performed according to the rules of limits.

Precision. Real number precision is a number of bits in a floating-point mantissa, including the high-order one (non-zero bit). For example, if the precision of a given number is equal to 53, this number is of a standard double type. The AMP library allows to set any precision no less than 32 (which represents the signed long type) including numbers which are not powers of two. The restriction was added due to compatibility reasons: a number of mathematical programs suppose that we can set any integer value to the real variable without any loss of accuracy.

Compile time precision definition. The main feature of the AMP library is that class performing multiple precision real numbers operations is a template class with one parameter specifying precision of a stored value. Hence, the precision to perform the calculation with is known at the compile time (on template specialization) and could not be changed in run-time. This restriction was introduced knowingly. The reason is that the AMP library was designed to be used mainly within the ALGLIB project. In that case, the AMP library is used by the automatically translated programs. Automatically generated codes cannot work with dynamic precision real numbers due to the number of problems. For example, we can't say what precision ought to be used when some elements of the matrix have 128 bit precision and some have 256 bit. That's the reason why this restriction was introduced.

Mixed precision operands. There are several floating point types in C++: float, double, long double. At that, if operands have a different precision, it is usually performed by automatic casting from less to finer precision (e.g. sum of float and double is double). The AMP library does not allow automatic casting of operands with different precision. For example, the sum of amp::ampf<128> and amp::ampf<144> will cause compilation errors because it is required to cast operand types. It is partly defined by C++ restrictions (which could be overcome), and partly by the developer's conviction that in numerical analysis such functionality is very seldom required.

Garbage collection. Since in MPFR the memory for real numbers is allocated dynamically, real number initialization takes a great deal of time comparatively. In order to increase performance, the AMP library doesn't free used memory, but puts needless variables into the list. When it is necessary to create a new real number, it is taken from the list.

No multithreading. The AMP library doesn't support multithreading. The library can work in a multithreaded environment, but it can't be used by two or more threads simultaneously.

Exception classes

The AMP library includes the following exception classes:

  1. amp::incorrectPrecision - generated when the real number of an invalid precision is initialized (less than 32 or more than MPFR library limitations)
  2. amp::invalidConversion - generated when the real number to be rounded using trunc/round/ceil/floor doesn't fit in the signed long type
  3. amp::internalError - generated when an internal AMP library error occurs (it should not happen under normal conditions)

amp::mpfr_record and amp::mpfr_storage classes

These classes are service classes. Although these classes are available to the programmer, he shouldn't use them, because they aren't intended to be used outside the AMP library.

amp::ampf class

Key Concepts

amp::ampf class is a template class incapsulating real numbers in MPFR library format. Class implements memory allocation, memory release and garbage collection. Arithmetic operations are overloaded. In most cases, code using this class is almost the same as code working with real numbers.

amp::ampf members


    template<unsigned int Precision>
    class ampf
    {
    public:
        ~ampf();

        ampf ();        
        ampf (long double v);
        ampf (double v);
        ampf (float v);
        ampf (signed long v);
        ampf (unsigned long v);
        ampf (signed int v);
        ampf (unsigned int v);
        ampf (signed short v);
        ampf (unsigned short v);
        ampf (signed char v);
        ampf (unsigned char v);
        ampf (const std::string &s);
        ampf (const char *s);
        
        ampf(const ampf& r);
        template<unsigned int Precision2>
        ampf(const ampf<Precision2>& r);

        ampf& operator= (long double v);
        ampf& operator= (double v);
        ampf& operator= (float v);
        ampf& operator= (signed long v);
        ampf& operator= (unsigned long v);
        ampf& operator= (signed int v);
        ampf& operator= (unsigned int v);
        ampf& operator= (signed short v);
        ampf& operator= (unsigned short v);
        ampf& operator= (signed char v);
        ampf& operator= (unsigned char v);
        ampf& operator= (const char *s);
        ampf& operator= (const std::string &s);
        ampf& operator= (const ampf& r);
        template<unsigned int Precision2>
        ampf& operator= (const ampf<Precision2>& r);
        
        template<class T>
        ampf& operator+=(const T& v);
        template<class T>
        ampf& operator-=(const T& v);
        template<class T>
        ampf& operator*=(const T& v);
        template<class T>
        ampf& operator/=(const T& v);
        
        mpfr_srcptr getReadPtr() const;
        mpfr_ptr getWritePtr();
        
        bool isFiniteNumber() const;
        bool isPositiveNumber() const;
        bool isZero() const;
        bool isNegativeNumber() const;
        const ampf getUlpOf();

        double toDouble() const;
        std::string toHex() const;
        std::string toDec() const;
        
        static const ampf getUlpOf(const ampf &x);
        static const ampf getUlp();
        static const ampf getUlp256();
        static const ampf getUlp512();
        static const ampf getMaxNumber();
        static const ampf getMinNumber();
        static const ampf getAlgoPascalEpsilon();
        static const ampf getAlgoPascalMaxNumber();
        static const ampf getAlgoPascalMinNumber();
        static const ampf getRandom();
    };

Variable Declaration and Initialization

amp::ampf is a template class having one parameter which specifies the precision of a real number. Thus, it is necessary to determine what degree of precision is required before using this class. A number of constructors are defined for class initialization. They can take instance of the same class, integer numbers, standard real numbers (float, double, long double) and strings as the input. Strings can have fractional ("12.34") or exponential form, at that the mantissa could be decimal or hexadecimal (see MPFR library description).

Example:


    int i = 3;
    double d = 2.5;
    amp::ampf<128> a(i), b(d), c("2.33");
    a = "1.56E-3";
    b = a;

Text Conversions

To convert a number into a text form, you can use member functions toDec() and toHex(). They return decimal and hexadecimal exponential representation of a number (at that, the hexadecimal conversion is performed faster than the decimal one, therefore if performance is critical it is better to use the hexadecimal conversion). You can't set the format of the result: all significant digits of the number are output.

Example:


    amp::ampf<128> a;
    a = "11";
    printf("%s\n", amp::sqrt(a).toDec().c_str());

Properties Checking

In order to check whether a real number is finite (as opposed to one of the special values), positive, zero or negative, use member function isFiniteNumber(), isPositiveNumber(), isZero(), isNegativeNumber().

Information Functions

A number of functions were designed to get information about data type properties:

Access to the mpfr_t Pointer

The getReadPtr() and getWritePtr() functions provide access to mpfr_t pointer. The getReadPtr() function allows to get a read-only pointer, which can be passed into an MPFR library function as an argument. The getWritePtr() function allows to get a writable pointer, which can be passed into an MPFR library function as a place to store the result in.

It should be noted that the AMP library uses copy-on-write technique, i.e. pointer which has been received by getReadPtr() could be common to several objects. Thus, such a pointer must not be used in operations which can change its contents.

A pointer which has been received by getWritePtr() can be used in operations which can change its contents, but these operations should not change the precision of a pointer contents. You should take into account that due to its using copy-on-write technique, the pointer which has been received by getWritePtr() could differ from the pointer which had previously been received by getReadPtr().

Random Numbers Generation

In order to generate random numbers from [0, 1), use the static function getRandom(). At first, the call function initializes a random number generator using system time.

Arithmetic Operations

amp::ampf class overloads comparison, addition, subtraction, multiplication and division operations. The result is rounded by using exact rounding. Two objects of amp::ampf class with the same precision could be the operands. It is acceptable to use objects of amp::ampf class and variables of standard data types. You can also use two objects of amp::ampf class with different precision as operands, but one of them should be casted to precision of another one.

Example:


    amp::ampf<128> a(3), b(4), c(99);
    amp::ampf<256> d(1);

    a = b+c;          //
    b = c+2.3;        // acceptable  expressions
    c = (2*a-5)/b;    //

    a = b+d;          // compile time error -
                      // arguments of a different type

    a = b + amp::ampf<128>(d); // this is the right way

    a = d;            // this is also acceptable

Mathematical functions

All functions described here are members of the amp namespace. Thus, when you call, for example, the sqr function, you should write amp::sqr. Unless otherwise specified, the result is rounded up by using exact rounding.

template<unsigned int Precision>
const ampf<Precision> sqr(const ampf<Precision> &x)
Returns the square of x.

template<unsigned int Precision>
const int sign(const ampf<Precision> &x)
Returns:
+1, if X>0
-1, if X<0
0, if X=0

template<unsigned int Precision>
const ampf<Precision> abs(const ampf<Precision> &x)
Returns the absolute value of x.

template<unsigned int Precision>
const ampf<Precision> maximum(const ampf<Precision> &x, const ampf<Precision> &y)
Returns the maximum of two integers.

template<unsigned int Precision>
const ampf<Precision> minimum(const ampf<Precision> &x, const ampf<Precision> &y)
Returns the minimum of two real numbers.

template<unsigned int Precision>
const ampf<Precision> sqrt(const ampf<Precision> &x)
Returns the square root of x.

template<unsigned int Precision>
const signed long trunc(const ampf<Precision> &x)
Rounds towards zero. Throws amp::invalidConversion if result doesn't fit in the in the signed long type.

template<unsigned int Precision>
const ampf<Precision> frac(const ampf<Precision> &x)
Returns x-trunc(x).

template<unsigned int Precision>
const signed long floor(const ampf<Precision> &x)
Rounds towards minus infinity. Throws amp::invalidConversion if result doesn't fit in the in the signed long type.

template<unsigned int Precision>
const signed long ceil(const ampf<Precision> &x)
Rounds towards plus infinity. Throws amp::invalidConversion if result doesn't fit in the in the signed long type.

template<unsigned int Precision>
const signed long round(const ampf<Precision> &x)
Rounds towards nearest integer, rounding halfway cases away from zero. Throws amp::invalidConversion if result doesn't fit in the in the signed long type.

template<unsigned int Precision>
const ampf<Precision> frexp2(const ampf<Precision> &x, mp_exp_t *exponent)
Splits a number into mantissa and exponent.

template<unsigned int Precision>
const ampf<Precision> ldexp2(const ampf<Precision> &x, mp_exp_t exponent)
Calculates x*2exponent.

template<unsigned int Precision>
const ampf<Precision> pi()
Returns Pi.

template<unsigned int Precision>
const ampf<Precision> halfpi()
Returns 0.5*Pi.

template<unsigned int Precision>
const ampf<Precision> twopi()
Returns 2*Pi.

template<unsigned int Precision>
const ampf<Precision> sin(const ampf<Precision> &x)
Calculates the sine of a value.

template<unsigned int Precision>
const ampf<Precision> cos(const ampf<Precision> &x)
Calculates the cosine of a value.

template<unsigned int Precision>
const ampf<Precision> tan(const ampf<Precision> &x)
Calculates the tangent of a value.

template<unsigned int Precision>
const ampf<Precision> asin(const ampf<Precision> &x)
Calculates the arcsine of a value.

template<unsigned int Precision>
const ampf<Precision> acos(const ampf<Precision> &x)
Calculates the arccosine of a value.

template<unsigned int Precision>
const ampf<Precision> atan(const ampf<Precision> &x)
Calculates the arctangent of a value.

template<unsigned int Precision>
const ampf<Precision> atan2(const ampf<Precision> &y, const ampf<Precision> &x)
Calculates the arc tangent of y/x. It produces correct results even when the resulting angle is near pi/2 or -pi/2 (x near 0).

template<unsigned int Precision>
const ampf<Precision> log(const ampf<Precision> &x)
Calculates the natural logarithm of x.

template<unsigned int Precision>
const ampf<Precision> log2(const ampf<Precision> &x)
Calculates the binary logarithm of x.

template<unsigned int Precision>
const ampf<Precision> log10(const ampf<Precision> &x)
Calculates the base 10 logarithm of x.

template<unsigned int Precision>
const ampf<Precision> exp(const ampf<Precision> &_x)
Calculates the exponential e to the x.

template<unsigned int Precision>
const ampf<Precision> sinh(const ampf<Precision> &x)
Calculates the hyperbolic sine of a value.

template<unsigned int Precision>
const ampf<Precision> cosh(const ampf<Precision> &_x)
Calculates the hyperbolic cosine of a value.

template<unsigned int Precision>
const ampf<Precision> tanh(const ampf<Precision> &_x)
Calculates the hyperbolic tangent of a value.

template<unsigned int Precision>
const ampf<Precision> pow(const ampf<Precision> &x, const ampf<Precision> &y)
Calculates x to the power of y.

Vectors, matrices and basic linear algebra operations

Introduction

The AP library, part of the ALGLIB project, includes vector and matrix template classes as well as subroutines performing basic linear algebra operations (Level 1 BLAS). Since the AP library implements matrix classes using templates, you can use these classes when working with matrices of multiple precision numbers. Therefore, the AMP library doesn't include vector and matrix classes. Basic linear algebra operations are also implemented by using templates in the AP library, but AMP includes its own classes specially designed to work with multiple precision numbers. The AMP library subroutines are optimized to reduce memory manager loading, thus it is better to use these subroutines when performing vector operations over multiple precision numbers.

Description

template<unsigned int Precision>
ampf<Precision> vDotProduct(ap::const_raw_vector< ampf<Precision> > v1, ap::const_raw_vector< ampf<Precision> > v2)
The subroutine calculates the scalar product of two vectors. For more information about ap::const_raw_vector see description of AP library.

template<unsigned int Precision>
void vMove(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc)
This function copies vSrc vector to vDst.

template<unsigned int Precision>
void vMoveNeg(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc)
This function copies -vSrc vector to vDst.

template<unsigned int Precision, class T2>
void vMove(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc, T2 alpha)
This function copies alpha*vSrc vector to vDst.

template<unsigned int Precision>
void vAdd(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc)
This function adds vSrc vector to vDst.

template<unsigned int Precision, class T2>
void vAdd(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc, T2 alpha)
This function adds alpha*vSrc vector to vDst.

template<unsigned int Precision>
void vSub(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc)
This function subtracts vSrc vector from vDst vector.

template<unsigned int Precision, class T2>
void vSub(ap::raw_vector< ampf<Precision> > vDst, ap::const_raw_vector< ampf<Precision> > vSrc, T2 alpha)
This function subtracts alpha*vSrc from vDst vector.

template<unsigned int Precision, class T2>
void vMul(ap::raw_vector< ampf<Precision> > vDst, T2 alpha)
This function multiplies vector by a number and stores the result in the same place

amp::campf class

Key concepts

amp::campf class implements multiple precision complex number operations. Real and imaginary parts are stored in x and y fields of type amp::ampf. You can perform arithmetic and comparison operations over the objects of this class. Moreover, a number of standard mathematical functions are defined for this data type.

amp::ampf members


    template<unsigned int Precision>
    class campf
    {
    public:
        campf():x(0),y(0);
        campf(long double v);
        campf(double v);
        campf(float v);
        campf(signed long v);
        campf(unsigned long v);
        campf(signed int v);
        campf(unsigned int v);
        campf(signed short v);
        campf(unsigned short v);
        campf(signed char v);
        campf(unsigned char v);
        campf(const ampf<Precision> &_x);
        campf(const ampf<Precision> &_x, const ampf<Precision> &_y);
        campf(const campf &z):x(z.x),y(z.y);
        template<unsigned int Prec2>
        campf(const campf<Prec2> &z);

        campf& operator= (long double v);
        campf& operator= (double v);
        campf& operator= (float v);
        campf& operator= (signed long v);
        campf& operator= (unsigned long v);
        campf& operator= (signed int v);
        campf& operator= (unsigned int v);
        campf& operator= (signed short v);
        campf& operator= (unsigned short v);
        campf& operator= (signed char v);
        campf& operator= (unsigned char v);
        campf& operator= (const char *s);
        campf& operator= (const std::string &s);
        campf& operator= (const campf& r);
        template<unsigned int Precision2>
        campf& operator= (const campf<Precision2>& r);

        ampf<Precision> x, y;
    };

Arithmetic Operations

amp::campf class overrides comparison, addition, subtraction, multiplication and division operations. Two objects of amp::campf class of the same precision could be the operands. It is acceptable to use objects of amp::campf and amp::ampf types and variables of standard data types as operands. You can also use amp::campf instances with different precision, but they should be casted.

Standard Functions

template<unsigned int Precision>
const ampf<Precision> abscomplex(const campf<Precision> &z)
Returns the modulus of complex number z. It should be noted that the modulus calculation is performed using so called "safe" algorithm, that could never cause overflow when calculating intermediate results.

template<unsigned int Precision>
const campf<Precision> conj(const campf<Precision> &z)
Returns complex conjugate of z.

template<unsigned int Precision>
const campf<Precision> csqr(const campf<Precision> &z)
Returns the square of z.

ALGLIB reference manual

Packages and units

DiffEquations package
odesolver Ordinary differential equation solver
 
FastTransforms package
conv Fast real/complex convolution
corr Fast real/complex cross-correlation
fft Real/complex FFT
fht Real Fast Hartley Transform
 
Integration package
autogk Adaptive 1-dimensional integration
gkq Gauss-Kronrod quadrature generator
gq Gaussian quadrature generator
 
Interpolation package
idwint Inverse distance weighting: interpolation/fitting
lsfit Linear and nonlinear least-squares solvers
polint Polynomial interpolation/fitting
pspline Parametric spline interpolation
ratint Rational interpolation/fitting
spline1d 1D spline interpolation/fitting
spline2d 2D spline interpolation
 
LinAlg package
ablas Level 2 and Level 3 BLAS operations
bdsvd Bidiagonal SVD
evd Eigensolvers
inverseupdate Sherman-Morrison update of the inverse matrix
ldlt LDLT decomposition
matdet Determinant calculation
matgen Random matrix generation
matinv Matrix inverse
ortfac Real/complex QR, LQ, bi(tri)diagonal, Hessenberg decompositions
rcond Condition number estimate
schur Schur decomposition
sdet Determinant of a symmetric matrix
sinverse Symmetric inversion
spdgevd Generalized symmetric eigensolver
srcond Condition number estimate for symmetric matrices
svd Singular value decomposition
trfac LU and Cholesky decompositions
 
Optimization package
minasa ASA bound constrained optimizer
mincg Conjugate gradient optimizer
minlbfgs Limited memory BFGS optimizer
minlm Improved Levenberg-Marquardt optimizer
 
Solvers package
densesolver Dense linear system solver
ssolve Symmetric dense linear system solver
 

ablas unit

Functions

ablasblocksize
ablascomplexblocksize
ablascomplexsplitlength
ablasmicroblocksize
ablassplitlength
cmatrixcopy
cmatrixgemm
cmatrixlefttrsm
cmatrixmv
cmatrixrank1
cmatrixrighttrsm
cmatrixsyrk
cmatrixtranspose
rmatrixcopy
rmatrixgemm
rmatrixlefttrsm
rmatrixmv
rmatrixrank1
rmatrixrighttrsm
rmatrixsyrk
rmatrixtranspose

ablasblocksize function

/************************************************************************* Returns block size - subdivision size where cache-oblivious soubroutines switch to the optimized kernel. INPUT PARAMETERS A - real matrix, is passed to ensure that we didn't split complex matrix using real splitting subroutine. matrix itself is not changed. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> int ablasblocksize(const ap::template_2d_array< amp::ampf<Precision> >& a);

ablascomplexblocksize function

/************************************************************************* Block size for complex subroutines. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> int ablascomplexblocksize(const ap::template_2d_array< amp::campf<Precision> >& a);

ablascomplexsplitlength function

/************************************************************************* Complex ABLASSplitLength -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void ablascomplexsplitlength(const ap::template_2d_array< amp::campf<Precision> >& a, int n, int& n1, int& n2);

ablasmicroblocksize function

/************************************************************************* Microblock size -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> int ablasmicroblocksize();

ablassplitlength function

/************************************************************************* Splits matrix length in two parts, left part should match ABLAS block size INPUT PARAMETERS A - real matrix, is passed to ensure that we didn't split complex matrix using real splitting subroutine. matrix itself is not changed. N - length, N>0 OUTPUT PARAMETERS N1 - length N2 - length N1+N2=N, N1>=N2, N2 may be zero -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void ablassplitlength(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, int& n1, int& n2);

cmatrixcopy function

/************************************************************************* Copy Input parameters: M - number of rows N - number of columns A - source matrix, MxN submatrix is copied and transposed IA - submatrix offset (row index) JA - submatrix offset (column index) B - destination matrix IB - submatrix offset (row index) JB - submatrix offset (column index) *************************************************************************/
template<unsigned int Precision> void cmatrixcopy(int m, int n, const ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, ap::template_2d_array< amp::campf<Precision> >& b, int ib, int jb);

cmatrixgemm function

/************************************************************************* This subroutine calculates C = alpha*op1(A)*op2(B) +beta*C where: * C is MxN general matrix * op1(A) is MxK matrix * op2(B) is KxN matrix * "op" may be identity transformation, transposition, conjugate transposition Additional info: * cache-oblivious algorithm is used. * multiplication result replaces C. If Beta=0, C elements are not used in calculations (not multiplied by zero - just not referenced) * if Alpha=0, A is not used (not multiplied by zero - just not referenced) * if both Beta and Alpha are zero, C is filled by zeros. INPUT PARAMETERS N - matrix size, N>0 M - matrix size, N>0 K - matrix size, K>0 Alpha - coefficient A - matrix IA - submatrix offset JA - submatrix offset OpTypeA - transformation type: * 0 - no transformation * 1 - transposition * 2 - conjugate transposition B - matrix IB - submatrix offset JB - submatrix offset OpTypeB - transformation type: * 0 - no transformation * 1 - transposition * 2 - conjugate transposition Beta - coefficient C - matrix IC - submatrix offset JC - submatrix offset -- ALGLIB routine -- 16.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixgemm(int m, int n, int k, amp::campf<Precision> alpha, const ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, int optypea, const ap::template_2d_array< amp::campf<Precision> >& b, int ib, int jb, int optypeb, amp::campf<Precision> beta, ap::template_2d_array< amp::campf<Precision> >& c, int ic, int jc);

cmatrixlefttrsm function

/************************************************************************* This subroutine calculates op(A^-1)*X where: * X is MxN general matrix * A is MxM upper/lower triangular/unitriangular matrix * "op" may be identity transformation, transposition, conjugate transposition Multiplication result replaces X. Cache-oblivious algorithm is used. INPUT PARAMETERS N - matrix size, N>=0 M - matrix size, N>=0 A - matrix, actial matrix is stored in A[I1:I1+M-1,J1:J1+M-1] I1 - submatrix offset J1 - submatrix offset IsUpper - whether matrix is upper triangular IsUnit - whether matrix is unitriangular OpType - transformation type: * 0 - no transformation * 1 - transposition * 2 - conjugate transposition C - matrix, actial matrix is stored in C[I2:I2+M-1,J2:J2+N-1] I2 - submatrix offset J2 - submatrix offset -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlefttrsm(int m, int n, const ap::template_2d_array< amp::campf<Precision> >& a, int i1, int j1, bool isupper, bool isunit, int optype, ap::template_2d_array< amp::campf<Precision> >& x, int i2, int j2);

cmatrixmv function

/************************************************************************* Matrix-vector product: y := op(A)*x INPUT PARAMETERS: M - number of rows of op(A) M>=0 N - number of columns of op(A) N>=0 A - target matrix IA - submatrix offset (row index) JA - submatrix offset (column index) OpA - operation type: * OpA=0 => op(A) = A * OpA=1 => op(A) = A^T * OpA=2 => op(A) = A^H X - input vector IX - subvector offset IY - subvector offset OUTPUT PARAMETERS: Y - vector which stores result if M=0, then subroutine does nothing. if N=0, Y is filled by zeros. -- ALGLIB routine -- 28.01.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixmv(int m, int n, ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, int opa, ap::template_1d_array< amp::campf<Precision> >& x, int ix, ap::template_1d_array< amp::campf<Precision> >& y, int iy);

cmatrixrank1 function

/************************************************************************* Rank-1 correction: A := A + u*v' INPUT PARAMETERS: M - number of rows N - number of columns A - target matrix, MxN submatrix is updated IA - submatrix offset (row index) JA - submatrix offset (column index) U - vector #1 IU - subvector offset V - vector #2 IV - subvector offset *************************************************************************/
template<unsigned int Precision> void cmatrixrank1(int m, int n, ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, ap::template_1d_array< amp::campf<Precision> >& u, int iu, ap::template_1d_array< amp::campf<Precision> >& v, int iv);

cmatrixrighttrsm function

/************************************************************************* This subroutine calculates X*op(A^-1) where: * X is MxN general matrix * A is NxN upper/lower triangular/unitriangular matrix * "op" may be identity transformation, transposition, conjugate transposition Multiplication result replaces X. Cache-oblivious algorithm is used. INPUT PARAMETERS N - matrix size, N>=0 M - matrix size, N>=0 A - matrix, actial matrix is stored in A[I1:I1+N-1,J1:J1+N-1] I1 - submatrix offset J1 - submatrix offset IsUpper - whether matrix is upper triangular IsUnit - whether matrix is unitriangular OpType - transformation type: * 0 - no transformation * 1 - transposition * 2 - conjugate transposition C - matrix, actial matrix is stored in C[I2:I2+M-1,J2:J2+N-1] I2 - submatrix offset J2 - submatrix offset -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixrighttrsm(int m, int n, const ap::template_2d_array< amp::campf<Precision> >& a, int i1, int j1, bool isupper, bool isunit, int optype, ap::template_2d_array< amp::campf<Precision> >& x, int i2, int j2);

cmatrixsyrk function

/************************************************************************* This subroutine calculates C=alpha*A*A^H+beta*C or C=alpha*A^H*A+beta*C where: * C is NxN Hermitian matrix given by its upper/lower triangle * A is NxK matrix when A*A^H is calculated, KxN matrix otherwise Additional info: * cache-oblivious algorithm is used. * multiplication result replaces C. If Beta=0, C elements are not used in calculations (not multiplied by zero - just not referenced) * if Alpha=0, A is not used (not multiplied by zero - just not referenced) * if both Beta and Alpha are zero, C is filled by zeros. INPUT PARAMETERS N - matrix size, N>=0 K - matrix size, K>=0 Alpha - coefficient A - matrix IA - submatrix offset JA - submatrix offset OpTypeA - multiplication type: * 0 - A*A^H is calculated * 2 - A^H*A is calculated Beta - coefficient C - matrix IC - submatrix offset JC - submatrix offset IsUpper - whether C is upper triangular or lower triangular -- ALGLIB routine -- 16.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixsyrk(int n, int k, amp::ampf<Precision> alpha, const ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, int optypea, amp::ampf<Precision> beta, ap::template_2d_array< amp::campf<Precision> >& c, int ic, int jc, bool isupper);

cmatrixtranspose function

/************************************************************************* Cache-oblivous complex "copy-and-transpose" Input parameters: M - number of rows N - number of columns A - source matrix, MxN submatrix is copied and transposed IA - submatrix offset (row index) JA - submatrix offset (column index) A - destination matrix IB - submatrix offset (row index) JB - submatrix offset (column index) *************************************************************************/
template<unsigned int Precision> void cmatrixtranspose(int m, int n, const ap::template_2d_array< amp::campf<Precision> >& a, int ia, int ja, ap::template_2d_array< amp::campf<Precision> >& b, int ib, int jb);

rmatrixcopy function

/************************************************************************* Copy Input parameters: M - number of rows N - number of columns A - source matrix, MxN submatrix is copied and transposed IA - submatrix offset (row index) JA - submatrix offset (column index) B - destination matrix IB - submatrix offset (row index) JB - submatrix offset (column index) *************************************************************************/
template<unsigned int Precision> void rmatrixcopy(int m, int n, const ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, ap::template_2d_array< amp::ampf<Precision> >& b, int ib, int jb);

rmatrixgemm function

/************************************************************************* Same as CMatrixGEMM, but for real numbers. OpType may be only 0 or 1. -- ALGLIB routine -- 16.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixgemm(int m, int n, int k, amp::ampf<Precision> alpha, const ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, int optypea, const ap::template_2d_array< amp::ampf<Precision> >& b, int ib, int jb, int optypeb, amp::ampf<Precision> beta, ap::template_2d_array< amp::ampf<Precision> >& c, int ic, int jc);

rmatrixlefttrsm function

/************************************************************************* Same as CMatrixLeftTRSM, but for real matrices OpType may be only 0 or 1. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlefttrsm(int m, int n, const ap::template_2d_array< amp::ampf<Precision> >& a, int i1, int j1, bool isupper, bool isunit, int optype, ap::template_2d_array< amp::ampf<Precision> >& x, int i2, int j2);

rmatrixmv function

/************************************************************************* Matrix-vector product: y := op(A)*x INPUT PARAMETERS: M - number of rows of op(A) N - number of columns of op(A) A - target matrix IA - submatrix offset (row index) JA - submatrix offset (column index) OpA - operation type: * OpA=0 => op(A) = A * OpA=1 => op(A) = A^T X - input vector IX - subvector offset IY - subvector offset OUTPUT PARAMETERS: Y - vector which stores result if M=0, then subroutine does nothing. if N=0, Y is filled by zeros. -- ALGLIB routine -- 28.01.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixmv(int m, int n, ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, int opa, ap::template_1d_array< amp::ampf<Precision> >& x, int ix, ap::template_1d_array< amp::ampf<Precision> >& y, int iy);

rmatrixrank1 function

/************************************************************************* Rank-1 correction: A := A + u*v' INPUT PARAMETERS: M - number of rows N - number of columns A - target matrix, MxN submatrix is updated IA - submatrix offset (row index) JA - submatrix offset (column index) U - vector #1 IU - subvector offset V - vector #2 IV - subvector offset *************************************************************************/
template<unsigned int Precision> void rmatrixrank1(int m, int n, ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, ap::template_1d_array< amp::ampf<Precision> >& u, int iu, ap::template_1d_array< amp::ampf<Precision> >& v, int iv);

rmatrixrighttrsm function

/************************************************************************* Same as CMatrixRightTRSM, but for real matrices OpType may be only 0 or 1. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixrighttrsm(int m, int n, const ap::template_2d_array< amp::ampf<Precision> >& a, int i1, int j1, bool isupper, bool isunit, int optype, ap::template_2d_array< amp::ampf<Precision> >& x, int i2, int j2);

rmatrixsyrk function

/************************************************************************* Same as CMatrixSYRK, but for real matrices OpType may be only 0 or 1. -- ALGLIB routine -- 16.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixsyrk(int n, int k, amp::ampf<Precision> alpha, const ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, int optypea, amp::ampf<Precision> beta, ap::template_2d_array< amp::ampf<Precision> >& c, int ic, int jc, bool isupper);

rmatrixtranspose function

/************************************************************************* Cache-oblivous real "copy-and-transpose" Input parameters: M - number of rows N - number of columns A - source matrix, MxN submatrix is copied and transposed IA - submatrix offset (row index) JA - submatrix offset (column index) A - destination matrix IB - submatrix offset (row index) JB - submatrix offset (column index) *************************************************************************/
template<unsigned int Precision> void rmatrixtranspose(int m, int n, const ap::template_2d_array< amp::ampf<Precision> >& a, int ia, int ja, ap::template_2d_array< amp::ampf<Precision> >& b, int ib, int jb);

autogk unit

Structures

autogkreport
autogkstate

Functions

autogkiteration
autogkresults
autogksingular
autogksmooth
autogksmoothw

Examples

autogk_singular
autogk_smooth

autogkreport structure

/************************************************************************* Integration report: * TerminationType = completetion code: * -5 non-convergence of Gauss-Kronrod nodes calculation subroutine. * -1 incorrect parameters were specified * 1 OK * Rep.NFEV countains number of function calculations * Rep.NIntervals contains number of intervals [a,b] was partitioned into. *************************************************************************/
template<unsigned int Precision> class autogkreport { public: int terminationtype; int nfev; int nintervals; };

autogkstate structure

/************************************************************************* This structure stores internal state of the integration algorithm between subsequent calls of the AutoGKIteration() subroutine. *************************************************************************/
template<unsigned int Precision> class autogkstate { public: amp::ampf<Precision> a; amp::ampf<Precision> b; amp::ampf<Precision> alpha; amp::ampf<Precision> beta; amp::ampf<Precision> xwidth; amp::ampf<Precision> x; amp::ampf<Precision> xminusa; amp::ampf<Precision> bminusx; amp::ampf<Precision> f; int wrappermode; autogkinternalstate<Precision> internalstate; amp::rcommstate<Precision> rstate; amp::ampf<Precision> v; int terminationtype; int nfev; int nintervals; };

autogkiteration function

/************************************************************************* One step of adaptive integration process. Called after initialization with one of AutoGKXXX subroutines. See HTML documentation for examples. Input parameters: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with one of AutoGKXXX subroutines. If suborutine returned False, iterative proces has converged. If subroutine returned True, caller should calculate function value State.F at State.X and call AutoGKIteration again. NOTE: When integrating "difficult" functions with integrable singularities like F(x) = (x-A)^alpha * (B-x)^beta subroutine may require the value of F at points which are too close to A/B. Sometimes to calculate integral with high enough precision we may need to calculate F(A+delta) when delta is less than machine epsilon. In finite precision arithmetics A+delta will be effectively equal to A, so we may find us in situation when we are trying to calculate something like 1/sqrt(1-1). To avoid such situations, AutoGKIteration subroutine fills not only State.X field, but also State.XMinusA (which equals to X-A) and State.BMinusX (which equals to B-X) fields. If X is too close to A or B (X-A<0.001*A, or B-X<0.001*B, for example) use these fields instead of State.X -- ALGLIB -- Copyright 07.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool autogkiteration(autogkstate<Precision>& state);

Examples:   autogk_singular  autogk_smooth  

autogkresults function

/************************************************************************* Adaptive integration results Called after AutoGKIteration returned False. Input parameters: State - algorithm state (used by AutoGKIteration). Output parameters: V - integral(f(x)dx,a,b) Rep - optimization report (see AutoGKReport description) -- ALGLIB -- Copyright 14.11.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void autogkresults(const autogkstate<Precision>& state, amp::ampf<Precision>& v, autogkreport<Precision>& rep);

Examples:   autogk_singular  autogk_smooth  

autogksingular function

/************************************************************************* Integration on a finite interval [A,B]. Integrand have integrable singularities at A/B. F(X) must diverge as "(x-A)^alpha" at A, as "(B-x)^beta" at B, with known alpha/beta (alpha>-1, beta>-1). If alpha/beta are not known, estimates from below can be used (but these estimates should be greater than -1 too). One of alpha/beta variables (or even both alpha/beta) may be equal to 0, which means than function F(x) is non-singular at A/B. Anyway (singular at bounds or not), function F(x) is supposed to be continuous on (A,B). Fast-convergent algorithm based on a Gauss-Kronrod formula is used. Result is calculated with accuracy close to the machine precision. INPUT PARAMETERS: A, B - interval boundaries (A<B, A=B or A>B) Alpha - power-law coefficient of the F(x) at A, Alpha>-1 Beta - power-law coefficient of the F(x) at B, Beta>-1 OUTPUT PARAMETERS State - structure which stores algorithm state between subsequent calls of AutoGKIteration. Used for reverse communication. This structure should be passed to the AutoGKIteration subroutine. SEE ALSO AutoGKSmooth, AutoGKSmoothW, AutoGKIteration, AutoGKResults. -- ALGLIB -- Copyright 06.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void autogksingular(amp::ampf<Precision> a, amp::ampf<Precision> b, amp::ampf<Precision> alpha, amp::ampf<Precision> beta, autogkstate<Precision>& state);

Examples:   autogk_singular  

autogksmooth function

/************************************************************************* Integration of a smooth function F(x) on a finite interval [a,b]. Fast-convergent algorithm based on a Gauss-Kronrod formula is used. Result is calculated with accuracy close to the machine precision. Algorithm works well only with smooth integrands. It may be used with continuous non-smooth integrands, but with less performance. It should never be used with integrands which have integrable singularities at lower or upper limits - algorithm may crash. Use AutoGKSingular in such cases. INPUT PARAMETERS: A, B - interval boundaries (A<B, A=B or A>B) OUTPUT PARAMETERS State - structure which stores algorithm state between subsequent calls of AutoGKIteration. Used for reverse communication. This structure should be passed to the AutoGKIteration subroutine. SEE ALSO AutoGKSmoothW, AutoGKSingular, AutoGKIteration, AutoGKResults. -- ALGLIB -- Copyright 06.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void autogksmooth(amp::ampf<Precision> a, amp::ampf<Precision> b, autogkstate<Precision>& state);

Examples:   autogk_smooth  

autogksmoothw function

/************************************************************************* Integration of a smooth function F(x) on a finite interval [a,b]. This subroutine is same as AutoGKSmooth(), but it guarantees that interval [a,b] is partitioned into subintervals which have width at most XWidth. Subroutine can be used when integrating nearly-constant function with narrow "bumps" (about XWidth wide). If "bumps" are too narrow, AutoGKSmooth subroutine can overlook them. INPUT PARAMETERS: A, B - interval boundaries (A<B, A=B or A>B) OUTPUT PARAMETERS State - structure which stores algorithm state between subsequent calls of AutoGKIteration. Used for reverse communication. This structure should be passed to the AutoGKIteration subroutine. SEE ALSO AutoGKSmooth, AutoGKSingular, AutoGKIteration, AutoGKResults. -- ALGLIB -- Copyright 06.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void autogksmoothw(amp::ampf<Precision> a, amp::ampf<Precision> b, amp::ampf<Precision> xwidth, autogkstate<Precision>& state);

autogk_singular example

    autogk::autogkstate<Precision> state;
    amp::ampf<Precision> v;
    autogk::autogkreport<Precision> rep;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;
    amp::ampf<Precision> alpha;


    
    //
    // f1(x) = (1+x)*(x-a)^alpha, alpha=-0.3
    // Exact answer is (B-A)^(Alpha+2)/(Alpha+2) + (1+A)*(B-A)^(Alpha+1)/(Alpha+1)
    //
    // This code demonstrates use of the State.XMinusA (State.BMinusX) field.
    //
    // If we try to use State.X instead of State.XMinusA,
    // we will end up dividing by zero! (in 64-bit precision)
    //
    a = amp::ampf<Precision>("1.0");
    b = amp::ampf<Precision>("5.0");
    alpha = -amp::ampf<Precision>("0.9");
    autogk::autogksingular<Precision>(a, b, alpha, amp::ampf<Precision>("0.0"), state);
    while( autogk::autogkiteration<Precision>(state) )
    {
        state.f = amp::pow<Precision>(state.xminusa, alpha)*(1+state.x);
    }
    autogk::autogkresults<Precision>(state, v, rep);
    printf("integral((1+x)*(x-a)^alpha) on [%0.1lf; %0.1lf] = %0.2lf\nExact answer is %0.2lf\n",
        double(amp::ampf<Precision>(a).toDouble()),
        double(amp::ampf<Precision>(b).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(amp::pow<Precision>(b-a, alpha+2)/(alpha+2)+(1+a)*amp::pow<Precision>(b-a, alpha+1)/(alpha+1)).toDouble()));

autogk_smooth example

    autogk::autogkstate<Precision> state;
    amp::ampf<Precision> v;
    autogk::autogkreport<Precision> rep;


    
    //
    // f(x) = x*sin(x), integrated at [-pi, pi].
    // Exact answer is 2*pi
    //
    autogk::autogksmooth<Precision>(-amp::pi<Precision>(), +amp::pi<Precision>(), state);
    while( autogk::autogkiteration<Precision>(state) )
    {
        state.f = state.x*amp::sin<Precision>(state.x);
    }
    autogk::autogkresults<Precision>(state, v, rep);
    printf("integral(x*sin(x),-pi,+pi) = %0.2lf\nExact answer is %0.2lf\n",
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(2*amp::pi<Precision>()).toDouble()));

bdsvd unit

Functions

rmatrixbdsvd

rmatrixbdsvd function

/************************************************************************* Singular value decomposition of a bidiagonal matrix (extended algorithm) The algorithm performs the singular value decomposition of a bidiagonal matrix B (upper or lower) representing it as B = Q*S*P^T, where Q and P - orthogonal matrices, S - diagonal matrix with non-negative elements on the main diagonal, in descending order. The algorithm finds singular values. In addition, the algorithm can calculate matrices Q and P (more precisely, not the matrices, but their product with given matrices U and VT - U*Q and (P^T)*VT)). Of course, matrices U and VT can be of any type, including identity. Furthermore, the algorithm can calculate Q'*C (this product is calculated more effectively than U*Q, because this calculation operates with rows instead of matrix columns). The feature of the algorithm is its ability to find all singular values including those which are arbitrarily close to 0 with relative accuracy close to machine precision. If the parameter IsFractionalAccuracyRequired is set to True, all singular values will have high relative accuracy close to machine precision. If the parameter is set to False, only the biggest singular value will have relative accuracy close to machine precision. The absolute error of other singular values is equal to the absolute error of the biggest singular value. Input parameters: D - main diagonal of matrix B. Array whose index ranges within [0..N-1]. E - superdiagonal (or subdiagonal) of matrix B. Array whose index ranges within [0..N-2]. N - size of matrix B. IsUpper - True, if the matrix is upper bidiagonal. IsFractionalAccuracyRequired - accuracy to search singular values with. U - matrix to be multiplied by Q. Array whose indexes range within [0..NRU-1, 0..N-1]. The matrix can be bigger, in that case only the submatrix [0..NRU-1, 0..N-1] will be multiplied by Q. NRU - number of rows in matrix U. C - matrix to be multiplied by Q'. Array whose indexes range within [0..N-1, 0..NCC-1]. The matrix can be bigger, in that case only the submatrix [0..N-1, 0..NCC-1] will be multiplied by Q'. NCC - number of columns in matrix C. VT - matrix to be multiplied by P^T. Array whose indexes range within [0..N-1, 0..NCVT-1]. The matrix can be bigger, in that case only the submatrix [0..N-1, 0..NCVT-1] will be multiplied by P^T. NCVT - number of columns in matrix VT. Output parameters: D - singular values of matrix B in descending order. U - if NRU>0, contains matrix U*Q. VT - if NCVT>0, contains matrix (P^T)*VT. C - if NCC>0, contains matrix Q'*C. Result: True, if the algorithm has converged. False, if the algorithm hasn't converged (rare case). Additional information: The type of convergence is controlled by the internal parameter TOL. If the parameter is greater than 0, the singular values will have relative accuracy TOL. If TOL<0, the singular values will have absolute accuracy ABS(TOL)*norm(B). By default, |TOL| falls within the range of 10*Epsilon and 100*Epsilon, where Epsilon is the machine precision. It is not recommended to use TOL less than 10*Epsilon since this will considerably slow down the algorithm and may not lead to error decreasing. History: * 31 March, 2007. changed MAXITR from 6 to 12. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University October 31, 1999. *************************************************************************/
template<unsigned int Precision> bool rmatrixbdsvd(ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_1d_array< amp::ampf<Precision> > e, int n, bool isupper, bool isfractionalaccuracyrequired, ap::template_2d_array< amp::ampf<Precision> >& u, int nru, ap::template_2d_array< amp::ampf<Precision> >& c, int ncc, ap::template_2d_array< amp::ampf<Precision> >& vt, int ncvt);

conv unit

Functions

convc1d
convc1dcircular
convc1dcircularinv
convc1dinv
convc1dx
convr1d
convr1dcircular
convr1dcircularinv
convr1dinv
convr1dx

convc1d function

/************************************************************************* 1-dimensional complex convolution. For given A/B returns conv(A,B) (non-circular). Subroutine can automatically choose between three implementations: straightforward O(M*N) formula for very small N (or M), overlap-add algorithm for cases where max(M,N) is significantly larger than min(M,N), but O(M*N) algorithm is too slow, and general FFT-based formula for cases where two previois algorithms are too slow. Algorithm has max(M,N)*log(max(M,N)) complexity for any M/N. INPUT PARAMETERS A - array[0..M-1] - complex function to be transformed M - problem size B - array[0..N-1] - complex function to be transformed N - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..N+M-2]. NOTE: It is assumed that A is zero at T<0, B is zero too. If one or both functions have non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convc1d(const ap::template_1d_array< amp::campf<Precision> >& a, int m, const ap::template_1d_array< amp::campf<Precision> >& b, int n, ap::template_1d_array< amp::campf<Precision> >& r);

convc1dcircular function

/************************************************************************* 1-dimensional circular complex convolution. For given S/R returns conv(S,R) (circular). Algorithm has linearithmic complexity for any M/N. IMPORTANT: normal convolution is commutative, i.e. it is symmetric - conv(A,B)=conv(B,A). Cyclic convolution IS NOT. One function - S - is a signal, periodic function, and another - R - is a response, non-periodic function with limited length. INPUT PARAMETERS S - array[0..M-1] - complex periodic signal M - problem size B - array[0..N-1] - complex non-periodic response N - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..M-1]. NOTE: It is assumed that B is zero at T<0. If it has non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convc1dcircular(const ap::template_1d_array< amp::campf<Precision> >& s, int m, const ap::template_1d_array< amp::campf<Precision> >& r, int n, ap::template_1d_array< amp::campf<Precision> >& c);

convc1dcircularinv function

/************************************************************************* 1-dimensional circular complex deconvolution (inverse of ConvC1DCircular()). Algorithm has M*log(M)) complexity for any M (composite or prime). INPUT PARAMETERS A - array[0..M-1] - convolved periodic signal, A = conv(R, B) M - convolved signal length B - array[0..N-1] - non-periodic response N - response length OUTPUT PARAMETERS R - deconvolved signal. array[0..M-1]. NOTE: deconvolution is unstable process and may result in division by zero (if your response function is degenerate, i.e. has zero Fourier coefficient). NOTE: It is assumed that B is zero at T<0. If it has non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convc1dcircularinv(const ap::template_1d_array< amp::campf<Precision> >& a, int m, const ap::template_1d_array< amp::campf<Precision> >& b, int n, ap::template_1d_array< amp::campf<Precision> >& r);

convc1dinv function

/************************************************************************* 1-dimensional complex non-circular deconvolution (inverse of ConvC1D()). Algorithm has M*log(M)) complexity for any M (composite or prime). INPUT PARAMETERS A - array[0..M-1] - convolved signal, A = conv(R, B) M - convolved signal length B - array[0..N-1] - response N - response length, N<=M OUTPUT PARAMETERS R - deconvolved signal. array[0..M-N]. NOTE: deconvolution is unstable process and may result in division by zero (if your response function is degenerate, i.e. has zero Fourier coefficient). NOTE: It is assumed that A is zero at T<0, B is zero too. If one or both functions have non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convc1dinv(const ap::template_1d_array< amp::campf<Precision> >& a, int m, const ap::template_1d_array< amp::campf<Precision> >& b, int n, ap::template_1d_array< amp::campf<Precision> >& r);

convc1dx function

/************************************************************************* 1-dimensional complex convolution. Extended subroutine which allows to choose convolution algorithm. Intended for internal use, ALGLIB users should call ConvC1D()/ConvC1DCircular(). INPUT PARAMETERS A - array[0..M-1] - complex function to be transformed M - problem size B - array[0..N-1] - complex function to be transformed N - problem size, N<=M Alg - algorithm type: *-2 auto-select Q for overlap-add *-1 auto-select algorithm and parameters * 0 straightforward formula for small N's * 1 general FFT-based code * 2 overlap-add with length Q Q - length for overlap-add OUTPUT PARAMETERS R - convolution: A*B. array[0..N+M-1]. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convc1dx(const ap::template_1d_array< amp::campf<Precision> >& a, int m, const ap::template_1d_array< amp::campf<Precision> >& b, int n, bool circular, int alg, int q, ap::template_1d_array< amp::campf<Precision> >& r);

convr1d function

/************************************************************************* 1-dimensional real convolution. Analogous to ConvC1D(), see ConvC1D() comments for more details. INPUT PARAMETERS A - array[0..M-1] - real function to be transformed M - problem size B - array[0..N-1] - real function to be transformed N - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..N+M-2]. NOTE: It is assumed that A is zero at T<0, B is zero too. If one or both functions have non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convr1d(const ap::template_1d_array< amp::ampf<Precision> >& a, int m, const ap::template_1d_array< amp::ampf<Precision> >& b, int n, ap::template_1d_array< amp::ampf<Precision> >& r);

convr1dcircular function

/************************************************************************* 1-dimensional circular real convolution. Analogous to ConvC1DCircular(), see ConvC1DCircular() comments for more details. INPUT PARAMETERS S - array[0..M-1] - real signal M - problem size B - array[0..N-1] - real response N - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..M-1]. NOTE: It is assumed that B is zero at T<0. If it has non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convr1dcircular(const ap::template_1d_array< amp::ampf<Precision> >& s, int m, const ap::template_1d_array< amp::ampf<Precision> >& r, int n, ap::template_1d_array< amp::ampf<Precision> >& c);

convr1dcircularinv function

/************************************************************************* 1-dimensional complex deconvolution (inverse of ConvC1D()). Algorithm has M*log(M)) complexity for any M (composite or prime). INPUT PARAMETERS A - array[0..M-1] - convolved signal, A = conv(R, B) M - convolved signal length B - array[0..N-1] - response N - response length OUTPUT PARAMETERS R - deconvolved signal. array[0..M-N]. NOTE: deconvolution is unstable process and may result in division by zero (if your response function is degenerate, i.e. has zero Fourier coefficient). NOTE: It is assumed that B is zero at T<0. If it has non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convr1dcircularinv(const ap::template_1d_array< amp::ampf<Precision> >& a, int m, const ap::template_1d_array< amp::ampf<Precision> >& b, int n, ap::template_1d_array< amp::ampf<Precision> >& r);

convr1dinv function

/************************************************************************* 1-dimensional real deconvolution (inverse of ConvC1D()). Algorithm has M*log(M)) complexity for any M (composite or prime). INPUT PARAMETERS A - array[0..M-1] - convolved signal, A = conv(R, B) M - convolved signal length B - array[0..N-1] - response N - response length, N<=M OUTPUT PARAMETERS R - deconvolved signal. array[0..M-N]. NOTE: deconvolution is unstable process and may result in division by zero (if your response function is degenerate, i.e. has zero Fourier coefficient). NOTE: It is assumed that A is zero at T<0, B is zero too. If one or both functions have non-zero values at negative T's, you can still use this subroutine - just shift its result correspondingly. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convr1dinv(const ap::template_1d_array< amp::ampf<Precision> >& a, int m, const ap::template_1d_array< amp::ampf<Precision> >& b, int n, ap::template_1d_array< amp::ampf<Precision> >& r);

convr1dx function

/************************************************************************* 1-dimensional real convolution. Extended subroutine which allows to choose convolution algorithm. Intended for internal use, ALGLIB users should call ConvR1D(). INPUT PARAMETERS A - array[0..M-1] - complex function to be transformed M - problem size B - array[0..N-1] - complex function to be transformed N - problem size, N<=M Alg - algorithm type: *-2 auto-select Q for overlap-add *-1 auto-select algorithm and parameters * 0 straightforward formula for small N's * 1 general FFT-based code * 2 overlap-add with length Q Q - length for overlap-add OUTPUT PARAMETERS R - convolution: A*B. array[0..N+M-1]. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void convr1dx(const ap::template_1d_array< amp::ampf<Precision> >& a, int m, const ap::template_1d_array< amp::ampf<Precision> >& b, int n, bool circular, int alg, int q, ap::template_1d_array< amp::ampf<Precision> >& r);

corr unit

Functions

corrc1d
corrc1dcircular
corrr1d
corrr1dcircular

corrc1d function

/************************************************************************* 1-dimensional complex cross-correlation. For given Pattern/Signal returns corr(Pattern,Signal) (non-circular). Correlation is calculated using reduction to convolution. Algorithm with max(N,N)*log(max(N,N)) complexity is used (see ConvC1D() for more info about performance). IMPORTANT: for historical reasons subroutine accepts its parameters in reversed order: CorrC1D(Signal, Pattern) = Pattern x Signal (using traditional definition of cross-correlation, denoting cross-correlation as "x"). INPUT PARAMETERS Signal - array[0..N-1] - complex function to be transformed, signal containing pattern N - problem size Pattern - array[0..M-1] - complex function to be transformed, pattern to search withing signal M - problem size OUTPUT PARAMETERS R - cross-correlation, array[0..N+M-2]: * positive lags are stored in R[0..N-1], R[i] = sum(conj(pattern[j])*signal[i+j] * negative lags are stored in R[N..N+M-2], R[N+M-1-i] = sum(conj(pattern[j])*signal[-i+j] NOTE: It is assumed that pattern domain is [0..M-1]. If Pattern is non-zero on [-K..M-1], you can still use this subroutine, just shift result by K. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void corrc1d(const ap::template_1d_array< amp::campf<Precision> >& signal, int n, const ap::template_1d_array< amp::campf<Precision> >& pattern, int m, ap::template_1d_array< amp::campf<Precision> >& r);

corrc1dcircular function

/************************************************************************* 1-dimensional circular complex cross-correlation. For given Pattern/Signal returns corr(Pattern,Signal) (circular). Algorithm has linearithmic complexity for any M/N. IMPORTANT: for historical reasons subroutine accepts its parameters in reversed order: CorrC1DCircular(Signal, Pattern) = Pattern x Signal (using traditional definition of cross-correlation, denoting cross-correlation as "x"). INPUT PARAMETERS Signal - array[0..N-1] - complex function to be transformed, periodic signal containing pattern N - problem size Pattern - array[0..M-1] - complex function to be transformed, non-periodic pattern to search withing signal M - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..M-1]. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void corrc1dcircular(const ap::template_1d_array< amp::campf<Precision> >& signal, int m, const ap::template_1d_array< amp::campf<Precision> >& pattern, int n, ap::template_1d_array< amp::campf<Precision> >& c);

corrr1d function

/************************************************************************* 1-dimensional real cross-correlation. For given Pattern/Signal returns corr(Pattern,Signal) (non-circular). Correlation is calculated using reduction to convolution. Algorithm with max(N,N)*log(max(N,N)) complexity is used (see ConvC1D() for more info about performance). IMPORTANT: for historical reasons subroutine accepts its parameters in reversed order: CorrR1D(Signal, Pattern) = Pattern x Signal (using traditional definition of cross-correlation, denoting cross-correlation as "x"). INPUT PARAMETERS Signal - array[0..N-1] - real function to be transformed, signal containing pattern N - problem size Pattern - array[0..M-1] - real function to be transformed, pattern to search withing signal M - problem size OUTPUT PARAMETERS R - cross-correlation, array[0..N+M-2]: * positive lags are stored in R[0..N-1], R[i] = sum(pattern[j]*signal[i+j] * negative lags are stored in R[N..N+M-2], R[N+M-1-i] = sum(pattern[j]*signal[-i+j] NOTE: It is assumed that pattern domain is [0..M-1]. If Pattern is non-zero on [-K..M-1], you can still use this subroutine, just shift result by K. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void corrr1d(const ap::template_1d_array< amp::ampf<Precision> >& signal, int n, const ap::template_1d_array< amp::ampf<Precision> >& pattern, int m, ap::template_1d_array< amp::ampf<Precision> >& r);

corrr1dcircular function

/************************************************************************* 1-dimensional circular real cross-correlation. For given Pattern/Signal returns corr(Pattern,Signal) (circular). Algorithm has linearithmic complexity for any M/N. IMPORTANT: for historical reasons subroutine accepts its parameters in reversed order: CorrR1DCircular(Signal, Pattern) = Pattern x Signal (using traditional definition of cross-correlation, denoting cross-correlation as "x"). INPUT PARAMETERS Signal - array[0..N-1] - real function to be transformed, periodic signal containing pattern N - problem size Pattern - array[0..M-1] - real function to be transformed, non-periodic pattern to search withing signal M - problem size OUTPUT PARAMETERS R - convolution: A*B. array[0..M-1]. -- ALGLIB -- Copyright 21.07.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void corrr1dcircular(const ap::template_1d_array< amp::ampf<Precision> >& signal, int m, const ap::template_1d_array< amp::ampf<Precision> >& pattern, int n, ap::template_1d_array< amp::ampf<Precision> >& c);

densesolver unit

Functions

cmatrixlusolve
cmatrixlusolvem
cmatrixmixedsolve
cmatrixmixedsolvem
cmatrixsolve
cmatrixsolvem
hpdmatrixcholeskysolve
hpdmatrixcholeskysolvem
hpdmatrixsolve
hpdmatrixsolvem
rmatrixlusolve
rmatrixlusolvem
rmatrixmixedsolve
rmatrixmixedsolvem
rmatrixsolve
rmatrixsolvels
rmatrixsolvem
spdmatrixcholeskysolve
spdmatrixcholeskysolvem
spdmatrixsolve
spdmatrixsolvem

cmatrixlusolve function

/************************************************************************* Dense solver. Same as RMatrixLUSolve(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * O(N^2) complexity * condition number estimation No iterative refinement is provided because exact form of original matrix is not known to subroutine. Use CMatrixSolve or CMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS LUA - array[0..N-1,0..N-1], LU decomposition, CMatrixLU result P - array[0..N-1], pivots array, CMatrixLU result N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlusolve(const ap::template_2d_array< amp::campf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_1d_array< amp::campf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::campf<Precision> >& x);

cmatrixlusolvem function

/************************************************************************* Dense solver. Same as RMatrixLUSolveM(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * O(M*N^2) complexity * condition number estimation No iterative refinement is provided because exact form of original matrix is not known to subroutine. Use CMatrixSolve or CMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS LUA - array[0..N-1,0..N-1], LU decomposition, RMatrixLU result P - array[0..N-1], pivots array, RMatrixLU result N - size of A B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlusolvem(const ap::template_2d_array< amp::campf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_2d_array< amp::campf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::campf<Precision> >& x);

cmatrixmixedsolve function

/************************************************************************* Dense solver. Same as RMatrixMixedSolve(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix LUA - array[0..N-1,0..N-1], LU decomposition, CMatrixLU result P - array[0..N-1], pivots array, CMatrixLU result N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolveM Rep - same as in RMatrixSolveM X - same as in RMatrixSolveM -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixmixedsolve(const ap::template_2d_array< amp::campf<Precision> >& a, const ap::template_2d_array< amp::campf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_1d_array< amp::campf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::campf<Precision> >& x);

cmatrixmixedsolvem function

/************************************************************************* Dense solver. Same as RMatrixMixedSolveM(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(M*N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix LUA - array[0..N-1,0..N-1], LU decomposition, CMatrixLU result P - array[0..N-1], pivots array, CMatrixLU result N - size of A B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolveM Rep - same as in RMatrixSolveM X - same as in RMatrixSolveM -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixmixedsolvem(const ap::template_2d_array< amp::campf<Precision> >& a, const ap::template_2d_array< amp::campf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_2d_array< amp::campf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::campf<Precision> >& x);

cmatrixsolve function

/************************************************************************* Dense solver. Same as RMatrixSolve(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(N^3) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixsolve(const ap::template_2d_array< amp::campf<Precision> >& a, int n, const ap::template_1d_array< amp::campf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::campf<Precision> >& x);

cmatrixsolvem function

/************************************************************************* Dense solver. Same as RMatrixSolveM(), but for complex matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(N^3+M*N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A B - array[0..N-1,0..M-1], right part M - right part size RFS - iterative refinement switch: * True - refinement is used. Less performance, more precision. * False - refinement is not used. More performance, less precision. OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixsolvem(const ap::template_2d_array< amp::campf<Precision> >& a, int n, const ap::template_2d_array< amp::campf<Precision> >& b, int m, bool rfs, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::campf<Precision> >& x);

hpdmatrixcholeskysolve function

/************************************************************************* Dense solver. Same as RMatrixLUSolve(), but for HPD matrices represented by their Cholesky decomposition. Algorithm features: * automatic detection of degenerate cases * O(N^2) complexity * condition number estimation * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS CHA - array[0..N-1,0..N-1], Cholesky decomposition, SPDMatrixCholesky result N - size of A IsUpper - what half of CHA is provided B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixcholeskysolve(const ap::template_2d_array< amp::campf<Precision> >& cha, int n, bool isupper, const ap::template_1d_array< amp::campf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::campf<Precision> >& x);

hpdmatrixcholeskysolvem function

/************************************************************************* Dense solver. Same as RMatrixLUSolveM(), but for HPD matrices represented by their Cholesky decomposition. Algorithm features: * automatic detection of degenerate cases * O(M*N^2) complexity * condition number estimation * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS CHA - array[0..N-1,0..N-1], Cholesky decomposition, HPDMatrixCholesky result N - size of CHA IsUpper - what half of CHA is provided B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixcholeskysolvem(const ap::template_2d_array< amp::campf<Precision> >& cha, int n, bool isupper, const ap::template_2d_array< amp::campf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::campf<Precision> >& x);

hpdmatrixsolve function

/************************************************************************* Dense solver. Same as RMatrixSolve(), but for Hermitian positive definite matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * O(N^3) complexity * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A IsUpper - what half of A is provided B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Returns -3 for non-HPD matrices. Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixsolve(const ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, const ap::template_1d_array< amp::campf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::campf<Precision> >& x);

hpdmatrixsolvem function

/************************************************************************* Dense solver. Same as RMatrixSolveM(), but for Hermitian positive definite matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * O(N^3+M*N^2) complexity * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A IsUpper - what half of A is provided B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve. Returns -3 for non-HPD matrices. Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixsolvem(const ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, const ap::template_2d_array< amp::campf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::campf<Precision> >& x);

rmatrixlusolve function

/************************************************************************* Dense solver. This subroutine solves a system A*X=B, where A is NxN non-denegerate real matrix given by its LU decomposition, X and B are NxM real matrices. Algorithm features: * automatic detection of degenerate cases * O(N^2) complexity * condition number estimation No iterative refinement is provided because exact form of original matrix is not known to subroutine. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS LUA - array[0..N-1,0..N-1], LU decomposition, RMatrixLU result P - array[0..N-1], pivots array, RMatrixLU result N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlusolve(const ap::template_2d_array< amp::ampf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_1d_array< amp::ampf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

rmatrixlusolvem function

/************************************************************************* Dense solver. Similar to RMatrixLUSolve() but solves task with multiple right parts (where b and x are NxM matrices). Algorithm features: * automatic detection of degenerate cases * O(M*N^2) complexity * condition number estimation No iterative refinement is provided because exact form of original matrix is not known to subroutine. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS LUA - array[0..N-1,0..N-1], LU decomposition, RMatrixLU result P - array[0..N-1], pivots array, RMatrixLU result N - size of A B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlusolvem(const ap::template_2d_array< amp::ampf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_2d_array< amp::ampf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::ampf<Precision> >& x);

rmatrixmixedsolve function

/************************************************************************* Dense solver. This subroutine solves a system A*x=b, where BOTH ORIGINAL A AND ITS LU DECOMPOSITION ARE KNOWN. You can use it if for some reasons you have both A and its LU decomposition. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix LUA - array[0..N-1,0..N-1], LU decomposition, RMatrixLU result P - array[0..N-1], pivots array, RMatrixLU result N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolveM Rep - same as in RMatrixSolveM X - same as in RMatrixSolveM -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixmixedsolve(const ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_2d_array< amp::ampf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_1d_array< amp::ampf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

rmatrixmixedsolvem function

/************************************************************************* Dense solver. Similar to RMatrixMixedSolve() but solves task with multiple right parts (where b and x are NxM matrices). Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(M*N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix LUA - array[0..N-1,0..N-1], LU decomposition, RMatrixLU result P - array[0..N-1], pivots array, RMatrixLU result N - size of A B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolveM Rep - same as in RMatrixSolveM X - same as in RMatrixSolveM -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixmixedsolvem(const ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_2d_array< amp::ampf<Precision> >& lua, const ap::template_1d_array< int >& p, int n, const ap::template_2d_array< amp::ampf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::ampf<Precision> >& x);

rmatrixsolve function

/************************************************************************* Dense solver. This subroutine solves a system A*x=b, where A is NxN non-denegerate real matrix, x and b are vectors. Algorithm features: * automatic detection of degenerate cases * condition number estimation * iterative refinement * O(N^3) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A B - array[0..N-1], right part OUTPUT PARAMETERS Info - return code: * -3 A is singular, or VERY close to singular. X is filled by zeros in such cases. * -1 N<=0 was passed * 1 task is solved (but matrix A may be ill-conditioned, check R1/RInf parameters for condition numbers). Rep - solver report, see below for more info X - array[0..N-1], it contains: * solution of A*x=b if A is non-singular (well-conditioned or ill-conditioned, but not very close to singular) * zeros, if A is singular or VERY close to singular (in this case Info=-3). SOLVER REPORT Subroutine sets following fields of the Rep structure: * R1 reciprocal of condition number: 1/cond(A), 1-norm. * RInf reciprocal of condition number: 1/cond(A), inf-norm. -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixsolve(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, const ap::template_1d_array< amp::ampf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

rmatrixsolvels function

/************************************************************************* Dense solver. This subroutine finds solution of the linear system A*X=B with non-square, possibly degenerate A. System is solved in the least squares sense, and general least squares solution X = X0 + CX*y which minimizes |A*X-B| is returned. If A is non-degenerate, solution in the usual sense is returned Algorithm features: * automatic detection of degenerate cases * iterative refinement * O(N^3) complexity INPUT PARAMETERS A - array[0..NRows-1,0..NCols-1], system matrix NRows - vertical size of A NCols - horizontal size of A B - array[0..NCols-1], right part Threshold- a number in [0,1]. Singular values beyond Threshold are considered zero. Set it to 0.0, if you don't understand what it means, so the solver will choose good value on its own. OUTPUT PARAMETERS Info - return code: * -4 SVD subroutine failed * -1 if NRows<=0 or NCols<=0 or Threshold<0 was passed * 1 if task is solved Rep - solver report, see below for more info X - array[0..N-1,0..M-1], it contains: * solution of A*X=B if A is non-singular (well-conditioned or ill-conditioned, but not very close to singular) * zeros, if A is singular or VERY close to singular (in this case Info=-3). SOLVER REPORT Subroutine sets following fields of the Rep structure: * R2 reciprocal of condition number: 1/cond(A), 2-norm. * N = NCols * K dim(Null(A)) * CX array[0..N-1,0..K-1], kernel of A. Columns of CX store such vectors that A*CX[i]=0. -- ALGLIB -- Copyright 24.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixsolvels(const ap::template_2d_array< amp::ampf<Precision> >& a, int nrows, int ncols, const ap::template_1d_array< amp::ampf<Precision> >& b, amp::ampf<Precision> threshold, int& info, densesolverlsreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

rmatrixsolvem function

/************************************************************************* Dense solver. Similar to RMatrixSolve() but solves task with multiple right parts (where b and x are NxM matrices). Algorithm features: * automatic detection of degenerate cases * condition number estimation * optional iterative refinement * O(N^3+M*N^2) complexity INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A B - array[0..N-1,0..M-1], right part M - right part size RFS - iterative refinement switch: * True - refinement is used. Less performance, more precision. * False - refinement is not used. More performance, less precision. OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixsolvem(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, const ap::template_2d_array< amp::ampf<Precision> >& b, int m, bool rfs, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::ampf<Precision> >& x);

spdmatrixcholeskysolve function

/************************************************************************* Dense solver. Same as RMatrixLUSolve(), but for SPD matrices represented by their Cholesky decomposition. Algorithm features: * automatic detection of degenerate cases * O(N^2) complexity * condition number estimation * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS CHA - array[0..N-1,0..N-1], Cholesky decomposition, SPDMatrixCholesky result N - size of A IsUpper - what half of CHA is provided B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixcholeskysolve(const ap::template_2d_array< amp::ampf<Precision> >& cha, int n, bool isupper, const ap::template_1d_array< amp::ampf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

spdmatrixcholeskysolvem function

/************************************************************************* Dense solver. Same as RMatrixLUSolveM(), but for SPD matrices represented by their Cholesky decomposition. Algorithm features: * automatic detection of degenerate cases * O(M*N^2) complexity * condition number estimation * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS CHA - array[0..N-1,0..N-1], Cholesky decomposition, SPDMatrixCholesky result N - size of CHA IsUpper - what half of CHA is provided B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixcholeskysolvem(const ap::template_2d_array< amp::ampf<Precision> >& cha, int n, bool isupper, const ap::template_2d_array< amp::ampf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::ampf<Precision> >& x);

spdmatrixsolve function

/************************************************************************* Dense solver. Same as RMatrixSolve(), but for SPD matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * O(N^3) complexity * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A IsUpper - what half of A is provided B - array[0..N-1], right part OUTPUT PARAMETERS Info - same as in RMatrixSolve Returns -3 for non-SPD matrices. Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixsolve(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, const ap::template_1d_array< amp::ampf<Precision> >& b, int& info, densesolverreport<Precision>& rep, ap::template_1d_array< amp::ampf<Precision> >& x);

spdmatrixsolvem function

/************************************************************************* Dense solver. Same as RMatrixSolveM(), but for symmetric positive definite matrices. Algorithm features: * automatic detection of degenerate cases * condition number estimation * O(N^3+M*N^2) complexity * matrix is represented by its upper or lower triangle No iterative refinement is provided because such partial representation of matrix does not allow efficient calculation of extra-precise matrix-vector products for large matrices. Use RMatrixSolve or RMatrixMixedSolve if you need iterative refinement. INPUT PARAMETERS A - array[0..N-1,0..N-1], system matrix N - size of A IsUpper - what half of A is provided B - array[0..N-1,0..M-1], right part M - right part size OUTPUT PARAMETERS Info - same as in RMatrixSolve. Returns -3 for non-SPD matrices. Rep - same as in RMatrixSolve X - same as in RMatrixSolve -- ALGLIB -- Copyright 27.01.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixsolvem(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, const ap::template_2d_array< amp::ampf<Precision> >& b, int m, int& info, densesolverreport<Precision>& rep, ap::template_2d_array< amp::ampf<Precision> >& x);

evd unit

Functions

hmatrixevd
hmatrixevdi
hmatrixevdr
rmatrixevd
smatrixevd
smatrixevdi
smatrixevdr
smatrixtdevd
smatrixtdevdi
smatrixtdevdr

hmatrixevd function

/************************************************************************* Finding the eigenvalues and eigenvectors of a Hermitian matrix The algorithm finds eigen pairs of a Hermitian matrix by reducing it to real tridiagonal form and using the QL/QR algorithm. Input parameters: A - Hermitian matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. Output parameters: D - eigenvalues in ascending order. Array whose index ranges within [0..N-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains the eigenvectors. Array whose indexes range within [0..N-1, 0..N-1]. The eigenvectors are stored in the matrix columns. Result: True, if the algorithm has converged. False, if the algorithm hasn't converged (rare case). Note: eigenvectors of Hermitian matrix are defined up to multiplication by a complex number L, such that |L|=1. -- ALGLIB -- Copyright 2005, 23 March 2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool hmatrixevd(ap::template_2d_array< amp::campf<Precision> > a, int n, int zneeded, bool isupper, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_2d_array< amp::campf<Precision> >& z);

hmatrixevdi function

/************************************************************************* Subroutine for finding the eigenvalues and eigenvectors of a Hermitian matrix with given indexes by using bisection and inverse iteration methods Input parameters: A - Hermitian matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. IsUpperA - storage format of matrix A. I1, I2 - index interval for searching (from I1 to I2). 0 <= I1 <= I2 <= N-1. Output parameters: W - array of the eigenvalues found. Array whose index ranges within [0..I2-I1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains eigenvectors. Array whose indexes range within [0..N-1, 0..I2-I1]. In that case, the eigenvectors are stored in the matrix columns. Result: True, if successful. W contains the eigenvalues, Z contains the eigenvectors (if needed). False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned. Note: eigen vectors of Hermitian matrix are defined up to multiplication by a complex number L, such as |L|=1. -- ALGLIB -- Copyright 07.01.2006, 24.03.2007 by Bochkanov Sergey. *************************************************************************/
template<unsigned int Precision> bool hmatrixevdi(ap::template_2d_array< amp::campf<Precision> > a, int n, int zneeded, bool isupper, int i1, int i2, ap::template_1d_array< amp::ampf<Precision> >& w, ap::template_2d_array< amp::campf<Precision> >& z);

hmatrixevdr function

/************************************************************************* Subroutine for finding the eigenvalues (and eigenvectors) of a Hermitian matrix in a given half-interval (A, B] by using a bisection and inverse iteration Input parameters: A - Hermitian matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. IsUpperA - storage format of matrix A. B1, B2 - half-interval (B1, B2] to search eigenvalues in. Output parameters: M - number of eigenvalues found in a given half-interval, M>=0 W - array of the eigenvalues found. Array whose index ranges within [0..M-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains eigenvectors. Array whose indexes range within [0..N-1, 0..M-1]. The eigenvectors are stored in the matrix columns. Result: True, if successful. M contains the number of eigenvalues in the given half-interval (could be equal to 0), W contains the eigenvalues, Z contains the eigenvectors (if needed). False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned, M is equal to 0. Note: eigen vectors of Hermitian matrix are defined up to multiplication by a complex number L, such as |L|=1. -- ALGLIB -- Copyright 07.01.2006, 24.03.2007 by Bochkanov Sergey. *************************************************************************/
template<unsigned int Precision> bool hmatrixevdr(ap::template_2d_array< amp::campf<Precision> > a, int n, int zneeded, bool isupper, amp::ampf<Precision> b1, amp::ampf<Precision> b2, int& m, ap::template_1d_array< amp::ampf<Precision> >& w, ap::template_2d_array< amp::campf<Precision> >& z);

rmatrixevd function

/************************************************************************* Finding eigenvalues and eigenvectors of a general matrix The algorithm finds eigenvalues and eigenvectors of a general matrix by using the QR algorithm with multiple shifts. The algorithm can find eigenvalues and both left and right eigenvectors. The right eigenvector is a vector x such that A*x = w*x, and the left eigenvector is a vector y such that y'*A = w*y' (here y' implies a complex conjugate transposition of vector y). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. VNeeded - flag controlling whether eigenvectors are needed or not. If VNeeded is equal to: * 0, eigenvectors are not returned; * 1, right eigenvectors are returned; * 2, left eigenvectors are returned; * 3, both left and right eigenvectors are returned. Output parameters: WR - real parts of eigenvalues. Array whose index ranges within [0..N-1]. WR - imaginary parts of eigenvalues. Array whose index ranges within [0..N-1]. VL, VR - arrays of left and right eigenvectors (if they are needed). If WI[i]=0, the respective eigenvalue is a real number, and it corresponds to the column number I of matrices VL/VR. If WI[i]>0, we have a pair of complex conjugate numbers with positive and negative imaginary parts: the first eigenvalue WR[i] + sqrt(-1)*WI[i]; the second eigenvalue WR[i+1] + sqrt(-1)*WI[i+1]; WI[i]>0 WI[i+1] = -WI[i] < 0 In that case, the eigenvector corresponding to the first eigenvalue is located in i and i+1 columns of matrices VL/VR (the column number i contains the real part, and the column number i+1 contains the imaginary part), and the vector corresponding to the second eigenvalue is a complex conjugate to the first vector. Arrays whose indexes range within [0..N-1, 0..N-1]. Result: True, if the algorithm has converged. False, if the algorithm has not converged. Note 1: Some users may ask the following question: what if WI[N-1]>0? WI[N] must contain an eigenvalue which is complex conjugate to the N-th eigenvalue, but the array has only size N? The answer is as follows: such a situation cannot occur because the algorithm finds a pairs of eigenvalues, therefore, if WI[i]>0, I is strictly less than N-1. Note 2: The algorithm performance depends on the value of the internal parameter NS of the InternalSchurDecomposition subroutine which defines the number of shifts in the QR algorithm (similarly to the block width in block-matrix algorithms of linear algebra). If you require maximum performance on your machine, it is recommended to adjust this parameter manually. See also the InternalTREVC subroutine. The algorithm is based on the LAPACK 3.0 library. *************************************************************************/
template<unsigned int Precision> bool rmatrixevd(ap::template_2d_array< amp::ampf<Precision> > a, int n, int vneeded, ap::template_1d_array< amp::ampf<Precision> >& wr, ap::template_1d_array< amp::ampf<Precision> >& wi, ap::template_2d_array< amp::ampf<Precision> >& vl, ap::template_2d_array< amp::ampf<Precision> >& vr);

smatrixevd function

/************************************************************************* Finding the eigenvalues and eigenvectors of a symmetric matrix The algorithm finds eigen pairs of a symmetric matrix by reducing it to tridiagonal form and using the QL/QR algorithm. Input parameters: A - symmetric matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. Output parameters: D - eigenvalues in ascending order. Array whose index ranges within [0..N-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains the eigenvectors. Array whose indexes range within [0..N-1, 0..N-1]. The eigenvectors are stored in the matrix columns. Result: True, if the algorithm has converged. False, if the algorithm hasn't converged (rare case). -- ALGLIB -- Copyright 2005-2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixevd(ap::template_2d_array< amp::ampf<Precision> > a, int n, int zneeded, bool isupper, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixevdi function

/************************************************************************* Subroutine for finding the eigenvalues and eigenvectors of a symmetric matrix with given indexes by using bisection and inverse iteration methods. Input parameters: A - symmetric matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. IsUpperA - storage format of matrix A. I1, I2 - index interval for searching (from I1 to I2). 0 <= I1 <= I2 <= N-1. Output parameters: W - array of the eigenvalues found. Array whose index ranges within [0..I2-I1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains eigenvectors. Array whose indexes range within [0..N-1, 0..I2-I1]. In that case, the eigenvectors are stored in the matrix columns. Result: True, if successful. W contains the eigenvalues, Z contains the eigenvectors (if needed). False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned. -- ALGLIB -- Copyright 07.01.2006 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixevdi(ap::template_2d_array< amp::ampf<Precision> > a, int n, int zneeded, bool isupper, int i1, int i2, ap::template_1d_array< amp::ampf<Precision> >& w, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixevdr function

/************************************************************************* Subroutine for finding the eigenvalues (and eigenvectors) of a symmetric matrix in a given half open interval (A, B] by using a bisection and inverse iteration Input parameters: A - symmetric matrix which is given by its upper or lower triangular part. Array [0..N-1, 0..N-1]. N - size of matrix A. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. IsUpperA - storage format of matrix A. B1, B2 - half open interval (B1, B2] to search eigenvalues in. Output parameters: M - number of eigenvalues found in a given half-interval (M>=0). W - array of the eigenvalues found. Array whose index ranges within [0..M-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains eigenvectors. Array whose indexes range within [0..N-1, 0..M-1]. The eigenvectors are stored in the matrix columns. Result: True, if successful. M contains the number of eigenvalues in the given half-interval (could be equal to 0), W contains the eigenvalues, Z contains the eigenvectors (if needed). False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned, M is equal to 0. -- ALGLIB -- Copyright 07.01.2006 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixevdr(ap::template_2d_array< amp::ampf<Precision> > a, int n, int zneeded, bool isupper, amp::ampf<Precision> b1, amp::ampf<Precision> b2, int& m, ap::template_1d_array< amp::ampf<Precision> >& w, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixtdevd function

/************************************************************************* Finding the eigenvalues and eigenvectors of a tridiagonal symmetric matrix The algorithm finds the eigen pairs of a tridiagonal symmetric matrix by using an QL/QR algorithm with implicit shifts. Input parameters: D - the main diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-1]. E - the secondary diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-2]. N - size of matrix A. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not needed; * 1, the eigenvectors of a tridiagonal matrix are multiplied by the square matrix Z. It is used if the tridiagonal matrix is obtained by the similarity transformation of a symmetric matrix; * 2, the eigenvectors of a tridiagonal matrix replace the square matrix Z; * 3, matrix Z contains the first row of the eigenvectors matrix. Z - if ZNeeded=1, Z contains the square matrix by which the eigenvectors are multiplied. Array whose indexes range within [0..N-1, 0..N-1]. Output parameters: D - eigenvalues in ascending order. Array whose index ranges within [0..N-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains the product of a given matrix (from the left) and the eigenvectors matrix (from the right); * 2, Z contains the eigenvectors. * 3, Z contains the first row of the eigenvectors matrix. If ZNeeded<3, Z is the array whose indexes range within [0..N-1, 0..N-1]. In that case, the eigenvectors are stored in the matrix columns. If ZNeeded=3, Z is the array whose indexes range within [0..0, 0..N-1]. Result: True, if the algorithm has converged. False, if the algorithm hasn't converged. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University September 30, 1994 *************************************************************************/
template<unsigned int Precision> bool smatrixtdevd(ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_1d_array< amp::ampf<Precision> > e, int n, int zneeded, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixtdevdi function

/************************************************************************* Subroutine for finding tridiagonal matrix eigenvalues/vectors with given indexes (in ascending order) by using the bisection and inverse iteraion. Input parameters: D - the main diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-1]. E - the secondary diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-2]. N - size of matrix. N>=0. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not needed; * 1, the eigenvectors of a tridiagonal matrix are multiplied by the square matrix Z. It is used if the tridiagonal matrix is obtained by the similarity transformation of a symmetric matrix. * 2, the eigenvectors of a tridiagonal matrix replace matrix Z. I1, I2 - index interval for searching (from I1 to I2). 0 <= I1 <= I2 <= N-1. Z - if ZNeeded is equal to: * 0, Z isn't used and remains unchanged; * 1, Z contains the square matrix (array whose indexes range within [0..N-1, 0..N-1]) which reduces the given symmetric matrix to tridiagonal form; * 2, Z isn't used (but changed on the exit). Output parameters: D - array of the eigenvalues found. Array whose index ranges within [0..I2-I1]. Z - if ZNeeded is equal to: * 0, doesn't contain any information; * 1, contains the product of a given NxN matrix Z (from the left) and Nx(I2-I1) matrix of the eigenvectors found (from the right). Array whose indexes range within [0..N-1, 0..I2-I1]. * 2, contains the matrix of the eigenvalues found. Array whose indexes range within [0..N-1, 0..I2-I1]. Result: True, if successful. In that case, D contains the eigenvalues, Z contains the eigenvectors (if needed). It should be noted that the subroutine changes the size of arrays D and Z. False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned. -- ALGLIB -- Copyright 25.12.2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixtdevdi(ap::template_1d_array< amp::ampf<Precision> >& d, const ap::template_1d_array< amp::ampf<Precision> >& e, int n, int zneeded, int i1, int i2, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixtdevdr function

/************************************************************************* Subroutine for finding the tridiagonal matrix eigenvalues/vectors in a given half-interval (A, B] by using bisection and inverse iteration. Input parameters: D - the main diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-1]. E - the secondary diagonal of a tridiagonal matrix. Array whose index ranges within [0..N-2]. N - size of matrix, N>=0. ZNeeded - flag controlling whether the eigenvectors are needed or not. If ZNeeded is equal to: * 0, the eigenvectors are not needed; * 1, the eigenvectors of a tridiagonal matrix are multiplied by the square matrix Z. It is used if the tridiagonal matrix is obtained by the similarity transformation of a symmetric matrix. * 2, the eigenvectors of a tridiagonal matrix replace matrix Z. A, B - half-interval (A, B] to search eigenvalues in. Z - if ZNeeded is equal to: * 0, Z isn't used and remains unchanged; * 1, Z contains the square matrix (array whose indexes range within [0..N-1, 0..N-1]) which reduces the given symmetric matrix to tridiagonal form; * 2, Z isn't used (but changed on the exit). Output parameters: D - array of the eigenvalues found. Array whose index ranges within [0..M-1]. M - number of eigenvalues found in the given half-interval (M>=0). Z - if ZNeeded is equal to: * 0, doesn't contain any information; * 1, contains the product of a given NxN matrix Z (from the left) and NxM matrix of the eigenvectors found (from the right). Array whose indexes range within [0..N-1, 0..M-1]. * 2, contains the matrix of the eigenvectors found. Array whose indexes range within [0..N-1, 0..M-1]. Result: True, if successful. In that case, M contains the number of eigenvalues in the given half-interval (could be equal to 0), D contains the eigenvalues, Z contains the eigenvectors (if needed). It should be noted that the subroutine changes the size of arrays D and Z. False, if the bisection method subroutine wasn't able to find the eigenvalues in the given interval or if the inverse iteration subroutine wasn't able to find all the corresponding eigenvectors. In that case, the eigenvalues and eigenvectors are not returned, M is equal to 0. -- ALGLIB -- Copyright 31.03.2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixtdevdr(ap::template_1d_array< amp::ampf<Precision> >& d, const ap::template_1d_array< amp::ampf<Precision> >& e, int n, int zneeded, amp::ampf<Precision> a, amp::ampf<Precision> b, int& m, ap::template_2d_array< amp::ampf<Precision> >& z);

fft unit

Functions

fftc1d
fftc1dinv
fftr1d
fftr1dinternaleven
fftr1dinv
fftr1dinvinternaleven

fftc1d function

/************************************************************************* 1-dimensional complex FFT. Array size N may be arbitrary number (composite or prime). Composite N's are handled with cache-oblivious variation of a Cooley-Tukey algorithm. Small prime-factors are transformed using hard coded codelets (similar to FFTW codelets, but without low-level optimization), large prime-factors are handled with Bluestein's algorithm. Fastests transforms are for smooth N's (prime factors are 2, 3, 5 only), most fast for powers of 2. When N have prime factors larger than these, but orders of magnitude smaller than N, computations will be about 4 times slower than for nearby highly composite N's. When N itself is prime, speed will be 6 times lower. Algorithm has O(N*logN) complexity for any N (composite or prime). INPUT PARAMETERS A - array[0..N-1] - complex function to be transformed N - problem size OUTPUT PARAMETERS A - DFT of a input array, array[0..N-1] A_out[j] = SUM(A_in[k]*exp(-2*pi*sqrt(-1)*j*k/N), k = 0..N-1) -- ALGLIB -- Copyright 29.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftc1d(ap::template_1d_array< amp::campf<Precision> >& a, int n);

fftc1dinv function

/************************************************************************* 1-dimensional complex inverse FFT. Array size N may be arbitrary number (composite or prime). Algorithm has O(N*logN) complexity for any N (composite or prime). See FFTC1D() description for more information about algorithm performance. INPUT PARAMETERS A - array[0..N-1] - complex array to be transformed N - problem size OUTPUT PARAMETERS A - inverse DFT of a input array, array[0..N-1] A_out[j] = SUM(A_in[k]/N*exp(+2*pi*sqrt(-1)*j*k/N), k = 0..N-1) -- ALGLIB -- Copyright 29.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftc1dinv(ap::template_1d_array< amp::campf<Precision> >& a, int n);

fftr1d function

/************************************************************************* 1-dimensional real FFT. Algorithm has O(N*logN) complexity for any N (composite or prime). INPUT PARAMETERS A - array[0..N-1] - real function to be transformed N - problem size OUTPUT PARAMETERS F - DFT of a input array, array[0..N-1] F[j] = SUM(A[k]*exp(-2*pi*sqrt(-1)*j*k/N), k = 0..N-1) NOTE: F[] satisfies symmetry property F[k] = conj(F[N-k]), so just one half of array is usually needed. But for convinience subroutine returns full complex array (with frequencies above N/2), so its result may be used by other FFT-related subroutines. -- ALGLIB -- Copyright 01.06.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftr1d(const ap::template_1d_array< amp::ampf<Precision> >& a, int n, ap::template_1d_array< amp::campf<Precision> >& f);

fftr1dinternaleven function

/************************************************************************* Internal subroutine. Never call it directly! -- ALGLIB -- Copyright 01.06.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftr1dinternaleven(ap::template_1d_array< amp::ampf<Precision> >& a, int n, ap::template_1d_array< amp::ampf<Precision> >& buf, ftbase::ftplan<Precision>& plan);

fftr1dinv function

/************************************************************************* 1-dimensional real inverse FFT. Algorithm has O(N*logN) complexity for any N (composite or prime). INPUT PARAMETERS F - array[0..floor(N/2)] - frequencies from forward real FFT N - problem size OUTPUT PARAMETERS A - inverse DFT of a input array, array[0..N-1] NOTE: F[] should satisfy symmetry property F[k] = conj(F[N-k]), so just one half of frequencies array is needed - elements from 0 to floor(N/2). F[0] is ALWAYS real. If N is even F[floor(N/2)] is real too. If N is odd, then F[floor(N/2)] has no special properties. Relying on properties noted above, FFTR1DInv subroutine uses only elements from 0th to floor(N/2)-th. It ignores imaginary part of F[0], and in case N is even it ignores imaginary part of F[floor(N/2)] too. So you can pass either frequencies array with N elements or reduced array with roughly N/2 elements - subroutine will successfully transform both. -- ALGLIB -- Copyright 01.06.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftr1dinv(const ap::template_1d_array< amp::campf<Precision> >& f, int n, ap::template_1d_array< amp::ampf<Precision> >& a);

fftr1dinvinternaleven function

/************************************************************************* Internal subroutine. Never call it directly! -- ALGLIB -- Copyright 01.06.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fftr1dinvinternaleven(ap::template_1d_array< amp::ampf<Precision> >& a, int n, ap::template_1d_array< amp::ampf<Precision> >& buf, ftbase::ftplan<Precision>& plan);

fht unit

Functions

fhtr1d
fhtr1dinv

fhtr1d function

/************************************************************************* 1-dimensional Fast Hartley Transform. Algorithm has O(N*logN) complexity for any N (composite or prime). INPUT PARAMETERS A - array[0..N-1] - real function to be transformed N - problem size OUTPUT PARAMETERS A - FHT of a input array, array[0..N-1], A_out[k] = sum(A_in[j]*(cos(2*pi*j*k/N)+sin(2*pi*j*k/N)), j=0..N-1) -- ALGLIB -- Copyright 04.06.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fhtr1d(ap::template_1d_array< amp::ampf<Precision> >& a, int n);

fhtr1dinv function

/************************************************************************* 1-dimensional inverse FHT. Algorithm has O(N*logN) complexity for any N (composite or prime). INPUT PARAMETERS A - array[0..N-1] - complex array to be transformed N - problem size OUTPUT PARAMETERS A - inverse FHT of a input array, array[0..N-1] -- ALGLIB -- Copyright 29.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void fhtr1dinv(ap::template_1d_array< amp::ampf<Precision> >& a, int n);

gkq unit

Functions

gkqgenerategaussjacobi
gkqgenerategausslegendre
gkqgeneraterec
gkqlegendrecalc
gkqlegendretbl

gkqgenerategaussjacobi function

/************************************************************************* Returns Gauss and Gauss-Kronrod nodes/weights for Gauss-Jacobi quadrature on [-1,1] with weight function W(x)=Power(1-x,Alpha)*Power(1+x,Beta). INPUT PARAMETERS: N - number of Kronrod nodes, must be odd number, >=3. Alpha - power-law coefficient, Alpha>-1 Beta - power-law coefficient, Beta>-1 OUTPUT PARAMETERS: Info - error code: * -5 no real and positive Gauss-Kronrod formula can be created for such a weight function with a given number of nodes. * -4 an error was detected when calculating weights/nodes. Alpha or Beta are too close to -1 to obtain weights/nodes with high enough accuracy, or, may be, N is too large. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N was passed * +1 OK * +2 OK, but quadrature rule have exterior nodes, x[0]<-1 or x[n-1]>+1 X - array[0..N-1] - array of quadrature nodes, ordered in ascending order. WKronrod - array[0..N-1] - Kronrod weights WGauss - array[0..N-1] - Gauss weights (interleaved with zeros corresponding to extended Kronrod nodes). -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gkqgenerategaussjacobi(int n, amp::ampf<Precision> alpha, amp::ampf<Precision> beta, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& wkronrod, ap::template_1d_array< amp::ampf<Precision> >& wgauss);

gkqgenerategausslegendre function

/************************************************************************* Returns Gauss and Gauss-Kronrod nodes/weights for Gauss-Legendre quadrature with N points. GKQLegendreCalc (calculation) or GKQLegendreTbl (precomputed table) is used depending on machine precision and number of nodes. INPUT PARAMETERS: N - number of Kronrod nodes, must be odd number, >=3. OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. N is too large to obtain weights/nodes with high enough accuracy. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, ordered in ascending order. WKronrod - array[0..N-1] - Kronrod weights WGauss - array[0..N-1] - Gauss weights (interleaved with zeros corresponding to extended Kronrod nodes). -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gkqgenerategausslegendre(int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& wkronrod, ap::template_1d_array< amp::ampf<Precision> >& wgauss);

gkqgeneraterec function

/************************************************************************* Computation of nodes and weights of a Gauss-Kronrod quadrature formula The algorithm generates the N-point Gauss-Kronrod quadrature formula with weight function given by coefficients alpha and beta of a recurrence relation which generates a system of orthogonal polynomials: P-1(x) = 0 P0(x) = 1 Pn+1(x) = (x-alpha(n))*Pn(x) - beta(n)*Pn-1(x) and zero moment Mu0 Mu0 = integral(W(x)dx,a,b) INPUT PARAMETERS: Alpha – alpha coefficients, array[0..floor(3*K/2)]. Beta – beta coefficients, array[0..ceil(3*K/2)]. Beta[0] is not used and may be arbitrary. Beta[I]>0. Mu0 – zeroth moment of the weight function. N – number of nodes of the Gauss-Kronrod quadrature formula, N >= 3, N = 2*K+1. OUTPUT PARAMETERS: Info - error code: * -5 no real and positive Gauss-Kronrod formula can be created for such a weight function with a given number of nodes. * -4 N is too large, task may be ill conditioned - x[i]=x[i+1] found. * -3 internal eigenproblem solver hasn't converged * -2 Beta[i]<=0 * -1 incorrect N was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. WKronrod - array[0..N-1] - Kronrod weights WGauss - array[0..N-1] - Gauss weights (interleaved with zeros corresponding to extended Kronrod nodes). -- ALGLIB -- Copyright 08.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gkqgeneraterec(ap::template_1d_array< amp::ampf<Precision> > alpha, ap::template_1d_array< amp::ampf<Precision> > beta, amp::ampf<Precision> mu0, int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& wkronrod, ap::template_1d_array< amp::ampf<Precision> >& wgauss);

gkqlegendrecalc function

/************************************************************************* Returns Gauss and Gauss-Kronrod nodes for quadrature with N points. Reduction to tridiagonal eigenproblem is used. INPUT PARAMETERS: N - number of Kronrod nodes, must be odd number, >=3. OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. N is too large to obtain weights/nodes with high enough accuracy. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, ordered in ascending order. WKronrod - array[0..N-1] - Kronrod weights WGauss - array[0..N-1] - Gauss weights (interleaved with zeros corresponding to extended Kronrod nodes). -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gkqlegendrecalc(int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& wkronrod, ap::template_1d_array< amp::ampf<Precision> >& wgauss);

gkqlegendretbl function

/************************************************************************* Returns Gauss and Gauss-Kronrod nodes for quadrature with N points using pre-calculated table. Nodes/weights were computed with accuracy up to 1.0E-32 (if MPFR version of ALGLIB is used). In standard double precision accuracy reduces to something about 2.0E-16 (depending on your compiler's handling of long floating point constants). INPUT PARAMETERS: N - number of Kronrod nodes. N can be 15, 21, 31, 41, 51, 61. OUTPUT PARAMETERS: X - array[0..N-1] - array of quadrature nodes, ordered in ascending order. WKronrod - array[0..N-1] - Kronrod weights WGauss - array[0..N-1] - Gauss weights (interleaved with zeros corresponding to extended Kronrod nodes). -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gkqlegendretbl(int n, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& wkronrod, ap::template_1d_array< amp::ampf<Precision> >& wgauss, amp::ampf<Precision>& eps);

gq unit

Functions

gqgenerategausshermite
gqgenerategaussjacobi
gqgenerategausslaguerre
gqgenerategausslegendre
gqgenerategausslobattorec
gqgenerategaussradaurec
gqgeneraterec

gqgenerategausshermite function

/************************************************************************* Returns nodes/weights for Gauss-Hermite quadrature on (-inf,+inf) with weight function W(x)=Exp(-x*x) INPUT PARAMETERS: N - number of nodes, >=1 OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. May be, N is too large. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N/Alpha was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategausshermite(int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgenerategaussjacobi function

/************************************************************************* Returns nodes/weights for Gauss-Jacobi quadrature on [-1,1] with weight function W(x)=Power(1-x,Alpha)*Power(1+x,Beta). INPUT PARAMETERS: N - number of nodes, >=1 Alpha - power-law coefficient, Alpha>-1 Beta - power-law coefficient, Beta>-1 OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. Alpha or Beta are too close to -1 to obtain weights/nodes with high enough accuracy, or, may be, N is too large. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N/Alpha/Beta was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategaussjacobi(int n, amp::ampf<Precision> alpha, amp::ampf<Precision> beta, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgenerategausslaguerre function

/************************************************************************* Returns nodes/weights for Gauss-Laguerre quadrature on [0,+inf) with weight function W(x)=Power(x,Alpha)*Exp(-x) INPUT PARAMETERS: N - number of nodes, >=1 Alpha - power-law coefficient, Alpha>-1 OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. Alpha is too close to -1 to obtain weights/nodes with high enough accuracy or, may be, N is too large. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N/Alpha was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategausslaguerre(int n, amp::ampf<Precision> alpha, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgenerategausslegendre function

/************************************************************************* Returns nodes/weights for Gauss-Legendre quadrature on [-1,1] with N nodes. INPUT PARAMETERS: N - number of nodes, >=1 OUTPUT PARAMETERS: Info - error code: * -4 an error was detected when calculating weights/nodes. N is too large to obtain weights/nodes with high enough accuracy. Try to use multiple precision version. * -3 internal eigenproblem solver hasn't converged * -1 incorrect N was passed * +1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 12.05.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategausslegendre(int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgenerategausslobattorec function

/************************************************************************* Computation of nodes and weights for a Gauss-Lobatto quadrature formula The algorithm generates the N-point Gauss-Lobatto quadrature formula with weight function given by coefficients alpha and beta of a recurrence which generates a system of orthogonal polynomials. P-1(x) = 0 P0(x) = 1 Pn+1(x) = (x-alpha(n))*Pn(x) - beta(n)*Pn-1(x) and zeroth moment Mu0 Mu0 = integral(W(x)dx,a,b) INPUT PARAMETERS: Alpha – array[0..N-2], alpha coefficients Beta – array[0..N-2], beta coefficients. Zero-indexed element is not used, may be arbitrary. Beta[I]>0 Mu0 – zeroth moment of the weighting function. A – left boundary of the integration interval. B – right boundary of the integration interval. N – number of nodes of the quadrature formula, N>=3 (including the left and right boundary nodes). OUTPUT PARAMETERS: Info - error code: * -3 internal eigenproblem solver hasn't converged * -2 Beta[i]<=0 * -1 incorrect N was passed * 1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 2005-2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategausslobattorec(ap::template_1d_array< amp::ampf<Precision> > alpha, ap::template_1d_array< amp::ampf<Precision> > beta, amp::ampf<Precision> mu0, amp::ampf<Precision> a, amp::ampf<Precision> b, int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgenerategaussradaurec function

/************************************************************************* Computation of nodes and weights for a Gauss-Radau quadrature formula The algorithm generates the N-point Gauss-Radau quadrature formula with weight function given by the coefficients alpha and beta of a recurrence which generates a system of orthogonal polynomials. P-1(x) = 0 P0(x) = 1 Pn+1(x) = (x-alpha(n))*Pn(x) - beta(n)*Pn-1(x) and zeroth moment Mu0 Mu0 = integral(W(x)dx,a,b) INPUT PARAMETERS: Alpha – array[0..N-2], alpha coefficients. Beta – array[0..N-1], beta coefficients Zero-indexed element is not used. Beta[I]>0 Mu0 – zeroth moment of the weighting function. A – left boundary of the integration interval. N – number of nodes of the quadrature formula, N>=2 (including the left boundary node). OUTPUT PARAMETERS: Info - error code: * -3 internal eigenproblem solver hasn't converged * -2 Beta[i]<=0 * -1 incorrect N was passed * 1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 2005-2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgenerategaussradaurec(ap::template_1d_array< amp::ampf<Precision> > alpha, ap::template_1d_array< amp::ampf<Precision> > beta, amp::ampf<Precision> mu0, amp::ampf<Precision> a, int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

gqgeneraterec function

/************************************************************************* Computation of nodes and weights for a Gauss quadrature formula The algorithm generates the N-point Gauss quadrature formula with weight function given by coefficients alpha and beta of a recurrence relation which generates a system of orthogonal polynomials: P-1(x) = 0 P0(x) = 1 Pn+1(x) = (x-alpha(n))*Pn(x) - beta(n)*Pn-1(x) and zeroth moment Mu0 Mu0 = integral(W(x)dx,a,b) INPUT PARAMETERS: Alpha – array[0..N-1], alpha coefficients Beta – array[0..N-1], beta coefficients Zero-indexed element is not used and may be arbitrary. Beta[I]>0. Mu0 – zeroth moment of the weight function. N – number of nodes of the quadrature formula, N>=1 OUTPUT PARAMETERS: Info - error code: * -3 internal eigenproblem solver hasn't converged * -2 Beta[i]<=0 * -1 incorrect N was passed * 1 OK X - array[0..N-1] - array of quadrature nodes, in ascending order. W - array[0..N-1] - array of quadrature weights. -- ALGLIB -- Copyright 2005-2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void gqgeneraterec(const ap::template_1d_array< amp::ampf<Precision> >& alpha, const ap::template_1d_array< amp::ampf<Precision> >& beta, amp::ampf<Precision> mu0, int n, int& info, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& w);

idwint unit

Structures

idwinterpolant

Functions

idwbuildmodifiedshepard
idwbuildmodifiedshepardr
idwbuildnoisy
idwcalc

idwinterpolant structure

/************************************************************************* IDW interpolant. *************************************************************************/
template<unsigned int Precision> class idwinterpolant { public: int n; int nx; int d; amp::ampf<Precision> r; int nw; nearestneighbor::kdtree<Precision> tree; int modeltype; ap::template_2d_array< amp::ampf<Precision> > q; ap::template_1d_array< amp::ampf<Precision> > xbuf; ap::template_1d_array< int > tbuf; ap::template_1d_array< amp::ampf<Precision> > rbuf; ap::template_2d_array< amp::ampf<Precision> > xybuf; int debugsolverfailures; amp::ampf<Precision> debugworstrcond; amp::ampf<Precision> debugbestrcond; };

idwbuildmodifiedshepard function

/************************************************************************* IDW interpolant using modified Shepard method for uniform point distributions. INPUT PARAMETERS: XY - X and Y values, array[0..N-1,0..NX]. First NX columns contain X-values, last column contain Y-values. N - number of nodes, N>0. NX - space dimension, NX>=1. D - nodal function type, either: * 0 constant model. Just for demonstration only, worst model ever. * 1 linear model, least squares fitting. Simpe model for datasets too small for quadratic models * 2 quadratic model, least squares fitting. Best model available (if your dataset is large enough). * -1 "fast" linear model, use with caution!!! It is significantly faster than linear/quadratic and better than constant model. But it is less robust (especially in the presence of noise). NQ - number of points used to calculate nodal functions (ignored for constant models). NQ should be LARGER than: * max(1.5*(1+NX),2^NX+1) for linear model, * max(3/4*(NX+2)*(NX+1),2^NX+1) for quadratic model. Values less than this threshold will be silently increased. NW - number of points used to calculate weights and to interpolate. Required: >=2^NX+1, values less than this threshold will be silently increased. Recommended value: about 2*NQ OUTPUT PARAMETERS: Z - IDW interpolant. NOTES: * best results are obtained with quadratic models, worst - with constant models * when N is large, NQ and NW must be significantly smaller than N both to obtain optimal performance and to obtain optimal accuracy. In 2 or 3-dimensional tasks NQ=15 and NW=25 are good values to start with. * NQ and NW may be greater than N. In such cases they will be automatically decreased. * this subroutine is always succeeds (as long as correct parameters are passed). * see 'Multivariate Interpolation of Large Sets of Scattered Data' by Robert J. Renka for more information on this algorithm. * this subroutine assumes that point distribution is uniform at the small scales. If it isn't - for example, points are concentrated along "lines", but "lines" distribution is uniform at the larger scale - then you should use IDWBuildModifiedShepardR() -- ALGLIB PROJECT -- Copyright 02.03.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void idwbuildmodifiedshepard(const ap::template_2d_array< amp::ampf<Precision> >& xy, int n, int nx, int d, int nq, int nw, idwinterpolant<Precision>& z);

idwbuildmodifiedshepardr function

/************************************************************************* IDW interpolant using modified Shepard method for non-uniform datasets. This type of model uses constant nodal functions and interpolates using all nodes which are closer than user-specified radius R. It may be used when points distribution is non-uniform at the small scale, but it is at the distances as large as R. INPUT PARAMETERS: XY - X and Y values, array[0..N-1,0..NX]. First NX columns contain X-values, last column contain Y-values. N - number of nodes, N>0. NX - space dimension, NX>=1. R - radius, R>0 OUTPUT PARAMETERS: Z - IDW interpolant. NOTES: * if there is less than IDWKMin points within R-ball, algorithm selects IDWKMin closest ones, so that continuity properties of interpolant are preserved even far from points. -- ALGLIB PROJECT -- Copyright 11.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void idwbuildmodifiedshepardr(const ap::template_2d_array< amp::ampf<Precision> >& xy, int n, int nx, amp::ampf<Precision> r, idwinterpolant<Precision>& z);

idwbuildnoisy function

/************************************************************************* IDW model for noisy data. This subroutine may be used to handle noisy data, i.e. data with noise in OUTPUT values. It differs from IDWBuildModifiedShepard() in the following aspects: * nodal functions are not constrained to pass through nodes: Qi(xi)<>yi, i.e. we have fitting instead of interpolation. * weights which are used during least squares fitting stage are all equal to 1.0 (independently of distance) * "fast"-linear or constant nodal functions are not supported (either not robust enough or too rigid) This problem require far more complex tuning than interpolation problems. Below you can find some recommendations regarding this problem: * focus on tuning NQ; it controls noise reduction. As for NW, you can just make it equal to 2*NQ. * you can use cross-validation to determine optimal NQ. * optimal NQ is a result of complex tradeoff between noise level (more noise = larger NQ required) and underlying function complexity (given fixed N, larger NQ means smoothing of compex features in the data). For example, NQ=N will reduce noise to the minimum level possible, but you will end up with just constant/linear/quadratic (depending on D) least squares model for the whole dataset. INPUT PARAMETERS: XY - X and Y values, array[0..N-1,0..NX]. First NX columns contain X-values, last column contain Y-values. N - number of nodes, N>0. NX - space dimension, NX>=1. D - nodal function degree, either: * 1 linear model, least squares fitting. Simpe model for datasets too small for quadratic models (or for very noisy problems). * 2 quadratic model, least squares fitting. Best model available (if your dataset is large enough). NQ - number of points used to calculate nodal functions. NQ should be significantly larger than 1.5 times the number of coefficients in a nodal function to overcome effects of noise: * larger than 1.5*(1+NX) for linear model, * larger than 3/4*(NX+2)*(NX+1) for quadratic model. Values less than this threshold will be silently increased. NW - number of points used to calculate weights and to interpolate. Required: >=2^NX+1, values less than this threshold will be silently increased. Recommended value: about 2*NQ or larger OUTPUT PARAMETERS: Z - IDW interpolant. NOTES: * best results are obtained with quadratic models, linear models are not recommended to use unless you are pretty sure that it is what you want * this subroutine is always succeeds (as long as correct parameters are passed). * see 'Multivariate Interpolation of Large Sets of Scattered Data' by Robert J. Renka for more information on this algorithm. -- ALGLIB PROJECT -- Copyright 02.03.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void idwbuildnoisy(const ap::template_2d_array< amp::ampf<Precision> >& xy, int n, int nx, int d, int nq, int nw, idwinterpolant<Precision>& z);

idwcalc function

/************************************************************************* IDW interpolation INPUT PARAMETERS: Z - IDW interpolant built with one of model building subroutines. X - array[0..NX-1], interpolation point Result: IDW interpolant Z(X) -- ALGLIB -- Copyright 02.03.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> idwcalc(idwinterpolant<Precision>& z, const ap::template_1d_array< amp::ampf<Precision> >& x);

inverseupdate unit

Functions

rmatrixinvupdatecolumn
rmatrixinvupdaterow
rmatrixinvupdatesimple
rmatrixinvupdateuv

rmatrixinvupdatecolumn function

/************************************************************************* Inverse matrix update by the Sherman-Morrison formula The algorithm updates matrix A^-1 when adding a vector to a column of matrix A. Input parameters: InvA - inverse of matrix A. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. UpdColumn - the column of A whose vector U was added. 0 <= UpdColumn <= N-1 U - the vector to be added to a column. Array whose index ranges within [0..N-1]. Output parameters: InvA - inverse of modified matrix A. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixinvupdatecolumn(ap::template_2d_array< amp::ampf<Precision> >& inva, int n, int updcolumn, const ap::template_1d_array< amp::ampf<Precision> >& u);

rmatrixinvupdaterow function

/************************************************************************* Inverse matrix update by the Sherman-Morrison formula The algorithm updates matrix A^-1 when adding a vector to a row of matrix A. Input parameters: InvA - inverse of matrix A. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. UpdRow - the row of A whose vector V was added. 0 <= Row <= N-1 V - the vector to be added to a row. Array whose index ranges within [0..N-1]. Output parameters: InvA - inverse of modified matrix A. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixinvupdaterow(ap::template_2d_array< amp::ampf<Precision> >& inva, int n, int updrow, const ap::template_1d_array< amp::ampf<Precision> >& v);

rmatrixinvupdatesimple function

/************************************************************************* Inverse matrix update by the Sherman-Morrison formula The algorithm updates matrix A^-1 when adding a number to an element of matrix A. Input parameters: InvA - inverse of matrix A. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. UpdRow - row where the element to be updated is stored. UpdColumn - column where the element to be updated is stored. UpdVal - a number to be added to the element. Output parameters: InvA - inverse of modified matrix A. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixinvupdatesimple(ap::template_2d_array< amp::ampf<Precision> >& inva, int n, int updrow, int updcolumn, amp::ampf<Precision> updval);

rmatrixinvupdateuv function

/************************************************************************* Inverse matrix update by the Sherman-Morrison formula The algorithm computes the inverse of matrix A+u*v’ by using the given matrix A^-1 and the vectors u and v. Input parameters: InvA - inverse of matrix A. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. U - the vector modifying the matrix. Array whose index ranges within [0..N-1]. V - the vector modifying the matrix. Array whose index ranges within [0..N-1]. Output parameters: InvA - inverse of matrix A + u*v'. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixinvupdateuv(ap::template_2d_array< amp::ampf<Precision> >& inva, int n, const ap::template_1d_array< amp::ampf<Precision> >& u, const ap::template_1d_array< amp::ampf<Precision> >& v);

ldlt unit

Functions

smatrixldlt

smatrixldlt function

/************************************************************************* LDLTDecomposition of a symmetric matrix The algorithm represents a symmetric matrix (which is not necessarily positive definite) as A=L*D*L' or A = U*D*U', where D is a block-diagonal matrix with blocks 1x1 or 2x2, matrix L (matrix U) is a product of lower (upper) triangular matrices with unit diagonal and permutation matrices. Input parameters: A - factorized matrix, array with elements [0..N-1, 0..N-1]. If IsUpper – True, then the upper triangle contains elements of symmetric matrix A, and the lower triangle is not used. The same applies if IsUpper = False. N - size of factorized matrix. IsUpper - parameter which shows a method of matrix definition (lower or upper triangle). Output parameters: A - matrices D and U, if IsUpper = True, or L, if IsUpper = False, in compact form, replacing the upper (lower) triangle of matrix A. In that case, the elements under (over) the main diagonal are not used nor modified. Pivots - tables of performed permutations (see below). If IsUpper = True, then A = U*D*U', U = P(n)*U(n)*...*P(k)*U(k), where P(k) is the permutation matrix, U(k) - upper triangular matrix with its unit main diagonal and k decreases from n with step s which is equal to 1 or 2 (according to the size of the blocks of matrix D). ( I v 0 ) k-s+1 U(k) = ( 0 I 0 ) s ( 0 0 I ) n-k-1 k-s+1 s n-k-1 If Pivots[k]>=0, then s=1, P(k) - permutation of rows k and Pivots[k], the vectorv forming matrix U(k) is stored in elements A(0:k-1,k), D(k) replaces A(k,k). If Pivots[k]=Pivots[k-1]<0 then s=2, P(k) - permutation of rows k-1 and N+Pivots[k-1], the vector v forming matrix U(k) is stored in elements A(0:k-1,k:k+1), the upper triangle of block D(k) is stored in A(k,k), A(k,k+1) and A(k+1,k+1). If IsUpper = False, then A = L*D*L', L=P(0)*L(0)*...*P(k)*L(k), where P(k) is the permutation matrix, L(k) – lower triangular matrix with unit main diagonal and k decreases from 1 with step s which is equal to 1 or 2 (according to the size of the blocks of matrix D). ( I 0 0 ) k-1 L(k) = ( 0 I 0 ) s ( 0 v I ) n-k-s+1 k-1 s n-k-s+1 If Pivots[k]>=0 then s=1, P(k) – permutation of rows k and Pivots[k], the vector v forming matrix L(k) is stored in elements A(k+1:n-1,k), D(k) replaces A(k,k). If Pivots[k]=Pivots[k+1]<0 then s=2, P(k) - permutation of rows k+1 and N+Pivots[k+1], the vector v forming matrix L(k) is stored in elements A(k+2:n-1,k:k+1), the lower triangle of block D(k) is stored in A(k,k), A(k+1,k) and A(k+1,k+1). -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University June 30, 1999 *************************************************************************/
template<unsigned int Precision> void smatrixldlt(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, ap::template_1d_array< int >& pivots);

lsfit unit

Structures

lsfitreport

Functions

lsfitlinear
lsfitlinearc
lsfitlinearw
lsfitlinearwc
lsfitnonlinearfg
lsfitnonlinearfgh
lsfitnonlineariteration
lsfitnonlinearresults
lsfitnonlinearsetcond
lsfitnonlinearsetstpmax
lsfitnonlinearwfg
lsfitnonlinearwfgh

Examples

lsfit_linear
lsfit_nonlinear
lsfit_nonlinear2

lsfitreport structure

/************************************************************************* Least squares fitting report: TaskRCond reciprocal of task's condition number RMSError RMS error AvgError average error AvgRelError average relative error (for non-zero Y[I]) MaxError maximum error *************************************************************************/
template<unsigned int Precision> class lsfitreport { public: amp::ampf<Precision> taskrcond; amp::ampf<Precision> rmserror; amp::ampf<Precision> avgerror; amp::ampf<Precision> avgrelerror; amp::ampf<Precision> maxerror; };

lsfitlinear function

/************************************************************************* Linear least squares fitting, without weights. See LSFitLinearW for more information. -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitlinear(const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_2d_array< amp::ampf<Precision> >& fmatrix, int n, int m, int& info, ap::template_1d_array< amp::ampf<Precision> >& c, lsfitreport<Precision>& rep);

Examples:   lsfit_linear  

lsfitlinearc function

/************************************************************************* Constained linear least squares fitting, without weights. See LSFitLinearWC() for more information. -- ALGLIB -- Copyright 07.09.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitlinearc(ap::template_1d_array< amp::ampf<Precision> > y, const ap::template_2d_array< amp::ampf<Precision> >& fmatrix, const ap::template_2d_array< amp::ampf<Precision> >& cmatrix, int n, int m, int k, int& info, ap::template_1d_array< amp::ampf<Precision> >& c, lsfitreport<Precision>& rep);

Examples:   lsfit_linear  

lsfitlinearw function

/************************************************************************* Weighted linear least squares fitting. QR decomposition is used to reduce task to MxM, then triangular solver or SVD-based solver is used depending on condition number of the system. It allows to maximize speed and retain decent accuracy. INPUT PARAMETERS: Y - array[0..N-1] Function values in N points. W - array[0..N-1] Weights corresponding to function values. Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. FMatrix - a table of basis functions values, array[0..N-1, 0..M-1]. FMatrix[I, J] - value of J-th basis function in I-th point. N - number of points used. N>=1. M - number of basis functions, M>=1. OUTPUT PARAMETERS: Info - error code: * -4 internal SVD decomposition subroutine failed (very rare and for degenerate systems only) * -1 incorrect N/M were specified * 1 task is solved C - decomposition coefficients, array[0..M-1] Rep - fitting report. Following fields are set: * Rep.TaskRCond reciprocal of condition number * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED SEE ALSO LSFitLinear LSFitLinearC LSFitLinearWC -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitlinearw(const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, const ap::template_2d_array< amp::ampf<Precision> >& fmatrix, int n, int m, int& info, ap::template_1d_array< amp::ampf<Precision> >& c, lsfitreport<Precision>& rep);

Examples:   lsfit_linear  

lsfitlinearwc function

/************************************************************************* Weighted constained linear least squares fitting. This is variation of LSFitLinearW(), which searchs for min|A*x=b| given that K additional constaints C*x=bc are satisfied. It reduces original task to modified one: min|B*y-d| WITHOUT constraints, then LSFitLinearW() is called. INPUT PARAMETERS: Y - array[0..N-1] Function values in N points. W - array[0..N-1] Weights corresponding to function values. Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. FMatrix - a table of basis functions values, array[0..N-1, 0..M-1]. FMatrix[I,J] - value of J-th basis function in I-th point. CMatrix - a table of constaints, array[0..K-1,0..M]. I-th row of CMatrix corresponds to I-th linear constraint: CMatrix[I,0]*C[0] + ... + CMatrix[I,M-1]*C[M-1] = CMatrix[I,M] N - number of points used. N>=1. M - number of basis functions, M>=1. K - number of constraints, 0 <= K < M K=0 corresponds to absence of constraints. OUTPUT PARAMETERS: Info - error code: * -4 internal SVD decomposition subroutine failed (very rare and for degenerate systems only) * -3 either too many constraints (M or more), degenerate constraints (some constraints are repetead twice) or inconsistent constraints were specified. * -1 incorrect N/M/K were specified * 1 task is solved C - decomposition coefficients, array[0..M-1] Rep - fitting report. Following fields are set: * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED IMPORTANT: this subroitine doesn't calculate task's condition number for K<>0. SEE ALSO LSFitLinear LSFitLinearC LSFitLinearWC -- ALGLIB -- Copyright 07.09.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitlinearwc(ap::template_1d_array< amp::ampf<Precision> > y, const ap::template_1d_array< amp::ampf<Precision> >& w, const ap::template_2d_array< amp::ampf<Precision> >& fmatrix, ap::template_2d_array< amp::ampf<Precision> > cmatrix, int n, int m, int k, int& info, ap::template_1d_array< amp::ampf<Precision> >& c, lsfitreport<Precision>& rep);

Examples:   lsfit_linear  

lsfitnonlinearfg function

/************************************************************************* Nonlinear least squares fitting, no individual weights. See LSFitNonlinearWFG for more information. -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearfg(const ap::template_2d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& c, int n, int m, int k, bool cheapfg, lsfitstate<Precision>& state);

Examples:   lsfit_nonlinear  

lsfitnonlinearfgh function

/************************************************************************* Nonlinear least squares fitting using gradient/Hessian without individual weights. See LSFitNonlinearWFGH() for more information. -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearfgh(const ap::template_2d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& c, int n, int m, int k, lsfitstate<Precision>& state);

Examples:   lsfit_nonlinear2  

lsfitnonlineariteration function

/************************************************************************* Nonlinear least squares fitting. Algorithm iteration. Called after inialization of the State structure with LSFitNonlinearXXX() subroutine. See HTML docs for examples. INPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls and which is used for reverse communication. Must be initialized with LSFitNonlinearXXX() call first. RESULT 1. If subroutine returned False, iterative algorithm has converged. 2. If subroutine returned True, then if: * if State.NeedF=True, function value F(X,C) is required * if State.NeedFG=True, function value F(X,C) and gradient dF/dC(X,C) are required * if State.NeedFGH=True function value F(X,C), gradient dF/dC(X,C) and Hessian are required One and only one of this fields can be set at time. Function, its gradient and Hessian are calculated at (X,C), where X is stored in State.X[0..M-1] and C is stored in State.C[0..K-1]. Results are stored: * function value - in State.F * gradient - in State.G[0..K-1] * Hessian - in State.H[0..K-1,0..K-1] -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool lsfitnonlineariteration(lsfitstate<Precision>& state);

Examples:   lsfit_nonlinear  lsfit_nonlinear2  

lsfitnonlinearresults function

/************************************************************************* Nonlinear least squares fitting results. Called after LSFitNonlinearIteration() returned False. INPUT PARAMETERS: State - algorithm state (used by LSFitNonlinearIteration). OUTPUT PARAMETERS: Info - completetion code: * -1 incorrect parameters were specified * 1 relative function improvement is no more than EpsF. * 2 relative step is no more than EpsX. * 4 gradient norm is no more than EpsG * 5 MaxIts steps was taken C - array[0..K-1], solution Rep - optimization report. Following fields are set: * Rep.TerminationType completetion code: * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearresults(const lsfitstate<Precision>& state, int& info, ap::template_1d_array< amp::ampf<Precision> >& c, lsfitreport<Precision>& rep);

Examples:   lsfit_nonlinear  lsfit_nonlinear2  

lsfitnonlinearsetcond function

/************************************************************************* Stopping conditions for nonlinear least squares fitting. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with LSFitNonLinearCreate???() EpsF - stopping criterion. Algorithm stops if |F(k+1)-F(k)| <= EpsF*max{|F(k)|, |F(k+1)|, 1} EpsX - stopping criterion. Algorithm stops if |X(k+1)-X(k)| <= EpsX*(1+|X(k)|) MaxIts - stopping criterion. Algorithm stops after MaxIts iterations. MaxIts=0 means no stopping criterion. NOTE Passing EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (according to the scheme used by MINLM unit). -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearsetcond(lsfitstate<Precision>& state, amp::ampf<Precision> epsf, amp::ampf<Precision> epsx, int maxits);

lsfitnonlinearsetstpmax function

/************************************************************************* This function sets maximum step length INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with LSFitNonLinearCreate???() StpMax - maximum step length, >=0. Set StpMax to 0.0, if you don't want to limit step length. Use this subroutine when you optimize target function which contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow. This function allows us to reject steps that are too large (and therefore expose us to the possible overflow) without actually calculating function value at the x+stp*d. NOTE: non-zero StpMax leads to moderate performance degradation because intermediate step of preconditioned L-BFGS optimization is incompatible with limits on step size. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearsetstpmax(lsfitstate<Precision>& state, amp::ampf<Precision> stpmax);

lsfitnonlinearwfg function

/************************************************************************* Weighted nonlinear least squares fitting using gradient and Hessian. Nonlinear task min(F(c)) is solved, where F(c) = (w[0]*(f(x[0],c)-y[0]))^2 + ... + (w[n-1]*(f(x[n-1],c)-y[n-1]))^2, * N is a number of points, * M is a dimension of a space points belong to, * K is a dimension of a space of parameters being fitted, * w is an N-dimensional vector of weight coefficients, * x is a set of N points, each of them is an M-dimensional vector, * c is a K-dimensional vector of parameters being fitted This subroutine uses only f(x[i],c) and its gradient. INPUT PARAMETERS: X - array[0..N-1,0..M-1], points (one row = one point) Y - array[0..N-1], function values. W - weights, array[0..N-1] C - array[0..K-1], initial approximation to the solution, N - number of points, N>1 M - dimension of space K - number of parameters being fitted CheapFG - boolean flag, which is: * True if both function and gradient calculation complexity are less than O(M^2). An improved algorithm can be used which corresponds to FGJ scheme from MINLM unit. * False otherwise. Standard Jacibian-bases Levenberg-Marquardt algo will be used (FJ scheme). OUTPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls of LSFitNonlinearIteration. Used for reverse communication. This structure should be passed to LSFitNonlinearIteration subroutine. See also: LSFitNonlinearIteration LSFitNonlinearResults LSFitNonlinearFG (fitting without weights) LSFitNonlinearWFGH (fitting using Hessian) LSFitNonlinearFGH (fitting using Hessian, without weights) -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearwfg(const ap::template_2d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, const ap::template_1d_array< amp::ampf<Precision> >& c, int n, int m, int k, bool cheapfg, lsfitstate<Precision>& state);

Examples:   lsfit_nonlinear  

lsfitnonlinearwfgh function

/************************************************************************* Weighted nonlinear least squares fitting using gradient/Hessian. Nonlinear task min(F(c)) is solved, where F(c) = (w[0]*(f(x[0],c)-y[0]))^2 + ... + (w[n-1]*(f(x[n-1],c)-y[n-1]))^2, * N is a number of points, * M is a dimension of a space points belong to, * K is a dimension of a space of parameters being fitted, * w is an N-dimensional vector of weight coefficients, * x is a set of N points, each of them is an M-dimensional vector, * c is a K-dimensional vector of parameters being fitted This subroutine uses f(x[i],c), its gradient and its Hessian. See LSFitNonlinearWFG() subroutine for information about function parameters. -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void lsfitnonlinearwfgh(const ap::template_2d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, const ap::template_1d_array< amp::ampf<Precision> >& c, int n, int m, int k, lsfitstate<Precision>& state);

Examples:   lsfit_nonlinear2  

lsfit_linear example

    int m;
    int n;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_2d_array< amp::ampf<Precision> > fmatrix;
    ap::template_2d_array< amp::ampf<Precision> > cmatrix;
    lsfit::lsfitreport<Precision> rep;
    int info;
    ap::template_1d_array< amp::ampf<Precision> > c;
    int i;
    int j;
    amp::ampf<Precision> x;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;


    printf("\n\nFitting tan(x) by third degree polynomial\n\n");
    printf("Fit type             rms.err max.err    p(0)   dp(0)\n");
    
    //
    // Fitting tan(x) at [0, 0.4*pi] by third degree polynomial:
    // a) without constraints
    // b) constrained at x=0: p(0)=0
    // c) constrained at x=0: p'(0)=1
    // c) constrained at x=0: p(0)=0, p'(0)=1
    //
    m = 4;
    n = 100;
    a = 0;
    b = amp::ampf<Precision>("0.4")*amp::pi<Precision>();
    
    //
    // Prepare task matrix
    //
    y.setlength(n);
    fmatrix.setlength(n, m);
    for(i=0; i<=n-1; i++)
    {
        x = a+(b-a)*i/(n-1);
        y(i) = amp::tan<Precision>(x);
        fmatrix(i,0) = amp::ampf<Precision>("1.0");
        for(j=1; j<=m-1; j++)
        {
            fmatrix(i,j) = x*fmatrix(i,j-1);
        }
    }
    
    //
    // Solve unconstrained task
    //
    lsfit::lsfitlinear<Precision>(y, fmatrix, n, m, info, c, rep);
    printf("Unconstrained        %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(c(0)).toDouble()),
        double(amp::ampf<Precision>(c(1)).toDouble()));
    
    //
    // Solve constrained task, p(0)=0
    // Prepare constraints matrix:
    // * first M columns store values of basis functions at X=0
    // * last column stores zero (desired value at X=0)
    //
    cmatrix.setlength(1, m+1);
    cmatrix(0,0) = 1;
    for(i=1; i<=m-1; i++)
    {
        cmatrix(0,i) = 0;
    }
    cmatrix(0,m) = 0;
    lsfit::lsfitlinearc<Precision>(y, fmatrix, cmatrix, n, m, 1, info, c, rep);
    printf("Constrained, p(0)=0  %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(c(0)).toDouble()),
        double(amp::ampf<Precision>(c(1)).toDouble()));
    
    //
    // Solve constrained task, p'(0)=0
    // Prepare constraints matrix:
    // * first M columns store derivatives of basis functions at X=0
    // * last column stores 1.0 (desired derivative at X=0)
    //
    cmatrix.setlength(1, m+1);
    for(i=0; i<=m-1; i++)
    {
        cmatrix(0,i) = 0;
    }
    cmatrix(0,1) = 1;
    cmatrix(0,m) = 1;
    lsfit::lsfitlinearc<Precision>(y, fmatrix, cmatrix, n, m, 1, info, c, rep);
    printf("Constrained, dp(0)=1 %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(c(0)).toDouble()),
        double(amp::ampf<Precision>(c(1)).toDouble()));
    
    //
    // Solve constrained task, p(0)=0, p'(0)=0
    // Prepare constraints matrix:
    // * first M columns store values/derivatives of basis functions at X=0
    // * last column stores desired values/derivative at X=0
    //
    cmatrix.setlength(2, m+1);
    cmatrix(0,0) = 1;
    for(i=1; i<=m-1; i++)
    {
        cmatrix(0,i) = 0;
    }
    cmatrix(0,m) = 0;
    for(i=0; i<=m-1; i++)
    {
        cmatrix(1,i) = 0;
    }
    cmatrix(1,1) = 1;
    cmatrix(1,m) = 1;
    lsfit::lsfitlinearc<Precision>(y, fmatrix, cmatrix, n, m, 2, info, c, rep);
    printf("Constrained, both    %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(c(0)).toDouble()),
        double(amp::ampf<Precision>(c(1)).toDouble()));
    printf("\n\n");

lsfit_nonlinear example

    int m;
    int n;
    int k;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_2d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > c;
    lsfit::lsfitreport<Precision> rep;
    lsfit::lsfitstate<Precision> state;
    int info;
    amp::ampf<Precision> epsf;
    amp::ampf<Precision> epsx;
    int maxits;
    int i;
    int j;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;


    printf("Fitting 0.5(1+cos(x)) on [-pi,+pi] with exp(-alpha*x^2)\n");
    
    //
    // Fitting 0.5(1+cos(x)) on [-pi,+pi] with Gaussian exp(-alpha*x^2):
    // * without Hessian (gradient only)
    // * using alpha=1 as initial value
    // * using 1000 uniformly distributed points to fit to
    //
    // Notes:
    // * N - number of points
    // * M - dimension of space where points reside
    // * K - number of parameters being fitted
    //
    n = 1000;
    m = 1;
    k = 1;
    a = -amp::pi<Precision>();
    b = +amp::pi<Precision>();
    
    //
    // Prepare task matrix
    //
    y.setlength(n);
    x.setlength(n, m);
    c.setlength(k);
    for(i=0; i<=n-1; i++)
    {
        x(i,0) = a+(b-a)*i/(n-1);
        y(i) = amp::ampf<Precision>("0.5")*(1+amp::cos<Precision>(x(i,0)));
    }
    c(0) = amp::ampf<Precision>("1.0");
    epsf = amp::ampf<Precision>("0.0");
    epsx = amp::ampf<Precision>("0.0001");
    maxits = 0;
    
    //
    // Solve
    //
    lsfit::lsfitnonlinearfg<Precision>(x, y, c, n, m, k, true, state);
    lsfit::lsfitnonlinearsetcond<Precision>(state, epsf, epsx, maxits);
    while( lsfit::lsfitnonlineariteration<Precision>(state) )
    {
        if( state.needf )
        {
            
            //
            // F(x) = Exp(-alpha*x^2)
            //
            state.f = amp::exp<Precision>(-state.c(0)*amp::sqr<Precision>(state.x(0)));
        }
        if( state.needfg )
        {
            
            //
            // F(x)      = Exp(-alpha*x^2)
            // dF/dAlpha = (-x^2)*Exp(-alpha*x^2)
            //
            state.f = amp::exp<Precision>(-state.c(0)*amp::sqr<Precision>(state.x(0)));
            state.g(0) = -amp::sqr<Precision>(state.x(0))*state.f;
        }
    }
    lsfit::lsfitnonlinearresults<Precision>(state, info, c, rep);
    printf("alpha:   %0.3lf\n",
        double(amp::ampf<Precision>(c(0)).toDouble()));
    printf("rms.err: %0.3lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()));
    printf("max.err: %0.3lf\n",
        double(amp::ampf<Precision>(rep.maxerror).toDouble()));
    printf("Termination type: %0ld\n",
        long(info));
    printf("\n\n");

lsfit_nonlinear2 example

    int m;
    int n;
    int k;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_2d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > c;
    lsfit::lsfitreport<Precision> rep;
    lsfit::lsfitstate<Precision> state;
    int info;
    amp::ampf<Precision> epsf;
    amp::ampf<Precision> epsx;
    int maxits;
    int i;
    int j;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;


    printf("Fitting 1-x^2 on [-1,+1] with cos(alpha*pi*x)+beta\n");
    
    //
    // Fitting 1-x^2 on [-1,+1] with cos(alpha*pi*x)+beta:
    // * using Hessian
    // * using alpha=1 and beta=0 as initial values
    // * using 1000 uniformly distributed points to fit to
    //
    // Notes:
    // * N - number of points
    // * M - dimension of space where points reside
    // * K - number of parameters being fitted
    //
    n = 1000;
    m = 1;
    k = 2;
    a = -1;
    b = +1;
    
    //
    // Prepare task matrix
    //
    y.setlength(n);
    x.setlength(n, m);
    c.setlength(k);
    for(i=0; i<=n-1; i++)
    {
        x(i,0) = a+(b-a)*i/(n-1);
        y(i) = 1-amp::sqr<Precision>(x(i,0));
    }
    c(0) = amp::ampf<Precision>("1.0");
    c(1) = amp::ampf<Precision>("0.0");
    epsf = amp::ampf<Precision>("0.0");
    epsx = amp::ampf<Precision>("0.0001");
    maxits = 0;
    
    //
    // Solve
    //
    lsfit::lsfitnonlinearfgh<Precision>(x, y, c, n, m, k, state);
    lsfit::lsfitnonlinearsetcond<Precision>(state, epsf, epsx, maxits);
    while( lsfit::lsfitnonlineariteration<Precision>(state) )
    {
        
        //
        // F(x) = Cos(alpha*pi*x)+beta
        //
        state.f = amp::cos<Precision>(state.c(0)*amp::pi<Precision>()*state.x(0))+state.c(1);
        
        //
        // F(x)      = Cos(alpha*pi*x)+beta
        // dF/dAlpha = -pi*x*Sin(alpha*pi*x)
        // dF/dBeta  = 1.0
        //
        if( state.needfg || state.needfgh )
        {
            state.g(0) = -amp::pi<Precision>()*state.x(0)*amp::sin<Precision>(state.c(0)*amp::pi<Precision>()*state.x(0));
            state.g(1) = amp::ampf<Precision>("1.0");
        }
        
        //
        // F(x)            = Cos(alpha*pi*x)+beta
        // d2F/dAlpha2     = -(pi*x)^2*Cos(alpha*pi*x)
        // d2F/dAlphadBeta = 0
        // d2F/dBeta2     =  0
        //
        if( state.needfgh )
        {
            state.h(0,0) = -amp::sqr<Precision>(amp::pi<Precision>()*state.x(0))*amp::cos<Precision>(state.c(0)*amp::pi<Precision>()*state.x(0));
            state.h(0,1) = amp::ampf<Precision>("0.0");
            state.h(1,0) = amp::ampf<Precision>("0.0");
            state.h(1,1) = amp::ampf<Precision>("0.0");
        }
    }
    lsfit::lsfitnonlinearresults<Precision>(state, info, c, rep);
    printf("alpha:   %0.3lf\n",
        double(amp::ampf<Precision>(c(0)).toDouble()));
    printf("beta:    %0.3lf\n",
        double(amp::ampf<Precision>(c(1)).toDouble()));
    printf("rms.err: %0.3lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()));
    printf("max.err: %0.3lf\n",
        double(amp::ampf<Precision>(rep.maxerror).toDouble()));
    printf("Termination type: %0ld\n",
        long(info));
    printf("\n\n");

matdet unit

Functions

cmatrixdet
cmatrixludet
rmatrixdet
rmatrixludet
spdmatrixcholeskydet
spdmatrixdet

cmatrixdet function

/************************************************************************* Calculation of the determinant of a general matrix Input parameters: A - matrix, array[0..N-1, 0..N-1] N - size of matrix A. Result: determinant of matrix A. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::campf<Precision> cmatrixdet(ap::template_2d_array< amp::campf<Precision> > a, int n);

cmatrixludet function

/************************************************************************* Determinant calculation of the matrix given by its LU decomposition. Input parameters: A - LU decomposition of the matrix (output of RMatrixLU subroutine). Pivots - table of permutations which were made during the LU decomposition. Output of RMatrixLU subroutine. N - size of matrix A. Result: matrix determinant. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::campf<Precision> cmatrixludet(const ap::template_2d_array< amp::campf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n);

rmatrixdet function

/************************************************************************* Calculation of the determinant of a general matrix Input parameters: A - matrix, array[0..N-1, 0..N-1] N - size of matrix A. Result: determinant of matrix A. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixdet(ap::template_2d_array< amp::ampf<Precision> > a, int n);

rmatrixludet function

/************************************************************************* Determinant calculation of the matrix given by its LU decomposition. Input parameters: A - LU decomposition of the matrix (output of RMatrixLU subroutine). Pivots - table of permutations which were made during the LU decomposition. Output of RMatrixLU subroutine. N - size of matrix A. Result: matrix determinant. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixludet(const ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n);

spdmatrixcholeskydet function

/************************************************************************* Determinant calculation of the matrix given by the Cholesky decomposition. Input parameters: A - Cholesky decomposition, output of SMatrixCholesky subroutine. N - size of matrix A. As the determinant is equal to the product of squares of diagonal elements, it’s not necessary to specify which triangle - lower or upper - the matrix is stored in. Result: matrix determinant. -- ALGLIB -- Copyright 2005-2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spdmatrixcholeskydet(const ap::template_2d_array< amp::ampf<Precision> >& a, int n);

spdmatrixdet function

/************************************************************************* Determinant calculation of the symmetric positive definite matrix. Input parameters: A - matrix. Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - if IsUpper = True, then the symmetric matrix A is given by its upper triangle, and the lower triangle isn’t used by subroutine. Similarly, if IsUpper = False, then A is given by its lower triangle. Result: determinant of matrix A. If matrix A is not positive definite, then subroutine returns -1. -- ALGLIB -- Copyright 2005-2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spdmatrixdet(ap::template_2d_array< amp::ampf<Precision> > a, int n, bool isupper);

matgen unit

Functions

cmatrixrndcond
cmatrixrndorthogonal
cmatrixrndorthogonalfromtheleft
cmatrixrndorthogonalfromtheright
hmatrixrndcond
hmatrixrndmultiply
hpdmatrixrndcond
rmatrixrndcond
rmatrixrndorthogonal
rmatrixrndorthogonalfromtheleft
rmatrixrndorthogonalfromtheright
smatrixrndcond
smatrixrndmultiply
spdmatrixrndcond

cmatrixrndcond function

/************************************************************************* Generation of random NxN complex matrix with given condition number C and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::campf<Precision> >& a);

cmatrixrndorthogonal function

/************************************************************************* Generation of a random Haar distributed orthogonal complex matrix INPUT PARAMETERS: N - matrix size, N>=1 OUTPUT PARAMETERS: A - orthogonal NxN matrix, array[0..N-1,0..N-1] -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixrndorthogonal(int n, ap::template_2d_array< amp::campf<Precision> >& a);

cmatrixrndorthogonalfromtheleft function

/************************************************************************* Multiplication of MxN complex matrix by MxM random Haar distributed complex orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..M-1, 0..N-1] M, N- matrix size OUTPUT PARAMETERS: A - Q*A, where Q is random MxM orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixrndorthogonalfromtheleft(ap::template_2d_array< amp::campf<Precision> >& a, int m, int n);

cmatrixrndorthogonalfromtheright function

/************************************************************************* Multiplication of MxN complex matrix by NxN random Haar distributed complex orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..M-1, 0..N-1] M, N- matrix size OUTPUT PARAMETERS: A - A*Q, where Q is random NxN orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixrndorthogonalfromtheright(ap::template_2d_array< amp::campf<Precision> >& a, int m, int n);

hmatrixrndcond function

/************************************************************************* Generation of random NxN Hermitian matrix with given condition number and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hmatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::campf<Precision> >& a);

hmatrixrndmultiply function

/************************************************************************* Hermitian multiplication of NxN matrix by random Haar distributed complex orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..N-1, 0..N-1] N - matrix size OUTPUT PARAMETERS: A - Q^H*A*Q, where Q is random NxN orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hmatrixrndmultiply(ap::template_2d_array< amp::campf<Precision> >& a, int n);

hpdmatrixrndcond function

/************************************************************************* Generation of random NxN Hermitian positive definite matrix with given condition number and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random HPD matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::campf<Precision> >& a);

rmatrixrndcond function

/************************************************************************* Generation of random NxN matrix with given condition number and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::ampf<Precision> >& a);

rmatrixrndorthogonal function

/************************************************************************* Generation of a random uniformly distributed (Haar) orthogonal matrix INPUT PARAMETERS: N - matrix size, N>=1 OUTPUT PARAMETERS: A - orthogonal NxN matrix, array[0..N-1,0..N-1] -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixrndorthogonal(int n, ap::template_2d_array< amp::ampf<Precision> >& a);

rmatrixrndorthogonalfromtheleft function

/************************************************************************* Multiplication of MxN matrix by MxM random Haar distributed orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..M-1, 0..N-1] M, N- matrix size OUTPUT PARAMETERS: A - Q*A, where Q is random MxM orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixrndorthogonalfromtheleft(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n);

rmatrixrndorthogonalfromtheright function

/************************************************************************* Multiplication of MxN matrix by NxN random Haar distributed orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..M-1, 0..N-1] M, N- matrix size OUTPUT PARAMETERS: A - A*Q, where Q is random NxN orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixrndorthogonalfromtheright(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n);

smatrixrndcond function

/************************************************************************* Generation of random NxN symmetric matrix with given condition number and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void smatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::ampf<Precision> >& a);

smatrixrndmultiply function

/************************************************************************* Symmetric multiplication of NxN matrix by random Haar distributed orthogonal matrix INPUT PARAMETERS: A - matrix, array[0..N-1, 0..N-1] N - matrix size OUTPUT PARAMETERS: A - Q'*A*Q, where Q is random NxN orthogonal matrix -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void smatrixrndmultiply(ap::template_2d_array< amp::ampf<Precision> >& a, int n);

spdmatrixrndcond function

/************************************************************************* Generation of random NxN symmetric positive definite matrix with given condition number and norm2(A)=1 INPUT PARAMETERS: N - matrix size C - condition number (in 2-norm) OUTPUT PARAMETERS: A - random SPD matrix with norm2(A)=1 and cond(A)=C -- ALGLIB routine -- 04.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixrndcond(int n, amp::ampf<Precision> c, ap::template_2d_array< amp::ampf<Precision> >& a);

matinv unit

Functions

cmatrixinverse
cmatrixluinverse
cmatrixtrinverse
hpdmatrixcholeskyinverse
hpdmatrixinverse
rmatrixinverse
rmatrixluinverse
rmatrixtrinverse
spdmatrixcholeskyinverse
spdmatrixinverse

cmatrixinverse function

/************************************************************************* Inversion of a general matrix. Input parameters: A - matrix, array[0..N-1,0..N-1]. N - size of A. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixinverse(ap::template_2d_array< amp::campf<Precision> >& a, int n, int& info, matinvreport<Precision>& rep);

cmatrixluinverse function

/************************************************************************* Inversion of a matrix given by its LU decomposition. INPUT PARAMETERS: A - LU decomposition of the matrix (output of CMatrixLU subroutine). Pivots - table of permutations which were made during the LU decomposition (the output of CMatrixLU subroutine). N - size of matrix A. OUTPUT PARAMETERS: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB routine -- 05.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixluinverse(ap::template_2d_array< amp::campf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n, int& info, matinvreport<Precision>& rep);

cmatrixtrinverse function

/************************************************************************* Triangular matrix inverse (complex) The subroutine inverts the following types of matrices: * upper triangular * upper triangular with unit diagonal * lower triangular * lower triangular with unit diagonal In case of an upper (lower) triangular matrix, the inverse matrix will also be upper (lower) triangular, and after the end of the algorithm, the inverse matrix replaces the source matrix. The elements below (above) the main diagonal are not changed by the algorithm. If the matrix has a unit diagonal, the inverse matrix also has a unit diagonal, and the diagonal elements are not passed to the algorithm. Input parameters: A - matrix, array[0..N-1, 0..N-1]. N - size of A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Output parameters: Info - same as for RMatrixLUInverse Rep - same as for RMatrixLUInverse A - same as for RMatrixLUInverse. -- ALGLIB -- Copyright 05.02.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixtrinverse(ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, bool isunit, int& info, matinvreport<Precision>& rep);

hpdmatrixcholeskyinverse function

/************************************************************************* Inversion of a Hermitian positive definite matrix which is given by Cholesky decomposition. Input parameters: A - Cholesky decomposition of the matrix to be inverted: A=U’*U or A = L*L'. Output of HPDMatrixCholesky subroutine. N - size of matrix A. IsUpper – storage format. If IsUpper = True, then matrix A is given as A = U'*U (matrix contains upper triangle). Similarly, if IsUpper = False, then A = L*L'. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB routine -- 10.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixcholeskyinverse(ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, int& info, matinvreport<Precision>& rep);

hpdmatrixinverse function

/************************************************************************* Inversion of a Hermitian positive definite matrix. Given an upper or lower triangle of a Hermitian positive definite matrix, the algorithm generates matrix A^-1 and saves the upper or lower triangle depending on the input. Input parameters: A - matrix to be inverted (upper or lower triangle). Array with elements [0..N-1,0..N-1]. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then the upper triangle of matrix A is given, otherwise the lower triangle is given. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB routine -- 10.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hpdmatrixinverse(ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, int& info, matinvreport<Precision>& rep);

rmatrixinverse function

/************************************************************************* Inversion of a general matrix. Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse Result: True, if the matrix is not singular. False, if the matrix is singular. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixinverse(ap::template_2d_array< amp::ampf<Precision> >& a, int n, int& info, matinvreport<Precision>& rep);

rmatrixluinverse function

/************************************************************************* Inversion of a matrix given by its LU decomposition. INPUT PARAMETERS: A - LU decomposition of the matrix (output of RMatrixLU subroutine). Pivots - table of permutations which were made during the LU decomposition (the output of RMatrixLU subroutine). N - size of matrix A. OUTPUT PARAMETERS: Info - return code: * -3 A is singular, or VERY close to singular. it is filled by zeros in such cases. * -1 N<=0 was passed, or incorrect Pivots was passed * 1 task is solved (but matrix A may be ill-conditioned, check R1/RInf parameters for condition numbers). Rep - solver report, see below for more info A - inverse of matrix A. Array whose indexes range within [0..N-1, 0..N-1]. SOLVER REPORT Subroutine sets following fields of the Rep structure: * R1 reciprocal of condition number: 1/cond(A), 1-norm. * RInf reciprocal of condition number: 1/cond(A), inf-norm. -- ALGLIB routine -- 05.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixluinverse(ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n, int& info, matinvreport<Precision>& rep);

rmatrixtrinverse function

/************************************************************************* Triangular matrix inverse (real) The subroutine inverts the following types of matrices: * upper triangular * upper triangular with unit diagonal * lower triangular * lower triangular with unit diagonal In case of an upper (lower) triangular matrix, the inverse matrix will also be upper (lower) triangular, and after the end of the algorithm, the inverse matrix replaces the source matrix. The elements below (above) the main diagonal are not changed by the algorithm. If the matrix has a unit diagonal, the inverse matrix also has a unit diagonal, and the diagonal elements are not passed to the algorithm. Input parameters: A - matrix, array[0..N-1, 0..N-1]. N - size of A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Output parameters: Info - same as for RMatrixLUInverse Rep - same as for RMatrixLUInverse A - same as for RMatrixLUInverse. -- ALGLIB -- Copyright 05.02.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixtrinverse(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, bool isunit, int& info, matinvreport<Precision>& rep);

spdmatrixcholeskyinverse function

/************************************************************************* Inversion of a symmetric positive definite matrix which is given by Cholesky decomposition. Input parameters: A - Cholesky decomposition of the matrix to be inverted: A=U’*U or A = L*L'. Output of SPDMatrixCholesky subroutine. N - size of matrix A. IsUpper – storage format. If IsUpper = True, then matrix A is given as A = U'*U (matrix contains upper triangle). Similarly, if IsUpper = False, then A = L*L'. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB routine -- 10.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixcholeskyinverse(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, int& info, matinvreport<Precision>& rep);

spdmatrixinverse function

/************************************************************************* Inversion of a symmetric positive definite matrix. Given an upper or lower triangle of a symmetric positive definite matrix, the algorithm generates matrix A^-1 and saves the upper or lower triangle depending on the input. Input parameters: A - matrix to be inverted (upper or lower triangle). Array with elements [0..N-1,0..N-1]. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then the upper triangle of matrix A is given, otherwise the lower triangle is given. Output parameters: Info - return code, same as in RMatrixLUInverse Rep - solver report, same as in RMatrixLUInverse A - inverse of matrix A, same as in RMatrixLUInverse -- ALGLIB routine -- 10.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spdmatrixinverse(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, int& info, matinvreport<Precision>& rep);

minasa unit

Functions

minasacreate
minasaiteration
minasaresults
minasasetalgorithm
minasasetcond
minasasetstpmax
minasasetxrep

Examples

minasa_1
minasa_2

minasacreate function

/************************************************************************* NONLINEAR BOUND CONSTRAINED OPTIMIZATION USING MODIFIED WILLIAM W. HAGER AND HONGCHAO ZHANG ACTIVE SET ALGORITHM The subroutine minimizes function F(x) of N arguments with bound constraints: BndL[i] <= x[i] <= BndU[i] This method is globally convergent as long as grad(f) is Lipschitz continuous on a level set: L = { x : f(x)<=f(x0) }. INPUT PARAMETERS: N - problem dimension. N>0 X - initial solution approximation, array[0..N-1]. BndL - lower bounds, array[0..N-1]. all elements MUST be specified, i.e. all variables are bounded. However, if some (all) variables are unbounded, you may specify very small number as bound: -1000, -1.0E6 or -1.0E300, or something like that. BndU - upper bounds, array[0..N-1]. all elements MUST be specified, i.e. all variables are bounded. However, if some (all) variables are unbounded, you may specify very large number as bound: +1000, +1.0E6 or +1.0E300, or something like that. EpsG - positive number which defines a precision of search. The subroutine finishes its work if the condition ||G|| < EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient, X - current approximation. EpsF - positive number which defines a precision of search. The subroutine finishes its work if on iteration number k+1 the condition |F(k+1)-F(k)| <= EpsF*max{|F(k)|, |F(k+1)|, 1} is satisfied. EpsX - positive number which defines a precision of search. The subroutine finishes its work if on iteration number k+1 the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. OUTPUT PARAMETERS: State - structure used for reverse communication. This function initializes State structure with default optimization parameters (stopping conditions, step size, etc.). Use MinASASet??????() functions to tune optimization parameters. After all optimization parameters are tuned, you should use MinASAIteration() function to advance algorithm iterations. NOTES: 1. you may tune stopping conditions with MinASASetCond() function 2. if target function contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow, use MinASASetStpMax() function to bound algorithm's steps. -- ALGLIB -- Copyright 25.03.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasacreate(int n, const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& bndl, const ap::template_1d_array< amp::ampf<Precision> >& bndu, minasastate<Precision>& state);

Examples:   minasa_1  minasa_2  

minasaiteration function

/************************************************************************* One ASA iteration Called after initialization with MinASACreate. See HTML documentation for examples. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinASACreate. RESULT: * if function returned False, iterative proces has converged. Use MinLBFGSResults() to obtain optimization results. * if subroutine returned True, then, depending on structure fields, we have one of the following situations === FUNC/GRAD REQUEST === State.NeedFG is True => function value/gradient are needed. Caller should calculate function value State.F and gradient State.G[0..N-1] at State.X[0..N-1] and call MinLBFGSIteration() again. === NEW INTERATION IS REPORTED === State.XUpdated is True => one more iteration was made. State.X contains current position, State.F contains function value at X. You can read info from these fields, but never modify them because they contain the only copy of optimization algorithm state. One and only one of these fields (NeedFG, XUpdated) is true on return. New iterations are reported only when reports are explicitly turned on by MinLBFGSSetXRep() function, so if you never called it, you can expect that NeedFG is always True. -- ALGLIB -- Copyright 20.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool minasaiteration(minasastate<Precision>& state);

Examples:   minasa_1  minasa_2  

minasaresults function

/************************************************************************* Conjugate gradient results Called after MinASA returned False. INPUT PARAMETERS: State - algorithm state (used by MinASAIteration). OUTPUT PARAMETERS: X - array[0..N-1], solution Rep - optimization report: * Rep.TerminationType completetion code: * -2 rounding errors prevent further improvement. X contains best point found. * -1 incorrect parameters were specified * 1 relative function improvement is no more than EpsF. * 2 relative step is no more than EpsX. * 4 gradient norm is no more than EpsG * 5 MaxIts steps was taken * 7 stopping conditions are too stringent, further improvement is impossible * Rep.IterationsCount contains iterations count * NFEV countains number of function calculations * ActiveConstraints contains number of active constraints -- ALGLIB -- Copyright 20.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasaresults(const minasastate<Precision>& state, ap::template_1d_array< amp::ampf<Precision> >& x, minasareport<Precision>& rep);

Examples:   minasa_1  minasa_2  

minasasetalgorithm function

/************************************************************************* This function sets optimization algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinASACreate() UAType - algorithm type: * -1 automatic selection of the best algorithm * 0 DY (Dai and Yuan) algorithm * 1 Hybrid DY-HS algorithm -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasasetalgorithm(minasastate<Precision>& state, int algotype);

minasasetcond function

/************************************************************************* This function sets stopping conditions for the ASA optimization algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinASACreate() EpsG - >=0 The subroutine finishes its work if the condition ||G||<EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient. EpsF - >=0 The subroutine finishes its work if on k+1-th iteration the condition |F(k+1)-F(k)|<=EpsF*max{|F(k)|,|F(k+1)|,1} is satisfied. EpsX - >=0 The subroutine finishes its work if on k+1-th iteration the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. Passing EpsG=0, EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (small EpsX). -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasasetcond(minasastate<Precision>& state, amp::ampf<Precision> epsg, amp::ampf<Precision> epsf, amp::ampf<Precision> epsx, int maxits);

minasasetstpmax function

/************************************************************************* This function sets maximum step length INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate() StpMax - maximum step length, >=0. Set StpMax to 0.0, if you don't want to limit step length. Use this subroutine when you optimize target function which contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow. This function allows us to reject steps that are too large (and therefore expose us to the possible overflow) without actually calculating function value at the x+stp*d. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasasetstpmax(minasastate<Precision>& state, amp::ampf<Precision> stpmax);

minasasetxrep function

/************************************************************************* This function turns on/off reporting. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinASACreate() NeedXRep- whether iteration reports are needed or not Usually algorithm returns from MinASAIteration() only when it needs function/gradient. However, with this function we can let it stop after each iteration (one iteration may include more than one function evaluation), which is indicated by XUpdated field. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minasasetxrep(minasastate<Precision>& state, bool needxrep);

minasa_1 example

    int n;
    int i;
    minasa::minasastate<Precision> state;
    minasa::minasareport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    ap::template_1d_array< amp::ampf<Precision> > bndl;
    ap::template_1d_array< amp::ampf<Precision> > bndu;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;
    amp::ampf<Precision> z;


    
    //
    // Function being minimized:
    //     F = x+2y+3z subject to 0<=x<=1, 0<=y<=1, 0<=z<=1.
    //
    n = 3;
    s.setlength(n);
    bndl.setlength(n);
    bndu.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        s(i) = 1;
        bndl(i) = 0;
        bndu(i) = 1;
    }
    minasa::minasacreate<Precision>(n, s, bndl, bndu, state);
    minasa::minasasetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.00001"), 0);
    minasa::minasasetxrep<Precision>(state, true);
    printf("\n\nF = x+2y+3z subject to 0<=x<=1, 0<=y<=1, 0<=z<=1\n");
    printf("OPTIMIZATION STARTED\n");
    while( minasa::minasaiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            z = state.x(2);
            state.f = x+2*y+3*z;
            state.g(0) = 1;
            state.g(1) = 2;
            state.g(2) = 3;
        }
        if( state.xupdated )
        {
            printf("    F(%4.2lf,%4.2lf,%4.2lf)=%0.3lf\n",
                double(amp::ampf<Precision>(state.x(0)).toDouble()),
                double(amp::ampf<Precision>(state.x(1)).toDouble()),
                double(amp::ampf<Precision>(state.x(2)).toDouble()),
                double(amp::ampf<Precision>(state.f).toDouble()));
        }
    }
    printf("OPTIMIZATION STOPPED\n");
    minasa::minasaresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (should be 0.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 0.00)\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("Z = %4.2lf (should be 0.00)\n\n\n",
        double(amp::ampf<Precision>(s(2)).toDouble()));

minasa_2 example

    int n;
    int i;
    minasa::minasastate<Precision> state;
    minasa::minasareport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    ap::template_1d_array< amp::ampf<Precision> > bndl;
    ap::template_1d_array< amp::ampf<Precision> > bndu;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;
    amp::ampf<Precision> z;


    
    //
    // Function being minimized:
    //     F = x+4y+9z subject to 0<=x<=1, 0<=y<=1, 0<=z<=1.
    //
    // Take a look at MinASASetStpMax() - it restricts step length by
    // a small value, so we can see the current point traveling through
    // a feasible set, sticking to its bounds.
    //
    n = 3;
    s.setlength(n);
    bndl.setlength(n);
    bndu.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        s(i) = 1;
        bndl(i) = 0;
        bndu(i) = 1;
    }
    minasa::minasacreate<Precision>(n, s, bndl, bndu, state);
    minasa::minasasetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.00001"), 0);
    minasa::minasasetxrep<Precision>(state, true);
    minasa::minasasetstpmax<Precision>(state, amp::ampf<Precision>("0.2"));
    printf("\n\nF = x+4y+9z subject to 0<=x<=1, 0<=y<=1, 0<=z<=1\n");
    printf("OPTIMIZATION STARTED\n");
    while( minasa::minasaiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            z = state.x(2);
            state.f = x+4*y+9*z;
            state.g(0) = 1;
            state.g(1) = 4;
            state.g(2) = 9;
        }
        if( state.xupdated )
        {
            printf("    F(%4.2lf, %4.2lf, %4.2lf) = %0.3lf\n",
                double(amp::ampf<Precision>(state.x(0)).toDouble()),
                double(amp::ampf<Precision>(state.x(1)).toDouble()),
                double(amp::ampf<Precision>(state.x(2)).toDouble()),
                double(amp::ampf<Precision>(state.f).toDouble()));
        }
    }
    printf("OPTIMIZATION STOPPED\n");
    minasa::minasaresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (should be 0.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 0.00)\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("Z = %4.2lf (should be 0.00)\n\n\n",
        double(amp::ampf<Precision>(s(2)).toDouble()));

mincg unit

Functions

mincgcreate
mincgiteration
mincgresults
mincgsetcgtype
mincgsetcond
mincgsetstpmax
mincgsetxrep

Examples

mincg_1
mincg_2

mincgcreate function

/************************************************************************* NONLINEAR CONJUGATE GRADIENT METHOD The subroutine minimizes function F(x) of N arguments by using one of the nonlinear conjugate gradient methods. These CG methods are globally convergent (even on non-convex functions) as long as grad(f) is Lipschitz continuous in a some neighborhood of the L = { x : f(x)<=f(x0) }. INPUT PARAMETERS: N - problem dimension. N>0 X - initial solution approximation, array[0..N-1]. EpsG - positive number which defines a precision of search. The subroutine finishes its work if the condition ||G|| < EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient, X - current approximation. EpsF - positive number which defines a precision of search. The subroutine finishes its work if on iteration number k+1 the condition |F(k+1)-F(k)| <= EpsF*max{|F(k)|, |F(k+1)|, 1} is satisfied. EpsX - positive number which defines a precision of search. The subroutine finishes its work if on iteration number k+1 the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. OUTPUT PARAMETERS: State - structure used for reverse communication. See also MinCGIteration, MinCGResults NOTE: Passing EpsG=0, EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (small EpsX). -- ALGLIB -- Copyright 25.03.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgcreate(int n, const ap::template_1d_array< amp::ampf<Precision> >& x, mincgstate<Precision>& state);

Examples:   mincg_2  mincg_1  

mincgiteration function

/************************************************************************* One conjugate gradient iteration Called after initialization with MinCG. See HTML documentation for examples. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCG. RESULT: * if function returned False, iterative proces has converged. Use MinLBFGSResults() to obtain optimization results. * if subroutine returned True, then, depending on structure fields, we have one of the following situations === FUNC/GRAD REQUEST === State.NeedFG is True => function value/gradient are needed. Caller should calculate function value State.F and gradient State.G[0..N-1] at State.X[0..N-1] and call MinLBFGSIteration() again. === NEW INTERATION IS REPORTED === State.XUpdated is True => one more iteration was made. State.X contains current position, State.F contains function value at X. You can read info from these fields, but never modify them because they contain the only copy of optimization algorithm state. One and only one of these fields (NeedFG, XUpdated) is true on return. New iterations are reported only when reports are explicitly turned on by MinLBFGSSetXRep() function, so if you never called it, you can expect that NeedFG is always True. -- ALGLIB -- Copyright 20.04.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool mincgiteration(mincgstate<Precision>& state);

Examples:   mincg_2  mincg_1  

mincgresults function

/************************************************************************* Conjugate gradient results Called after MinCG returned False. INPUT PARAMETERS: State - algorithm state (used by MinCGIteration). OUTPUT PARAMETERS: X - array[0..N-1], solution Rep - optimization report: * Rep.TerminationType completetion code: * -2 rounding errors prevent further improvement. X contains best point found. * -1 incorrect parameters were specified * 1 relative function improvement is no more than EpsF. * 2 relative step is no more than EpsX. * 4 gradient norm is no more than EpsG * 5 MaxIts steps was taken * 7 stopping conditions are too stringent, further improvement is impossible * Rep.IterationsCount contains iterations count * NFEV countains number of function calculations -- ALGLIB -- Copyright 20.04.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgresults(const mincgstate<Precision>& state, ap::template_1d_array< amp::ampf<Precision> >& x, mincgreport<Precision>& rep);

Examples:   mincg_2  mincg_1  

mincgsetcgtype function

/************************************************************************* This function sets CG algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate() CGType - algorithm type: * -1 automatic selection of the best algorithm * 0 DY (Dai and Yuan) algorithm * 1 Hybrid DY-HS algorithm -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgsetcgtype(mincgstate<Precision>& state, int cgtype);

mincgsetcond function

/************************************************************************* This function sets stopping conditions for CG optimization algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate() EpsG - >=0 The subroutine finishes its work if the condition ||G||<EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient. EpsF - >=0 The subroutine finishes its work if on k+1-th iteration the condition |F(k+1)-F(k)|<=EpsF*max{|F(k)|,|F(k+1)|,1} is satisfied. EpsX - >=0 The subroutine finishes its work if on k+1-th iteration the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. Passing EpsG=0, EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (small EpsX). -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgsetcond(mincgstate<Precision>& state, amp::ampf<Precision> epsg, amp::ampf<Precision> epsf, amp::ampf<Precision> epsx, int maxits);

mincgsetstpmax function

/************************************************************************* This function sets maximum step length INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate() StpMax - maximum step length, >=0. Set StpMax to 0.0, if you don't want to limit step length. Use this subroutine when you optimize target function which contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow. This function allows us to reject steps that are too large (and therefore expose us to the possible overflow) without actually calculating function value at the x+stp*d. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgsetstpmax(mincgstate<Precision>& state, amp::ampf<Precision> stpmax);

mincgsetxrep function

/************************************************************************* This function turns on/off reporting. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate() NeedXRep- whether iteration reports are needed or not Usually algorithm returns from MinCGIteration() only when it needs function/gradient. However, with this function we can let it stop after each iteration (one iteration may include more than one function evaluation), which is indicated by XUpdated field. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void mincgsetxrep(mincgstate<Precision>& state, bool needxrep);

mincg_1 example

    int n;
    mincg::mincgstate<Precision> state;
    mincg::mincgreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Function minimized:
    //     F = (x-1)^4 + (y-x)^2
    // N = 2 - task dimension.
    //
    n = 2;
    s.setlength(2);
    s(0) = 10;
    s(1) = 11;
    mincg::mincgcreate<Precision>(n, s, state);
    mincg::mincgsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.00001"), 0);
    mincg::mincgsetxrep<Precision>(state, true);
    printf("\n\nF = (x-1)^4 + (y-x)^2\n");
    printf("OPTIMIZATION STARTED\n");
    while( mincg::mincgiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            state.f = amp::sqr<Precision>(amp::sqr<Precision>(x-1))+amp::sqr<Precision>(y-x);
            state.g(0) = 4*amp::sqr<Precision>(x-1)*(x-1)+2*(x-y);
            state.g(1) = 2*(y-x);
        }
        if( state.xupdated )
        {
            printf("    F(%8.5lf,%8.5lf)=%0.5lf\n",
                double(amp::ampf<Precision>(state.x(0)).toDouble()),
                double(amp::ampf<Precision>(state.x(1)).toDouble()),
                double(amp::ampf<Precision>(state.f).toDouble()));
        }
    }
    printf("OPTIMIZATION STOPPED\n");
    mincg::mincgresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (should be 1.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 1.00)\n\n\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));

mincg_2 example

    int n;
    mincg::mincgstate<Precision> state;
    mincg::mincgreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Function minimized:
    //     F = exp(x-1) + exp(1-x) + (y-x)^2
    // N = 2 - task dimension.
    //
    // Take a look at MinCGSetStpMax() call - it prevents us
    // from overflow (which may be result of too large step).
    // Try to comment it and see what will happen.
    //
    n = 2;
    s.setlength(2);
    s(0) = 10;
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    mincg::mincgcreate<Precision>(n, s, state);
    mincg::mincgsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0001"), 0);
    mincg::mincgsetxrep<Precision>(state, true);
    mincg::mincgsetstpmax<Precision>(state, amp::ampf<Precision>("1.0"));
    printf("\n\nF = exp(x-1) + exp(1-x) + (y-x)^2\n");
    printf("OPTIMIZATION STARTED\n");
    while( mincg::mincgiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            state.f = amp::exp<Precision>(x-1)+amp::exp<Precision>(1-x)+amp::sqr<Precision>(y-x);
            state.g(0) = amp::exp<Precision>(x-1)-amp::exp<Precision>(1-x)+2*(x-y);
            state.g(1) = 2*(y-x);
        }
        if( state.xupdated )
        {
            printf("    F(%8.5lf,%8.5lf)=%0.5lf\n",
                double(amp::ampf<Precision>(state.x(0)).toDouble()),
                double(amp::ampf<Precision>(state.x(1)).toDouble()),
                double(amp::ampf<Precision>(state.f).toDouble()));
        }
    }
    printf("OPTIMIZATION STOPPED\n");
    mincg::mincgresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (should be 1.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 1.00)\n\n\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));

minlbfgs unit

Functions

minlbfgscreate
minlbfgscreatex
minlbfgsiteration
minlbfgsresults
minlbfgssetcond
minlbfgssetstpmax
minlbfgssetxrep

Examples

minlbfgs_1
minlbfgs_2

minlbfgscreate function

/************************************************************************* LIMITED MEMORY BFGS METHOD FOR LARGE SCALE OPTIMIZATION The subroutine minimizes function F(x) of N arguments by using a quasi- Newton method (LBFGS scheme) which is optimized to use a minimum amount of memory. The subroutine generates the approximation of an inverse Hessian matrix by using information about the last M steps of the algorithm (instead of N). It lessens a required amount of memory from a value of order N^2 to a value of order 2*N*M. INPUT PARAMETERS: N - problem dimension. N>0 M - number of corrections in the BFGS scheme of Hessian approximation update. Recommended value: 3<=M<=7. The smaller value causes worse convergence, the bigger will not cause a considerably better convergence, but will cause a fall in the performance. M<=N. X - initial solution approximation, array[0..N-1]. OUTPUT PARAMETERS: State - structure used for reverse communication. This function initializes State structure with default optimization parameters (stopping conditions, step size, etc.). Use MinLBFGSSet??????() functions to tune optimization parameters. After all optimization parameters are tuned, you should use MinLBFGSIteration() function to advance algorithm iterations. NOTES: 1. you may tune stopping conditions with MinLBFGSSetCond() function 2. if target function contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow, use MinLBFGSSetStpMax() function to bound algorithm's steps. However, L-BFGS rarely needs such a tuning. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgscreate(int n, int m, const ap::template_1d_array< amp::ampf<Precision> >& x, minlbfgsstate<Precision>& state);

Examples:   minlbfgs_1  minlbfgs_2  

minlbfgscreatex function

/************************************************************************* Extended subroutine for internal use only. Accepts additional parameters: Flags - additional settings: * Flags = 0 means no additional settings * Flags = 1 "do not allocate memory". used when solving a many subsequent tasks with same N/M values. First call MUST be without this flag bit set, subsequent calls of MinLBFGS with same MinLBFGSState structure can set Flags to 1. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgscreatex(int n, int m, const ap::template_1d_array< amp::ampf<Precision> >& x, int flags, minlbfgsstate<Precision>& state);

minlbfgsiteration function

/************************************************************************* L-BFGS iterations Called after initialization with MinLBFGSCreate() function. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLBFGSCreate() RESULT: * if function returned False, iterative proces has converged. Use MinLBFGSResults() to obtain optimization results. * if subroutine returned True, then, depending on structure fields, we have one of the following situations === FUNC/GRAD REQUEST === State.NeedFG is True => function value/gradient are needed. Caller should calculate function value State.F and gradient State.G[0..N-1] at State.X[0..N-1] and call MinLBFGSIteration() again. === NEW INTERATION IS REPORTED === State.XUpdated is True => one more iteration was made. State.X contains current position, State.F contains function value at X. You can read info from these fields, but never modify them because they contain the only copy of optimization algorithm state. One and only one of these fields (NeedFG, XUpdated) is true on return. New iterations are reported only when reports are explicitly turned on by MinLBFGSSetXRep() function, so if you never called it, you can expect that NeedFG is always True. -- ALGLIB -- Copyright 20.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool minlbfgsiteration(minlbfgsstate<Precision>& state);

Examples:   minlbfgs_1  minlbfgs_2  

minlbfgsresults function

/************************************************************************* L-BFGS algorithm results Called after MinLBFGSIteration() returned False. INPUT PARAMETERS: State - algorithm state (used by MinLBFGSIteration). OUTPUT PARAMETERS: X - array[0..N-1], solution Rep - optimization report: * Rep.TerminationType completetion code: * -2 rounding errors prevent further improvement. X contains best point found. * -1 incorrect parameters were specified * 1 relative function improvement is no more than EpsF. * 2 relative step is no more than EpsX. * 4 gradient norm is no more than EpsG * 5 MaxIts steps was taken * 7 stopping conditions are too stringent, further improvement is impossible * Rep.IterationsCount contains iterations count * NFEV countains number of function calculations -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgsresults(const minlbfgsstate<Precision>& state, ap::template_1d_array< amp::ampf<Precision> >& x, minlbfgsreport<Precision>& rep);

Examples:   minlbfgs_1  minlbfgs_2  

minlbfgssetcond function

/************************************************************************* This function sets stopping conditions for L-BFGS optimization algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLBFGSCreate() EpsG - >=0 The subroutine finishes its work if the condition ||G||<EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient. EpsF - >=0 The subroutine finishes its work if on k+1-th iteration the condition |F(k+1)-F(k)|<=EpsF*max{|F(k)|,|F(k+1)|,1} is satisfied. EpsX - >=0 The subroutine finishes its work if on k+1-th iteration the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. Passing EpsG=0, EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (small EpsX). -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgssetcond(minlbfgsstate<Precision>& state, amp::ampf<Precision> epsg, amp::ampf<Precision> epsf, amp::ampf<Precision> epsx, int maxits);

minlbfgssetstpmax function

/************************************************************************* This function sets maximum step length INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLBFGSCreate() StpMax - maximum step length, >=0. Set StpMax to 0.0, if you don't want to limit step length. Use this subroutine when you optimize target function which contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow. This function allows us to reject steps that are too large (and therefore expose us to the possible overflow) without actually calculating function value at the x+stp*d. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgssetstpmax(minlbfgsstate<Precision>& state, amp::ampf<Precision> stpmax);

minlbfgssetxrep function

/************************************************************************* This function turns on/off reporting. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLBFGSCreate() NeedXRep- whether iteration reports are needed or not Usually algorithm returns from MinLBFGSIteration() only when it needs function/gradient/ (which is indicated by NeedFG field. However, with this function we can let it stop after each iteration (one iteration may include more than one function evaluation), which is indicated by XUpdated field. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlbfgssetxrep(minlbfgsstate<Precision>& state, bool needxrep);

minlbfgs_1 example

    int n;
    int m;
    minlbfgs::minlbfgsstate<Precision> state;
    minlbfgs::minlbfgsreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Function minimized:
    //     F = exp(x-1) + exp(1-x) + (y-x)^2
    // N = 2 - task dimension
    // M = 1 - build tank-1 model
    //
    n = 2;
    m = 1;
    s.setlength(2);
    s(0) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    minlbfgs::minlbfgscreate<Precision>(n, m, s, state);
    minlbfgs::minlbfgssetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0001"), 0);
    while( minlbfgs::minlbfgsiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            state.f = amp::exp<Precision>(x-1)+amp::exp<Precision>(1-x)+amp::sqr<Precision>(y-x);
            state.g(0) = amp::exp<Precision>(x-1)-amp::exp<Precision>(1-x)+2*(x-y);
            state.g(1) = 2*(y-x);
        }
    }
    minlbfgs::minlbfgsresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("\n\nF = exp(x-1) + exp(1-x) + (y-x)^2\n");
    printf("X = %4.2lf (should be 1.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 1.00)\n\n\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));

minlbfgs_2 example

    int n;
    int m;
    minlbfgs::minlbfgsstate<Precision> state;
    minlbfgs::minlbfgsreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Function minimized:
    //     F = exp(x-1) + exp(1-x) + (y-x)^2
    // N = 2 - task dimension
    // M = 1 - build tank-1 model
    //
    n = 2;
    m = 1;
    s.setlength(2);
    s(0) = 10;
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    minlbfgs::minlbfgscreate<Precision>(n, m, s, state);
    minlbfgs::minlbfgssetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0001"), 0);
    minlbfgs::minlbfgssetxrep<Precision>(state, true);
    printf("\n\nF = exp(x-1) + exp(1-x) + (y-x)^2\n");
    printf("OPTIMIZATION STARTED\n");
    while( minlbfgs::minlbfgsiteration<Precision>(state) )
    {
        if( state.needfg )
        {
            x = state.x(0);
            y = state.x(1);
            state.f = amp::exp<Precision>(x-1)+amp::exp<Precision>(1-x)+amp::sqr<Precision>(y-x);
            state.g(0) = amp::exp<Precision>(x-1)-amp::exp<Precision>(1-x)+2*(x-y);
            state.g(1) = 2*(y-x);
        }
        if( state.xupdated )
        {
            printf("    F(%8.5lf,%8.5lf)=%0.5lf\n",
                double(amp::ampf<Precision>(state.x(0)).toDouble()),
                double(amp::ampf<Precision>(state.x(1)).toDouble()),
                double(amp::ampf<Precision>(state.f).toDouble()));
        }
    }
    printf("OPTIMIZATION STOPPED\n");
    minlbfgs::minlbfgsresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (should be 1.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (should be 1.00)\n\n\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));

minlm unit

Functions

minlmcreatefgh
minlmcreatefgj
minlmcreatefj
minlmiteration
minlmresults
minlmsetcond
minlmsetstpmax
minlmsetxrep

Examples

minlm_fgh
minlm_fgj
minlm_fj
minlm_fj2

minlmcreatefgh function

/************************************************************************* LEVENBERG-MARQUARDT-LIKE METHOD FOR NON-LINEAR OPTIMIZATION Optimization using function gradient and Hessian. Algorithm - Levenberg- Marquardt modification with L-BFGS pre-optimization and internal pre-conditioned L-BFGS optimization after each Levenberg-Marquardt step. Function F has general form (not "sum-of-squares"): F = F(x[0], ..., x[n-1]) EXAMPLE See HTML-documentation. INPUT PARAMETERS: N - dimension, N>1 X - initial solution, array[0..N-1] OUTPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls of MinLMIteration. Used for reverse communication. This structure should be passed to MinLMIteration subroutine. See also MinLMIteration, MinLMResults. NOTES: 1. you may tune stopping conditions with MinLMSetCond() function 2. if target function contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow, use MinLMSetStpMax() function to bound algorithm's steps. -- ALGLIB -- Copyright 30.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmcreatefgh(const int& n, const ap::template_1d_array< amp::ampf<Precision> >& x, minlmstate<Precision>& state);

Examples:   minlm_fgh  

minlmcreatefgj function

/************************************************************************* LEVENBERG-MARQUARDT-LIKE METHOD FOR NON-LINEAR OPTIMIZATION Optimization using function gradient and Jacobian. Algorithm - Levenberg- Marquardt modification with L-BFGS pre-optimization and internal pre-conditioned L-BFGS optimization after each Levenberg-Marquardt step. Function F is represented as sum of squares: F = f[0]^2(x[0],...,x[n-1]) + ... + f[m-1]^2(x[0],...,x[n-1]) EXAMPLE See HTML-documentation. INPUT PARAMETERS: N - dimension, N>1 M - number of functions f[i] X - initial solution, array[0..N-1] OUTPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls of MinLMIteration. Used for reverse communication. This structure should be passed to MinLMIteration subroutine. See also MinLMIteration, MinLMResults. NOTES: 1. you may tune stopping conditions with MinLMSetCond() function 2. if target function contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow, use MinLMSetStpMax() function to bound algorithm's steps. -- ALGLIB -- Copyright 30.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmcreatefgj(const int& n, const int& m, const ap::template_1d_array< amp::ampf<Precision> >& x, minlmstate<Precision>& state);

Examples:   minlm_fgj  

minlmcreatefj function

/************************************************************************* CLASSIC LEVENBERG-MARQUARDT METHOD FOR NON-LINEAR OPTIMIZATION Optimization using Jacobi matrix. Algorithm - classic Levenberg-Marquardt method. Function F is represented as sum of squares: F = f[0]^2(x[0],...,x[n-1]) + ... + f[m-1]^2(x[0],...,x[n-1]) EXAMPLE See HTML-documentation. INPUT PARAMETERS: N - dimension, N>1 M - number of functions f[i] X - initial solution, array[0..N-1] OUTPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls of MinLMIteration. Used for reverse communication. This structure should be passed to MinLMIteration subroutine. See also MinLMIteration, MinLMResults. NOTES: 1. you may tune stopping conditions with MinLMSetCond() function 2. if target function contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow, use MinLMSetStpMax() function to bound algorithm's steps. -- ALGLIB -- Copyright 30.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmcreatefj(const int& n, const int& m, const ap::template_1d_array< amp::ampf<Precision> >& x, minlmstate<Precision>& state);

Examples:   minlm_fj  minlm_fj2  

minlmiteration function

/************************************************************************* One Levenberg-Marquardt iteration. Called after inialization of State structure with MinLMXXX subroutine. See HTML docs for examples. Input parameters: State - structure which stores algorithm state between subsequent calls and which is used for reverse communication. Must be initialized with MinLMXXX call first. If subroutine returned False, iterative algorithm has converged. If subroutine returned True, then: * if State.NeedF=True, - function value F at State.X[0..N-1] is required * if State.NeedFG=True - function value F and gradient G are required * if State.NeedFiJ=True - function vector f[i] and Jacobi matrix J are required * if State.NeedFGH=True - function value F, gradient G and Hesian H are required * if State.XUpdated=True - algorithm reports about new iteration, State.X contains current point, State.F contains function value. One and only one of this fields can be set at time. Results are stored: * function value - in MinLMState.F * gradient - in MinLMState.G[0..N-1] * Jacobi matrix - in MinLMState.J[0..M-1,0..N-1] * Hessian - in MinLMState.H[0..N-1,0..N-1] -- ALGLIB -- Copyright 10.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool minlmiteration(minlmstate<Precision>& state);

Examples:   minlm_fgh  minlm_fgj  minlm_fj  minlm_fj2  

minlmresults function

/************************************************************************* Levenberg-Marquardt algorithm results Called after MinLMIteration returned False. Input parameters: State - algorithm state (used by MinLMIteration). Output parameters: X - array[0..N-1], solution Rep - optimization report: * Rep.TerminationType completetion code: * -1 incorrect parameters were specified * 1 relative function improvement is no more than EpsF. * 2 relative step is no more than EpsX. * 4 gradient is no more than EpsG. * 5 MaxIts steps was taken * 7 stopping conditions are too stringent, further improvement is impossible * Rep.IterationsCount contains iterations count * Rep.NFunc - number of function calculations * Rep.NJac - number of Jacobi matrix calculations * Rep.NGrad - number of gradient calculations * Rep.NHess - number of Hessian calculations * Rep.NCholesky - number of Cholesky decomposition calculations -- ALGLIB -- Copyright 10.03.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmresults(const minlmstate<Precision>& state, ap::template_1d_array< amp::ampf<Precision> >& x, minlmreport<Precision>& rep);

Examples:   minlm_fgh  minlm_fgj  minlm_fj  minlm_fj2  

minlmsetcond function

/************************************************************************* This function sets stopping conditions for Levenberg-Marquardt optimization algorithm. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLMCreate???() EpsG - >=0 The subroutine finishes its work if the condition ||G||<EpsG is satisfied, where ||.|| means Euclidian norm, G - gradient. EpsF - >=0 The subroutine finishes its work if on k+1-th iteration the condition |F(k+1)-F(k)|<=EpsF*max{|F(k)|,|F(k+1)|,1} is satisfied. EpsX - >=0 The subroutine finishes its work if on k+1-th iteration the condition |X(k+1)-X(k)| <= EpsX is fulfilled. MaxIts - maximum number of iterations. If MaxIts=0, the number of iterations is unlimited. Only Levenberg-Marquardt iterations are counted (L-BFGS/CG iterations are NOT counted because their cost is very low copared to that of LM). Passing EpsG=0, EpsF=0, EpsX=0 and MaxIts=0 (simultaneously) will lead to automatic stopping criterion selection (small EpsX). -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmsetcond(minlmstate<Precision>& state, amp::ampf<Precision> epsg, amp::ampf<Precision> epsf, amp::ampf<Precision> epsx, int maxits);

minlmsetstpmax function

/************************************************************************* This function sets maximum step length INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinCGCreate???() StpMax - maximum step length, >=0. Set StpMax to 0.0, if you don't want to limit step length. Use this subroutine when you optimize target function which contains exp() or other fast growing functions, and optimization algorithm makes too large steps which leads to overflow. This function allows us to reject steps that are too large (and therefore expose us to the possible overflow) without actually calculating function value at the x+stp*d. NOTE: non-zero StpMax leads to moderate performance degradation because intermediate step of preconditioned L-BFGS optimization is incompatible with limits on step size. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmsetstpmax(minlmstate<Precision>& state, amp::ampf<Precision> stpmax);

minlmsetxrep function

/************************************************************************* This function turns on/off reporting. INPUT PARAMETERS: State - structure which stores algorithm state between calls and which is used for reverse communication. Must be initialized with MinLMCreate???() NeedXRep- whether iteration reports are needed or not Usually algorithm returns from MinLMIteration() only when it needs function/gradient/Hessian. However, with this function we can let it stop after each iteration (one iteration may include more than one function evaluation), which is indicated by XUpdated field. Both Levenberg-Marquardt and L-BFGS iterations are reported. -- ALGLIB -- Copyright 02.04.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void minlmsetxrep(minlmstate<Precision>& state, bool needxrep);

minlm_fgh example

    minlm::minlmstate<Precision> state;
    minlm::minlmreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Example of solving simple task using FGH scheme.
    //
    // Function minimized:
    //     F = (x-2*y)^2 + (x-2)^2 + (y-1)^2
    // exact solution is (2,1).
    //
    s.setlength(2);
    s(0) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    minlm::minlmcreatefgh<Precision>(2, s, state);
    minlm::minlmsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.001"), 0);
    while( minlm::minlmiteration<Precision>(state) )
    {
        x = state.x(0);
        y = state.x(1);
        if( state.needf )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
        }
        if( state.needfg )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
            state.g(0) = 2*(x-2*y)+2*(x-2)+0;
            state.g(1) = -4*(x-2*y)+0+2*(y-1);
        }
        if( state.needfgh )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
            state.g(0) = 2*(x-2*y)+2*(x-2)+0;
            state.g(1) = -4*(x-2*y)+0+2*(y-1);
            state.h(0,0) = 4;
            state.h(1,0) = -4;
            state.h(0,1) = -4;
            state.h(1,1) = 10;
        }
    }
    minlm::minlmresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (correct value - 2.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (correct value - 1.00)\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("TerminationType = %0ld (should be 2 - stopping when step is small enough)\n",
        long(rep.terminationtype));
    printf("NFunc = %0ld\n",
        long(rep.nfunc));
    printf("NJac  = %0ld\n",
        long(rep.njac));
    printf("NGrad = %0ld\n",
        long(rep.ngrad));
    printf("NHess = %0ld\n",
        long(rep.nhess));

minlm_fgj example

    minlm::minlmstate<Precision> state;
    minlm::minlmreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Example of solving simple task using FGJ scheme.
    //
    // Function minimized:
    //     F = (x-2*y)^2 + (x-2)^2 + (y-1)^2
    // exact solution is (2,1).
    //
    s.setlength(2);
    s(0) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    minlm::minlmcreatefgj<Precision>(2, 3, s, state);
    minlm::minlmsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.001"), 0);
    while( minlm::minlmiteration<Precision>(state) )
    {
        x = state.x(0);
        y = state.x(1);
        if( state.needf )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
        }
        if( state.needfg )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
            state.g(0) = 2*(x-2*y)+2*(x-2)+0;
            state.g(1) = -4*(x-2*y)+0+2*(y-1);
        }
        if( state.needfij )
        {
            state.fi(0) = x-2*y;
            state.fi(1) = x-2;
            state.fi(2) = y-1;
            state.j(0,0) = 1;
            state.j(0,1) = -2;
            state.j(1,0) = 1;
            state.j(1,1) = 0;
            state.j(2,0) = 0;
            state.j(2,1) = 1;
        }
    }
    minlm::minlmresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (correct value - 2.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (correct value - 1.00)\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("TerminationType = %0ld (should be 2 - stopping when step is small enough)\n",
        long(rep.terminationtype));
    printf("NFunc = %0ld\n",
        long(rep.nfunc));
    printf("NJac  = %0ld\n",
        long(rep.njac));
    printf("NGrad = %0ld\n",
        long(rep.ngrad));
    printf("NHess = %0ld\n",
        long(rep.nhess));

minlm_fj example

    minlm::minlmstate<Precision> state;
    minlm::minlmreport<Precision> rep;
    ap::template_1d_array< amp::ampf<Precision> > s;
    amp::ampf<Precision> x;
    amp::ampf<Precision> y;


    
    //
    // Example of solving simple task using FJ scheme.
    //
    // Function minimized:
    //     F = (x-2*y)^2 + (x-2)^2 + (y-1)^2
    // exact solution is (2,1).
    //
    s.setlength(2);
    s(0) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    s(1) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    minlm::minlmcreatefj<Precision>(2, 3, s, state);
    minlm::minlmsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.001"), 0);
    while( minlm::minlmiteration<Precision>(state) )
    {
        x = state.x(0);
        y = state.x(1);
        if( state.needf )
        {
            state.f = amp::sqr<Precision>(x-2*y)+amp::sqr<Precision>(x-2)+amp::sqr<Precision>(y-1);
        }
        if( state.needfij )
        {
            state.fi(0) = x-2*y;
            state.fi(1) = x-2;
            state.fi(2) = y-1;
            state.j(0,0) = 1;
            state.j(0,1) = -2;
            state.j(1,0) = 1;
            state.j(1,1) = 0;
            state.j(2,0) = 0;
            state.j(2,1) = 1;
        }
    }
    minlm::minlmresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("X = %4.2lf (correct value - 2.00)\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("Y = %4.2lf (correct value - 1.00)\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("TerminationType = %0ld (should be 2 - stopping when step is small enough)\n",
        long(rep.terminationtype));
    printf("NFunc = %0ld\n",
        long(rep.nfunc));
    printf("NJac  = %0ld\n",
        long(rep.njac));
    printf("NGrad = %0ld\n",
        long(rep.ngrad));
    printf("NHess = %0ld\n",
        long(rep.nhess));

minlm_fj2 example

    minlm::minlmstate<Precision> state;
    minlm::minlmreport<Precision> rep;
    int i;
    ap::template_1d_array< amp::ampf<Precision> > s;
    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    amp::ampf<Precision> fi;
    int n;
    int m;


    
    //
    // Example of solving polynomial approximation task using FJ scheme.
    //
    // Data points:
    //     xi are random numbers from [-1,+1],
    //
    // Function being fitted:
    //     yi = exp(xi) - sin(xi) - x^3/3
    //
    // Function being minimized:
    //     F(a,b,c) =
    //         (a + b*x0 + c*x0^2 - y0)^2 +
    //         (a + b*x1 + c*x1^2 - y1)^2 + ...
    //
    n = 3;
    s.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        s(i) = amp::ampf<Precision>::getRandom()-amp::ampf<Precision>("0.5");
    }
    m = 100;
    x.setlength(m);
    y.setlength(m);
    for(i=0; i<=m-1; i++)
    {
        x(i) = amp::ampf<Precision>(2*i)/(amp::ampf<Precision>(m-1))-1;
        y(i) = amp::exp<Precision>(x(i))-amp::sin<Precision>(x(i))-x(i)*x(i)*x(i)/3;
    }
    
    //
    // Now S stores starting point, X and Y store points being fitted.
    //
    minlm::minlmcreatefj<Precision>(n, m, s, state);
    minlm::minlmsetcond<Precision>(state, amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.0"), amp::ampf<Precision>("0.001"), 0);
    while( minlm::minlmiteration<Precision>(state) )
    {
        if( state.needf )
        {
            state.f = 0;
        }
        for(i=0; i<=m-1; i++)
        {
            
            //
            // "a" is stored in State.X[0]
            // "b" - State.X[1]
            // "c" - State.X[2]
            //
            fi = state.x(0)+state.x(1)*x(i)+state.x(2)*amp::sqr<Precision>(x(i))-y(i);
            if( state.needf )
            {
                
                //
                // F is equal to sum of fi squared.
                //
                state.f = state.f+amp::sqr<Precision>(fi);
            }
            if( state.needfij )
            {
                
                //
                // Fi
                //
                state.fi(i) = fi;
                
                //
                // dFi/da
                //
                state.j(i,0) = 1;
                
                //
                // dFi/db
                //
                state.j(i,1) = x(i);
                
                //
                // dFi/dc
                //
                state.j(i,2) = amp::sqr<Precision>(x(i));
            }
        }
    }
    minlm::minlmresults<Precision>(state, s, rep);
    
    //
    // output results
    //
    printf("A = %4.2lf\n",
        double(amp::ampf<Precision>(s(0)).toDouble()));
    printf("B = %4.2lf\n",
        double(amp::ampf<Precision>(s(1)).toDouble()));
    printf("C = %4.2lf\n",
        double(amp::ampf<Precision>(s(2)).toDouble()));
    printf("TerminationType = %0ld (should be 2 - stopping when step is small enough)\n",
        long(rep.terminationtype));

odesolver unit

Functions

odesolveriteration
odesolverresults
odesolverrkck

Examples

ode_example1
ode_example2

odesolveriteration function

/************************************************************************* One iteration of ODE solver. Called after inialization of State structure with OdeSolverXXX subroutine. See HTML docs for examples. INPUT PARAMETERS: State - structure which stores algorithm state between subsequent calls and which is used for reverse communication. Must be initialized with OdeSolverXXX() call first. If subroutine returned False, algorithm have finished its work. If subroutine returned True, then user should: * calculate F(State.X, State.Y) * store it in State.DY Here State.X is real, State.Y and State.DY are arrays[0..N-1] of reals. -- ALGLIB -- Copyright 01.09.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool odesolveriteration(odesolverstate<Precision>& state);

Examples:   ode_example1  ode_example2  

odesolverresults function

/************************************************************************* ODE solver results Called after OdeSolverIteration returned False. INPUT PARAMETERS: State - algorithm state (used by OdeSolverIteration). OUTPUT PARAMETERS: M - number of tabulated values, M>=1 XTbl - array[0..M-1], values of X YTbl - array[0..M-1,0..N-1], values of Y in X[i] Rep - solver report: * Rep.TerminationType completetion code: * -2 X is not ordered by ascending/descending or there are non-distinct X[], i.e. X[i]=X[i+1] * -1 incorrect parameters were specified * 1 task has been solved * Rep.NFEV contains number of function calculations -- ALGLIB -- Copyright 01.09.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void odesolverresults(const odesolverstate<Precision>& state, int& m, ap::template_1d_array< amp::ampf<Precision> >& xtbl, ap::template_2d_array< amp::ampf<Precision> >& ytbl, odesolverreport<Precision>& rep);

Examples:   ode_example1  ode_example2  

odesolverrkck function

/************************************************************************* Cash-Karp adaptive ODE solver. This subroutine solves ODE Y'=f(Y,x) with initial conditions Y(xs)=Ys (here Y may be single variable or vector of N variables). INPUT PARAMETERS: Y - initial conditions, array[0..N-1]. contains values of Y[] at X[0] N - system size X - points at which Y should be tabulated, array[0..M-1] integrations starts at X[0], ends at X[M-1], intermediate values at X[i] are returned too. SHOULD BE ORDERED BY ASCENDING OR BY DESCENDING!!!! M - number of intermediate points + first point + last point: * M>2 means that you need both Y(X[M-1]) and M-2 values at intermediate points * M=2 means that you want just to integrate from X[0] to X[1] and don't interested in intermediate values. * M=1 means that you don't want to integrate :) it is degenerate case, but it will be handled correctly. * M<1 means error Eps - tolerance (absolute/relative error on each step will be less than Eps). When passing: * Eps>0, it means desired ABSOLUTE error * Eps<0, it means desired RELATIVE error. Relative errors are calculated with respect to maximum values of Y seen so far. Be careful to use this criterion when starting from Y[] that are close to zero. H - initial step lenth, it will be adjusted automatically after the first step. If H=0, step will be selected automatically (usualy it will be equal to 0.001 of min(x[i]-x[j])). OUTPUT PARAMETERS State - structure which stores algorithm state between subsequent calls of OdeSolverIteration. Used for reverse communication. This structure should be passed to the OdeSolverIteration subroutine. SEE ALSO AutoGKSmoothW, AutoGKSingular, AutoGKIteration, AutoGKResults. -- ALGLIB -- Copyright 01.09.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void odesolverrkck(const ap::template_1d_array< amp::ampf<Precision> >& y, int n, const ap::template_1d_array< amp::ampf<Precision> >& x, int m, amp::ampf<Precision> eps, amp::ampf<Precision> h, odesolverstate<Precision>& state);

Examples:   ode_example1  ode_example2  

ode_example1 example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_2d_array< amp::ampf<Precision> > ytbl;
    amp::ampf<Precision> eps;
    amp::ampf<Precision> h;
    int m;
    int i;
    odesolver::odesolverstate<Precision> state;
    odesolver::odesolverreport<Precision> rep;


    
    //
    // ODESolver unit is used to solve simple ODE:
    // y' = y, y(0) = 1.
    //
    // Its solution is well known in academic circles :)
    //
    // No intermediate values are calculated,
    // just starting and final points.
    //
    y.setlength(1);
    y(0) = 1;
    x.setlength(2);
    x(0) = 0;
    x(1) = 1;
    eps = amp::ampf<Precision>("1.0E-4");
    h = amp::ampf<Precision>("0.01");
    odesolver::odesolverrkck<Precision>(y, 1, x, 2, eps, h, state);
    while( odesolver::odesolveriteration<Precision>(state) )
    {
        state.dy(0) = state.y(0);
    }
    odesolver::odesolverresults<Precision>(state, m, x, ytbl, rep);
    printf("    X  Y(X)\n");
    for(i=0; i<=m-1; i++)
    {
        printf("%5.3lf %5.3lf\n",
            double(amp::ampf<Precision>(x(i)).toDouble()),
            double(amp::ampf<Precision>(ytbl(i,0)).toDouble()));
    }

ode_example2 example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_2d_array< amp::ampf<Precision> > ytbl;
    amp::ampf<Precision> eps;
    amp::ampf<Precision> h;
    int m;
    int i;
    odesolver::odesolverstate<Precision> state;
    odesolver::odesolverreport<Precision> rep;


    
    //
    // ODESolver unit is used to solve simple ODE:
    // y'' = -y, y(0) = 0, y'(0)=1.
    //
    // This ODE may be written as first-order system:
    // y' =  z
    // z' = -y
    //
    // Its solution is well known in academic circles :)
    //
    // Three intermediate values are calculated,
    // plus starting and final points.
    //
    y.setlength(2);
    y(0) = 0;
    y(1) = 1;
    x.setlength(5);
    x(0) = amp::pi<Precision>()*0/4;
    x(1) = amp::pi<Precision>()*1/4;
    x(2) = amp::pi<Precision>()*2/4;
    x(3) = amp::pi<Precision>()*3/4;
    x(4) = amp::pi<Precision>()*4/4;
    eps = amp::ampf<Precision>("1.0E-8");
    h = amp::ampf<Precision>("0.01");
    odesolver::odesolverrkck<Precision>(y, 2, x, 5, eps, h, state);
    while( odesolver::odesolveriteration<Precision>(state) )
    {
        state.dy(0) = state.y(1);
        state.dy(1) = -state.y(0);
    }
    odesolver::odesolverresults<Precision>(state, m, x, ytbl, rep);
    printf("     X   Y(X)     Error\n");
    for(i=0; i<=m-1; i++)
    {
        printf("%6.3lf %6.3lf  %8.1le\n",
            double(amp::ampf<Precision>(x(i)).toDouble()),
            double(amp::ampf<Precision>(ytbl(i,0)).toDouble()),
            double(amp::ampf<Precision>(amp::abs<Precision>(ytbl(i,0)-amp::sin<Precision>(x(i)))).toDouble()));
    }

ortfac unit

Functions

cmatrixlq
cmatrixlqunpackl
cmatrixlqunpackq
cmatrixqr
cmatrixqrunpackq
cmatrixqrunpackr
hmatrixtd
hmatrixtdunpackq
rmatrixbd
rmatrixbdmultiplybyp
rmatrixbdmultiplybyq
rmatrixbdunpackdiagonals
rmatrixbdunpackpt
rmatrixbdunpackq
rmatrixhessenberg
rmatrixhessenbergunpackh
rmatrixhessenbergunpackq
rmatrixlq
rmatrixlqunpackl
rmatrixlqunpackq
rmatrixqr
rmatrixqrunpackq
rmatrixqrunpackr
smatrixtd
smatrixtdunpackq

cmatrixlq function

/************************************************************************* LQ decomposition of a rectangular complex matrix of size MxN Input parameters: A - matrix A whose indexes range within [0..M-1, 0..N-1] M - number of rows in matrix A. N - number of columns in matrix A. Output parameters: A - matrices Q and L in compact form Tau - array of scalar factors which are used to form matrix Q. Array whose indexes range within [0.. Min(M,N)-1] Matrix A is represented as A = LQ, where Q is an orthogonal matrix of size MxM, L - lower triangular (or lower trapezoid) matrix of size MxN. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University September 30, 1994 *************************************************************************/
template<unsigned int Precision> void cmatrixlq(ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, ap::template_1d_array< amp::campf<Precision> >& tau);

cmatrixlqunpackl function

/************************************************************************* Unpacking of matrix L from the LQ decomposition of a matrix A Input parameters: A - matrices Q and L in compact form. Output of CMatrixLQ subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Output parameters: L - matrix L, array[0..M-1, 0..N-1]. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlqunpackl(const ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, ap::template_2d_array< amp::campf<Precision> >& l);

cmatrixlqunpackq function

/************************************************************************* Partial unpacking of matrix Q from LQ decomposition of a complex matrix A. Input parameters: A - matrices Q and R in compact form. Output of CMatrixLQ subroutine . M - number of rows in matrix A. M>=0. N - number of columns in matrix A. N>=0. Tau - scalar factors which are used to form Q. Output of CMatrixLQ subroutine . QRows - required number of rows in matrix Q. N>=QColumns>=0. Output parameters: Q - first QRows rows of matrix Q. Array whose index ranges within [0..QRows-1, 0..N-1]. If QRows=0, array isn't changed. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlqunpackq(const ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, const ap::template_1d_array< amp::campf<Precision> >& tau, int qrows, ap::template_2d_array< amp::campf<Precision> >& q);

cmatrixqr function

/************************************************************************* QR decomposition of a rectangular complex matrix of size MxN Input parameters: A - matrix A whose indexes range within [0..M-1, 0..N-1] M - number of rows in matrix A. N - number of columns in matrix A. Output parameters: A - matrices Q and R in compact form Tau - array of scalar factors which are used to form matrix Q. Array whose indexes range within [0.. Min(M,N)-1] Matrix A is represented as A = QR, where Q is an orthogonal matrix of size MxM, R - upper triangular (or upper trapezoid) matrix of size MxN. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University September 30, 1994 *************************************************************************/
template<unsigned int Precision> void cmatrixqr(ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, ap::template_1d_array< amp::campf<Precision> >& tau);

cmatrixqrunpackq function

/************************************************************************* Partial unpacking of matrix Q from QR decomposition of a complex matrix A. Input parameters: A - matrices Q and R in compact form. Output of CMatrixQR subroutine . M - number of rows in matrix A. M>=0. N - number of columns in matrix A. N>=0. Tau - scalar factors which are used to form Q. Output of CMatrixQR subroutine . QColumns - required number of columns in matrix Q. M>=QColumns>=0. Output parameters: Q - first QColumns columns of matrix Q. Array whose index ranges within [0..M-1, 0..QColumns-1]. If QColumns=0, array isn't changed. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixqrunpackq(const ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, const ap::template_1d_array< amp::campf<Precision> >& tau, int qcolumns, ap::template_2d_array< amp::campf<Precision> >& q);

cmatrixqrunpackr function

/************************************************************************* Unpacking of matrix R from the QR decomposition of a matrix A Input parameters: A - matrices Q and R in compact form. Output of CMatrixQR subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Output parameters: R - matrix R, array[0..M-1, 0..N-1]. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixqrunpackr(const ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, ap::template_2d_array< amp::campf<Precision> >& r);

hmatrixtd function

/************************************************************************* Reduction of a Hermitian matrix which is given by its higher or lower triangular part to a real tridiagonal matrix using unitary similarity transformation: Q'*A*Q = T. Input parameters: A - matrix to be transformed array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then matrix A is given by its upper triangle, and the lower triangle is not used and not modified by the algorithm, and vice versa if IsUpper = False. Output parameters: A - matrices T and Q in compact form (see lower) Tau - array of factors which are forming matrices H(i) array with elements [0..N-2]. D - main diagonal of real symmetric matrix T. array with elements [0..N-1]. E - secondary diagonal of real symmetric matrix T. array with elements [0..N-2]. If IsUpper=True, the matrix Q is represented as a product of elementary reflectors Q = H(n-2) . . . H(2) H(0). Each H(i) has the form H(i) = I - tau * v * v' where tau is a complex scalar, and v is a complex vector with v(i+1:n-1) = 0, v(i) = 1, v(0:i-1) is stored on exit in A(0:i-1,i+1), and tau in TAU(i). If IsUpper=False, the matrix Q is represented as a product of elementary reflectors Q = H(0) H(2) . . . H(n-2). Each H(i) has the form H(i) = I - tau * v * v' where tau is a complex scalar, and v is a complex vector with v(0:i) = 0, v(i+1) = 1, v(i+2:n-1) is stored on exit in A(i+2:n-1,i), and tau in TAU(i). The contents of A on exit are illustrated by the following examples with n = 5: if UPLO = 'U': if UPLO = 'L': ( d e v1 v2 v3 ) ( d ) ( d e v2 v3 ) ( e d ) ( d e v3 ) ( v0 e d ) ( d e ) ( v0 v1 e d ) ( d ) ( v0 v1 v2 e d ) where d and e denote diagonal and off-diagonal elements of T, and vi denotes an element of the vector defining H(i). -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University October 31, 1992 *************************************************************************/
template<unsigned int Precision> void hmatrixtd(ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, ap::template_1d_array< amp::campf<Precision> >& tau, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_1d_array< amp::ampf<Precision> >& e);

hmatrixtdunpackq function

/************************************************************************* Unpacking matrix Q which reduces a Hermitian matrix to a real tridiagonal form. Input parameters: A - the result of a HMatrixTD subroutine N - size of matrix A. IsUpper - storage format (a parameter of HMatrixTD subroutine) Tau - the result of a HMatrixTD subroutine Output parameters: Q - transformation matrix. array with elements [0..N-1, 0..N-1]. -- ALGLIB -- Copyright 2005-2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void hmatrixtdunpackq(const ap::template_2d_array< amp::campf<Precision> >& a, const int& n, const bool& isupper, const ap::template_1d_array< amp::campf<Precision> >& tau, ap::template_2d_array< amp::campf<Precision> >& q);

rmatrixbd function

/************************************************************************* Reduction of a rectangular matrix to bidiagonal form The algorithm reduces the rectangular matrix A to bidiagonal form by orthogonal transformations P and Q: A = Q*B*P. Input parameters: A - source matrix. array[0..M-1, 0..N-1] M - number of rows in matrix A. N - number of columns in matrix A. Output parameters: A - matrices Q, B, P in compact form (see below). TauQ - scalar factors which are used to form matrix Q. TauP - scalar factors which are used to form matrix P. The main diagonal and one of the secondary diagonals of matrix A are replaced with bidiagonal matrix B. Other elements contain elementary reflections which form MxM matrix Q and NxN matrix P, respectively. If M>=N, B is the upper bidiagonal MxN matrix and is stored in the corresponding elements of matrix A. Matrix Q is represented as a product of elementary reflections Q = H(0)*H(1)*...*H(n-1), where H(i) = 1-tau*v*v'. Here tau is a scalar which is stored in TauQ[i], and vector v has the following structure: v(0:i-1)=0, v(i)=1, v(i+1:m-1) is stored in elements A(i+1:m-1,i). Matrix P is as follows: P = G(0)*G(1)*...*G(n-2), where G(i) = 1 - tau*u*u'. Tau is stored in TauP[i], u(0:i)=0, u(i+1)=1, u(i+2:n-1) is stored in elements A(i,i+2:n-1). If M<N, B is the lower bidiagonal MxN matrix and is stored in the corresponding elements of matrix A. Q = H(0)*H(1)*...*H(m-2), where H(i) = 1 - tau*v*v', tau is stored in TauQ, v(0:i)=0, v(i+1)=1, v(i+2:m-1) is stored in elements A(i+2:m-1,i). P = G(0)*G(1)*...*G(m-1), G(i) = 1-tau*u*u', tau is stored in TauP, u(0:i-1)=0, u(i)=1, u(i+1:n-1) is stored in A(i,i+1:n-1). EXAMPLE: m=6, n=5 (m > n): m=5, n=6 (m < n): ( d e u1 u1 u1 ) ( d u1 u1 u1 u1 u1 ) ( v1 d e u2 u2 ) ( e d u2 u2 u2 u2 ) ( v1 v2 d e u3 ) ( v1 e d u3 u3 u3 ) ( v1 v2 v3 d e ) ( v1 v2 e d u4 u4 ) ( v1 v2 v3 v4 d ) ( v1 v2 v3 e d u5 ) ( v1 v2 v3 v4 v5 ) Here vi and ui are vectors which form H(i) and G(i), and d and e - are the diagonal and off-diagonal elements of matrix B. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University September 30, 1994. Sergey Bochkanov, ALGLIB project, translation from FORTRAN to pseudocode, 2007-2010. *************************************************************************/
template<unsigned int Precision> void rmatrixbd(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_1d_array< amp::ampf<Precision> >& tauq, ap::template_1d_array< amp::ampf<Precision> >& taup);

rmatrixbdmultiplybyp function

/************************************************************************* Multiplication by matrix P which reduces matrix A to bidiagonal form. The algorithm allows pre- or post-multiply by P or P'. Input parameters: QP - matrices Q and P in compact form. Output of RMatrixBD subroutine. M - number of rows in matrix A. N - number of columns in matrix A. TAUP - scalar factors which are used to form P. Output of RMatrixBD subroutine. Z - multiplied matrix. Array whose indexes range within [0..ZRows-1,0..ZColumns-1]. ZRows - number of rows in matrix Z. If FromTheRight=False, ZRows=N, otherwise ZRows can be arbitrary. ZColumns - number of columns in matrix Z. If FromTheRight=True, ZColumns=N, otherwise ZColumns can be arbitrary. FromTheRight - pre- or post-multiply. DoTranspose - multiply by P or P'. Output parameters: Z - product of Z and P. Array whose indexes range within [0..ZRows-1,0..ZColumns-1]. If ZRows=0 or ZColumns=0, the array is not modified. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixbdmultiplybyp(const ap::template_2d_array< amp::ampf<Precision> >& qp, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& taup, ap::template_2d_array< amp::ampf<Precision> >& z, int zrows, int zcolumns, bool fromtheright, bool dotranspose);

rmatrixbdmultiplybyq function

/************************************************************************* Multiplication by matrix Q which reduces matrix A to bidiagonal form. The algorithm allows pre- or post-multiply by Q or Q'. Input parameters: QP - matrices Q and P in compact form. Output of ToBidiagonal subroutine. M - number of rows in matrix A. N - number of columns in matrix A. TAUQ - scalar factors which are used to form Q. Output of ToBidiagonal subroutine. Z - multiplied matrix. array[0..ZRows-1,0..ZColumns-1] ZRows - number of rows in matrix Z. If FromTheRight=False, ZRows=M, otherwise ZRows can be arbitrary. ZColumns - number of columns in matrix Z. If FromTheRight=True, ZColumns=M, otherwise ZColumns can be arbitrary. FromTheRight - pre- or post-multiply. DoTranspose - multiply by Q or Q'. Output parameters: Z - product of Z and Q. Array[0..ZRows-1,0..ZColumns-1] If ZRows=0 or ZColumns=0, the array is not modified. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixbdmultiplybyq(const ap::template_2d_array< amp::ampf<Precision> >& qp, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& tauq, ap::template_2d_array< amp::ampf<Precision> >& z, int zrows, int zcolumns, bool fromtheright, bool dotranspose);

rmatrixbdunpackdiagonals function

/************************************************************************* Unpacking of the main and secondary diagonals of bidiagonal decomposition of matrix A. Input parameters: B - output of RMatrixBD subroutine. M - number of rows in matrix B. N - number of columns in matrix B. Output parameters: IsUpper - True, if the matrix is upper bidiagonal. otherwise IsUpper is False. D - the main diagonal. Array whose index ranges within [0..Min(M,N)-1]. E - the secondary diagonal (upper or lower, depending on the value of IsUpper). Array index ranges within [0..Min(M,N)-1], the last element is not used. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixbdunpackdiagonals(const ap::template_2d_array< amp::ampf<Precision> >& b, int m, int n, bool& isupper, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_1d_array< amp::ampf<Precision> >& e);

rmatrixbdunpackpt function

/************************************************************************* Unpacking matrix P which reduces matrix A to bidiagonal form. The subroutine returns transposed matrix P. Input parameters: QP - matrices Q and P in compact form. Output of ToBidiagonal subroutine. M - number of rows in matrix A. N - number of columns in matrix A. TAUP - scalar factors which are used to form P. Output of ToBidiagonal subroutine. PTRows - required number of rows of matrix P^T. N >= PTRows >= 0. Output parameters: PT - first PTRows columns of matrix P^T Array[0..PTRows-1, 0..N-1] If PTRows=0, the array is not modified. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixbdunpackpt(const ap::template_2d_array< amp::ampf<Precision> >& qp, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& taup, int ptrows, ap::template_2d_array< amp::ampf<Precision> >& pt);

rmatrixbdunpackq function

/************************************************************************* Unpacking matrix Q which reduces a matrix to bidiagonal form. Input parameters: QP - matrices Q and P in compact form. Output of ToBidiagonal subroutine. M - number of rows in matrix A. N - number of columns in matrix A. TAUQ - scalar factors which are used to form Q. Output of ToBidiagonal subroutine. QColumns - required number of columns in matrix Q. M>=QColumns>=0. Output parameters: Q - first QColumns columns of matrix Q. Array[0..M-1, 0..QColumns-1] If QColumns=0, the array is not modified. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixbdunpackq(const ap::template_2d_array< amp::ampf<Precision> >& qp, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& tauq, int qcolumns, ap::template_2d_array< amp::ampf<Precision> >& q);

rmatrixhessenberg function

/************************************************************************* Reduction of a square matrix to upper Hessenberg form: Q'*A*Q = H, where Q is an orthogonal matrix, H - Hessenberg matrix. Input parameters: A - matrix A with elements [0..N-1, 0..N-1] N - size of matrix A. Output parameters: A - matrices Q and P in compact form (see below). Tau - array of scalar factors which are used to form matrix Q. Array whose index ranges within [0..N-2] Matrix H is located on the main diagonal, on the lower secondary diagonal and above the main diagonal of matrix A. The elements which are used to form matrix Q are situated in array Tau and below the lower secondary diagonal of matrix A as follows: Matrix Q is represented as a product of elementary reflections Q = H(0)*H(2)*...*H(n-2), where each H(i) is given by H(i) = 1 - tau * v * (v^T) where tau is a scalar stored in Tau[I]; v - is a real vector, so that v(0:i) = 0, v(i+1) = 1, v(i+2:n-1) stored in A(i+2:n-1,i). -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University October 31, 1992 *************************************************************************/
template<unsigned int Precision> void rmatrixhessenberg(ap::template_2d_array< amp::ampf<Precision> >& a, int n, ap::template_1d_array< amp::ampf<Precision> >& tau);

rmatrixhessenbergunpackh function

/************************************************************************* Unpacking matrix H (the result of matrix A reduction to upper Hessenberg form) Input parameters: A - output of RMatrixHessenberg subroutine. N - size of matrix A. Output parameters: H - matrix H. Array whose indexes range within [0..N-1, 0..N-1]. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixhessenbergunpackh(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, ap::template_2d_array< amp::ampf<Precision> >& h);

rmatrixhessenbergunpackq function

/************************************************************************* Unpacking matrix Q which reduces matrix A to upper Hessenberg form Input parameters: A - output of RMatrixHessenberg subroutine. N - size of matrix A. Tau - scalar factors which are used to form Q. Output of RMatrixHessenberg subroutine. Output parameters: Q - matrix Q. Array whose indexes range within [0..N-1, 0..N-1]. -- ALGLIB -- 2005-2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixhessenbergunpackq(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, const ap::template_1d_array< amp::ampf<Precision> >& tau, ap::template_2d_array< amp::ampf<Precision> >& q);

rmatrixlq function

/************************************************************************* LQ decomposition of a rectangular matrix of size MxN Input parameters: A - matrix A whose indexes range within [0..M-1, 0..N-1]. M - number of rows in matrix A. N - number of columns in matrix A. Output parameters: A - matrices L and Q in compact form (see below) Tau - array of scalar factors which are used to form matrix Q. Array whose index ranges within [0..Min(M,N)-1]. Matrix A is represented as A = LQ, where Q is an orthogonal matrix of size MxM, L - lower triangular (or lower trapezoid) matrix of size M x N. The elements of matrix L are located on and below the main diagonal of matrix A. The elements which are located in Tau array and above the main diagonal of matrix A are used to form matrix Q as follows: Matrix Q is represented as a product of elementary reflections Q = H(k-1)*H(k-2)*...*H(1)*H(0), where k = min(m,n), and each H(i) is of the form H(i) = 1 - tau * v * (v^T) where tau is a scalar stored in Tau[I]; v - real vector, so that v(0:i-1)=0, v(i) = 1, v(i+1:n-1) stored in A(i,i+1:n-1). -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlq(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_1d_array< amp::ampf<Precision> >& tau);

rmatrixlqunpackl function

/************************************************************************* Unpacking of matrix L from the LQ decomposition of a matrix A Input parameters: A - matrices Q and L in compact form. Output of RMatrixLQ subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Output parameters: L - matrix L, array[0..M-1, 0..N-1]. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlqunpackl(const ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_2d_array< amp::ampf<Precision> >& l);

rmatrixlqunpackq function

/************************************************************************* Partial unpacking of matrix Q from the LQ decomposition of a matrix A Input parameters: A - matrices L and Q in compact form. Output of RMatrixLQ subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Tau - scalar factors which are used to form Q. Output of the RMatrixLQ subroutine. QRows - required number of rows in matrix Q. N>=QRows>=0. Output parameters: Q - first QRows rows of matrix Q. Array whose indexes range within [0..QRows-1, 0..N-1]. If QRows=0, the array remains unchanged. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlqunpackq(const ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& tau, int qrows, ap::template_2d_array< amp::ampf<Precision> >& q);

rmatrixqr function

/************************************************************************* QR decomposition of a rectangular matrix of size MxN Input parameters: A - matrix A whose indexes range within [0..M-1, 0..N-1]. M - number of rows in matrix A. N - number of columns in matrix A. Output parameters: A - matrices Q and R in compact form (see below). Tau - array of scalar factors which are used to form matrix Q. Array whose index ranges within [0.. Min(M-1,N-1)]. Matrix A is represented as A = QR, where Q is an orthogonal matrix of size MxM, R - upper triangular (or upper trapezoid) matrix of size M x N. The elements of matrix R are located on and above the main diagonal of matrix A. The elements which are located in Tau array and below the main diagonal of matrix A are used to form matrix Q as follows: Matrix Q is represented as a product of elementary reflections Q = H(0)*H(2)*...*H(k-1), where k = min(m,n), and each H(i) is in the form H(i) = 1 - tau * v * (v^T) where tau is a scalar stored in Tau[I]; v - real vector, so that v(0:i-1) = 0, v(i) = 1, v(i+1:m-1) stored in A(i+1:m-1,i). -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixqr(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_1d_array< amp::ampf<Precision> >& tau);

rmatrixqrunpackq function

/************************************************************************* Partial unpacking of matrix Q from the QR decomposition of a matrix A Input parameters: A - matrices Q and R in compact form. Output of RMatrixQR subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Tau - scalar factors which are used to form Q. Output of the RMatrixQR subroutine. QColumns - required number of columns of matrix Q. M>=QColumns>=0. Output parameters: Q - first QColumns columns of matrix Q. Array whose indexes range within [0..M-1, 0..QColumns-1]. If QColumns=0, the array remains unchanged. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixqrunpackq(const ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, const ap::template_1d_array< amp::ampf<Precision> >& tau, int qcolumns, ap::template_2d_array< amp::ampf<Precision> >& q);

rmatrixqrunpackr function

/************************************************************************* Unpacking of matrix R from the QR decomposition of a matrix A Input parameters: A - matrices Q and R in compact form. Output of RMatrixQR subroutine. M - number of rows in given matrix A. M>=0. N - number of columns in given matrix A. N>=0. Output parameters: R - matrix R, array[0..M-1, 0..N-1]. -- ALGLIB routine -- 17.02.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixqrunpackr(const ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_2d_array< amp::ampf<Precision> >& r);

smatrixtd function

/************************************************************************* Reduction of a symmetric matrix which is given by its higher or lower triangular part to a tridiagonal matrix using orthogonal similarity transformation: Q'*A*Q=T. Input parameters: A - matrix to be transformed array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then matrix A is given by its upper triangle, and the lower triangle is not used and not modified by the algorithm, and vice versa if IsUpper = False. Output parameters: A - matrices T and Q in compact form (see lower) Tau - array of factors which are forming matrices H(i) array with elements [0..N-2]. D - main diagonal of symmetric matrix T. array with elements [0..N-1]. E - secondary diagonal of symmetric matrix T. array with elements [0..N-2]. If IsUpper=True, the matrix Q is represented as a product of elementary reflectors Q = H(n-2) . . . H(2) H(0). Each H(i) has the form H(i) = I - tau * v * v' where tau is a real scalar, and v is a real vector with v(i+1:n-1) = 0, v(i) = 1, v(0:i-1) is stored on exit in A(0:i-1,i+1), and tau in TAU(i). If IsUpper=False, the matrix Q is represented as a product of elementary reflectors Q = H(0) H(2) . . . H(n-2). Each H(i) has the form H(i) = I - tau * v * v' where tau is a real scalar, and v is a real vector with v(0:i) = 0, v(i+1) = 1, v(i+2:n-1) is stored on exit in A(i+2:n-1,i), and tau in TAU(i). The contents of A on exit are illustrated by the following examples with n = 5: if UPLO = 'U': if UPLO = 'L': ( d e v1 v2 v3 ) ( d ) ( d e v2 v3 ) ( e d ) ( d e v3 ) ( v0 e d ) ( d e ) ( v0 v1 e d ) ( d ) ( v0 v1 v2 e d ) where d and e denote diagonal and off-diagonal elements of T, and vi denotes an element of the vector defining H(i). -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University October 31, 1992 *************************************************************************/
template<unsigned int Precision> void smatrixtd(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, ap::template_1d_array< amp::ampf<Precision> >& tau, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_1d_array< amp::ampf<Precision> >& e);

smatrixtdunpackq function

/************************************************************************* Unpacking matrix Q which reduces symmetric matrix to a tridiagonal form. Input parameters: A - the result of a SMatrixTD subroutine N - size of matrix A. IsUpper - storage format (a parameter of SMatrixTD subroutine) Tau - the result of a SMatrixTD subroutine Output parameters: Q - transformation matrix. array with elements [0..N-1, 0..N-1]. -- ALGLIB -- Copyright 2005-2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void smatrixtdunpackq(const ap::template_2d_array< amp::ampf<Precision> >& a, const int& n, const bool& isupper, const ap::template_1d_array< amp::ampf<Precision> >& tau, ap::template_2d_array< amp::ampf<Precision> >& q);

polint unit

Structures

polynomialfitreport

Functions

polynomialbuild
polynomialbuildcheb1
polynomialbuildcheb2
polynomialbuildeqdist
polynomialcalccheb1
polynomialcalccheb2
polynomialcalceqdist
polynomialfit
polynomialfitwc

Examples

polint_cheb1
polint_cheb2
polint_eqdist
polint_fit
polint_gen

polynomialfitreport structure

/************************************************************************* Polynomial fitting report: TaskRCond reciprocal of task's condition number RMSError RMS error AvgError average error AvgRelError average relative error (for non-zero Y[I]) MaxError maximum error *************************************************************************/
template<unsigned int Precision> class polynomialfitreport { public: amp::ampf<Precision> taskrcond; amp::ampf<Precision> rmserror; amp::ampf<Precision> avgerror; amp::ampf<Precision> avgrelerror; amp::ampf<Precision> maxerror; };

polynomialbuild function

/************************************************************************* Lagrange intepolant: generation of the model on the general grid. This function has O(N^2) complexity. INPUT PARAMETERS: X - abscissas, array[0..N-1] Y - function values, array[0..N-1] N - number of points, N>=1 OIYTPUT PARAMETERS P - barycentric model which represents Lagrange interpolant (see ratint unit info and BarycentricCalc() description for more information). -- ALGLIB -- Copyright 02.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialbuild(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, ratint::barycentricinterpolant<Precision>& p);

Examples:   polint_gen  

polynomialbuildcheb1 function

/************************************************************************* Lagrange intepolant on Chebyshev grid (first kind). This function has O(N) complexity. INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] Y - function values at the nodes, array[0..N-1], Y[I] = Y(0.5*(B+A) + 0.5*(B-A)*Cos(PI*(2*i+1)/(2*n))) N - number of points, N>=1 for N=1 a constant model is constructed. OIYTPUT PARAMETERS P - barycentric model which represents Lagrange interpolant (see ratint unit info and BarycentricCalc() description for more information). -- ALGLIB -- Copyright 03.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialbuildcheb1(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, ratint::barycentricinterpolant<Precision>& p);

Examples:   polint_cheb1  

polynomialbuildcheb2 function

/************************************************************************* Lagrange intepolant on Chebyshev grid (second kind). This function has O(N) complexity. INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] Y - function values at the nodes, array[0..N-1], Y[I] = Y(0.5*(B+A) + 0.5*(B-A)*Cos(PI*i/(n-1))) N - number of points, N>=1 for N=1 a constant model is constructed. OIYTPUT PARAMETERS P - barycentric model which represents Lagrange interpolant (see ratint unit info and BarycentricCalc() description for more information). -- ALGLIB -- Copyright 03.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialbuildcheb2(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, ratint::barycentricinterpolant<Precision>& p);

Examples:   polint_cheb2  

polynomialbuildeqdist function

/************************************************************************* Lagrange intepolant: generation of the model on equidistant grid. This function has O(N) complexity. INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] Y - function values at the nodes, array[0..N-1] N - number of points, N>=1 for N=1 a constant model is constructed. OIYTPUT PARAMETERS P - barycentric model which represents Lagrange interpolant (see ratint unit info and BarycentricCalc() description for more information). -- ALGLIB -- Copyright 03.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialbuildeqdist(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, ratint::barycentricinterpolant<Precision>& p);

Examples:   polint_eqdist  

polynomialcalccheb1 function

/************************************************************************* Fast polynomial interpolation function on Chebyshev points (first kind) with O(N) complexity. INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] F - function values, array[0..N-1] N - number of points on Chebyshev grid (first kind), X[i] = 0.5*(B+A) + 0.5*(B-A)*Cos(PI*(2*i+1)/(2*n)) for N=1 a constant model is constructed. T - position where P(x) is calculated RESULT value of the Lagrange interpolant at T IMPORTANT this function provides fast interface which is not overflow-safe nor it is very precise. the best option is to use PolIntBuildCheb1()/BarycentricCalc() subroutines unless you are pretty sure that your data will not result in overflow. -- ALGLIB -- Copyright 02.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> polynomialcalccheb1(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& f, int n, amp::ampf<Precision> t);

polynomialcalccheb2 function

/************************************************************************* Fast polynomial interpolation function on Chebyshev points (second kind) with O(N) complexity. INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] F - function values, array[0..N-1] N - number of points on Chebyshev grid (second kind), X[i] = 0.5*(B+A) + 0.5*(B-A)*Cos(PI*i/(n-1)) for N=1 a constant model is constructed. T - position where P(x) is calculated RESULT value of the Lagrange interpolant at T IMPORTANT this function provides fast interface which is not overflow-safe nor it is very precise. the best option is to use PolIntBuildCheb2()/BarycentricCalc() subroutines unless you are pretty sure that your data will not result in overflow. -- ALGLIB -- Copyright 02.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> polynomialcalccheb2(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& f, int n, amp::ampf<Precision> t);

polynomialcalceqdist function

/************************************************************************* Fast equidistant polynomial interpolation function with O(N) complexity INPUT PARAMETERS: A - left boundary of [A,B] B - right boundary of [A,B] F - function values, array[0..N-1] N - number of points on equidistant grid, N>=1 for N=1 a constant model is constructed. T - position where P(x) is calculated RESULT value of the Lagrange interpolant at T IMPORTANT this function provides fast interface which is not overflow-safe nor it is very precise. the best option is to use PolynomialBuildEqDist()/BarycentricCalc() subroutines unless you are pretty sure that your data will not result in overflow. -- ALGLIB -- Copyright 02.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> polynomialcalceqdist(amp::ampf<Precision> a, amp::ampf<Precision> b, const ap::template_1d_array< amp::ampf<Precision> >& f, int n, amp::ampf<Precision> t);

polynomialfit function

/************************************************************************* Least squares fitting by polynomial. This subroutine is "lightweight" alternative for more complex and feature- rich PolynomialFitWC(). See PolynomialFitWC() for more information about subroutine parameters (we don't duplicate it here because of length) -- ALGLIB PROJECT -- Copyright 12.10.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialfit(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, int m, int& info, ratint::barycentricinterpolant<Precision>& p, polynomialfitreport<Precision>& rep);

Examples:   polint_fit  

polynomialfitwc function

/************************************************************************* Weighted fitting by Chebyshev polynomial in barycentric form, with constraints on function values or first derivatives. Small regularizing term is used when solving constrained tasks (to improve stability). Task is linear, so linear least squares solver is used. Complexity of this computational scheme is O(N*M^2), mostly dominated by least squares solver SEE ALSO: PolynomialFit() INPUT PARAMETERS: X - points, array[0..N-1]. Y - function values, array[0..N-1]. W - weights, array[0..N-1] Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. Fill it by 1's if you don't want to solve weighted task. N - number of points, N>0. XC - points where polynomial values/derivatives are constrained, array[0..K-1]. YC - values of constraints, array[0..K-1] DC - array[0..K-1], types of constraints: * DC[i]=0 means that P(XC[i])=YC[i] * DC[i]=1 means that P'(XC[i])=YC[i] SEE BELOW FOR IMPORTANT INFORMATION ON CONSTRAINTS K - number of constraints, 0<=K<M. K=0 means no constraints (XC/YC/DC are not used in such cases) M - number of basis functions (= polynomial_degree + 1), M>=1 OUTPUT PARAMETERS: Info- same format as in LSFitLinearW() subroutine: * Info>0 task is solved * Info<=0 an error occured: -4 means inconvergence of internal SVD -3 means inconsistent constraints -1 means another errors in parameters passed (N<=0, for example) P - interpolant in barycentric form. Rep - report, same format as in LSFitLinearW() subroutine. Following fields are set: * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED IMPORTANT: this subroitine doesn't calculate task's condition number for K<>0. SETTING CONSTRAINTS - DANGERS AND OPPORTUNITIES: Setting constraints can lead to undesired results, like ill-conditioned behavior, or inconsistency being detected. From the other side, it allows us to improve quality of the fit. Here we summarize our experience with constrained regression splines: * even simple constraints can be inconsistent, see Wikipedia article on this subject: http://en.wikipedia.org/wiki/Birkhoff_interpolation * the greater is M (given fixed constraints), the more chances that constraints will be consistent * in the general case, consistency of constraints is NOT GUARANTEED. * in the one special cases, however, we can guarantee consistency. This case is: M>1 and constraints on the function values (NOT DERIVATIVES) Our final recommendation is to use constraints WHEN AND ONLY when you can't solve your task without them. Anything beyond special cases given above is not guaranteed and may result in inconsistency. -- ALGLIB PROJECT -- Copyright 10.12.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void polynomialfitwc(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, const ap::template_1d_array< amp::ampf<Precision> >& w, int n, ap::template_1d_array< amp::ampf<Precision> > xc, ap::template_1d_array< amp::ampf<Precision> > yc, const ap::template_1d_array< int >& dc, int k, int m, int& info, ratint::barycentricinterpolant<Precision>& p, polynomialfitreport<Precision>& rep);

Examples:   polint_fit  

polint_cheb1 example

    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    ratint::barycentricinterpolant<Precision> p;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;
    amp::ampf<Precision> d2v;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Demonstration
    //
    printf("POLYNOMIAL INTERPOLATION\n\n");
    printf("F(x)=sin(x), [0, pi]\n");
    printf("Second degree polynomial is used\n\n");
    
    //
    // Create polynomial interpolant
    //
    n = 3;
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        y(i) = amp::sin<Precision>(amp::ampf<Precision>("0.5")*amp::pi<Precision>()*(amp::ampf<Precision>("1.0")+amp::cos<Precision>(amp::pi<Precision>()*(2*i+1)/(2*n))));
    }
    polint::polynomialbuildcheb1<Precision>(amp::ampf<Precision>(0), amp::pi<Precision>(), y, n, p);
    
    //
    // Output results
    //
    ratint::barycentricdiff2<Precision>(p, amp::ampf<Precision>(0), v, dv, d2v);
    printf("                 P(x)    F(x) \n");
    printf("function       %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(ratint::barycentriccalc<Precision>(p, amp::ampf<Precision>(0))).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("d/dx(0)        %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(dv).toDouble()),
        double(amp::ampf<Precision>(1).toDouble()));
    printf("d2/dx2(0)      %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(d2v).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("\n\n");

polint_cheb2 example

    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    ratint::barycentricinterpolant<Precision> p;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;
    amp::ampf<Precision> d2v;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Demonstration
    //
    printf("POLYNOMIAL INTERPOLATION\n\n");
    printf("F(x)=sin(x), [0, pi]\n");
    printf("Second degree polynomial is used\n\n");
    
    //
    // Create polynomial interpolant
    //
    n = 3;
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        y(i) = amp::sin<Precision>(amp::ampf<Precision>("0.5")*amp::pi<Precision>()*(amp::ampf<Precision>("1.0")+amp::cos<Precision>(amp::pi<Precision>()*i/(n-1))));
    }
    polint::polynomialbuildcheb2<Precision>(amp::ampf<Precision>(0), amp::pi<Precision>(), y, n, p);
    
    //
    // Output results
    //
    ratint::barycentricdiff2<Precision>(p, amp::ampf<Precision>(0), v, dv, d2v);
    printf("                 P(x)    F(x) \n");
    printf("function       %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(ratint::barycentriccalc<Precision>(p, amp::ampf<Precision>(0))).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("d/dx(0)        %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(dv).toDouble()),
        double(amp::ampf<Precision>(1).toDouble()));
    printf("d2/dx2(0)      %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(d2v).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("\n\n");

polint_eqdist example

    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    ratint::barycentricinterpolant<Precision> p;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;
    amp::ampf<Precision> d2v;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Demonstration
    //
    printf("POLYNOMIAL INTERPOLATION\n\n");
    printf("F(x)=sin(x), [0, pi]\n");
    printf("Second degree polynomial is used\n\n");
    
    //
    // Create polynomial interpolant
    //
    n = 3;
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        y(i) = amp::sin<Precision>(amp::pi<Precision>()*i/(n-1));
    }
    polint::polynomialbuildeqdist<Precision>(amp::ampf<Precision>(0), amp::pi<Precision>(), y, n, p);
    
    //
    // Output results
    //
    ratint::barycentricdiff2<Precision>(p, amp::ampf<Precision>(0), v, dv, d2v);
    printf("                 P(x)    F(x) \n");
    printf("function       %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(ratint::barycentriccalc<Precision>(p, amp::ampf<Precision>(0))).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("d/dx(0)        %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(dv).toDouble()),
        double(amp::ampf<Precision>(1).toDouble()));
    printf("d2/dx2(0)      %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(d2v).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("\n\n");

polint_fit example

    int m;
    int n;
    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_1d_array< amp::ampf<Precision> > w;
    ap::template_1d_array< amp::ampf<Precision> > xc;
    ap::template_1d_array< amp::ampf<Precision> > yc;
    ap::template_1d_array< int > dc;
    polint::polynomialfitreport<Precision> rep;
    int info;
    ratint::barycentricinterpolant<Precision> p;
    int i;
    int j;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;


    printf("\n\nFitting exp(2*x) at [-1,+1] by polinomial\n\n");
    printf("Fit type             rms.err max.err    p(0)   dp(0)\n");
    
    //
    // Prepare points
    //
    m = 5;
    a = -1;
    b = +1;
    n = 1000;
    x.setlength(n);
    y.setlength(n);
    w.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = a+(b-a)*i/(n-1);
        y(i) = amp::exp<Precision>(2*x(i));
        w(i) = amp::ampf<Precision>("1.0");
    }
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5th degree polynomial
    // c) without constraints
    //
    polint::polynomialfit<Precision>(x, y, n, m, info, p, rep);
    ratint::barycentricdiff1<Precision>(p, amp::ampf<Precision>("0.0"), v, dv);
    printf("Unconstrained        %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5th degree polynomial
    // c) constrained: p(0)=1
    //
    xc.setlength(1);
    yc.setlength(1);
    dc.setlength(1);
    xc(0) = 0;
    yc(0) = 1;
    dc(0) = 0;
    polint::polynomialfitwc<Precision>(x, y, w, n, xc, yc, dc, 1, m, info, p, rep);
    ratint::barycentricdiff1<Precision>(p, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained, p(0)=1  %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5th degree polynomial
    // c) constrained: dp(0)=2
    //
    xc.setlength(1);
    yc.setlength(1);
    dc.setlength(1);
    xc(0) = 0;
    yc(0) = 2;
    dc(0) = 1;
    polint::polynomialfitwc<Precision>(x, y, w, n, xc, yc, dc, 1, m, info, p, rep);
    ratint::barycentricdiff1<Precision>(p, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained, dp(0)=2 %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5th degree polynomial
    // c) constrained: p(0)=1, dp(0)=2
    //
    xc.setlength(2);
    yc.setlength(2);
    dc.setlength(2);
    xc(0) = 0;
    yc(0) = 1;
    dc(0) = 0;
    xc(1) = 0;
    yc(1) = 2;
    dc(1) = 1;
    polint::polynomialfitwc<Precision>(x, y, w, n, xc, yc, dc, 2, m, info, p, rep);
    ratint::barycentricdiff1<Precision>(p, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained, both    %7.4lf %7.4lf %7.4lf %7.4lf\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()));
    printf("\n\n");

polint_gen example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    ratint::barycentricinterpolant<Precision> p;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;
    amp::ampf<Precision> d2v;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Demonstration
    //
    printf("POLYNOMIAL INTERPOLATION\n\n");
    printf("F(x)=sin(x), [0, pi]\n");
    printf("Second degree polynomial is used\n\n");
    
    //
    // Create polynomial interpolant
    //
    n = 3;
    x.setlength(n);
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
    }
    polint::polynomialbuild<Precision>(x, y, n, p);
    
    //
    // Output results
    //
    ratint::barycentricdiff2<Precision>(p, amp::ampf<Precision>(0), v, dv, d2v);
    printf("                 P(x)    F(x) \n");
    printf("function       %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(ratint::barycentriccalc<Precision>(p, amp::ampf<Precision>(0))).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("d/dx(0)        %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(dv).toDouble()),
        double(amp::ampf<Precision>(1).toDouble()));
    printf("d2/dx2(0)      %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(d2v).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("\n\n");

pspline unit

Structures

pspline2interpolant
pspline3interpolant

Functions

pspline2arclength
pspline2build
pspline2buildperiodic
pspline2calc
pspline2diff
pspline2diff2
pspline2parametervalues
pspline2tangent
pspline3arclength
pspline3build
pspline3buildperiodic
pspline3calc
pspline3diff
pspline3diff2
pspline3parametervalues
pspline3tangent

pspline2interpolant structure

/************************************************************************* Parametric spline inteprolant: 2-dimensional curve. You should not try to access its members directly - use PSpline2XXXXXXXX() functions instead. *************************************************************************/
template<unsigned int Precision> class pspline2interpolant { public: int n; bool periodic; ap::template_1d_array< amp::ampf<Precision> > p; spline1d::spline1dinterpolant<Precision> x; spline1d::spline1dinterpolant<Precision> y; };

pspline3interpolant structure

/************************************************************************* Parametric spline inteprolant: 3-dimensional curve. You should not try to access its members directly - use PSpline3XXXXXXXX() functions instead. *************************************************************************/
template<unsigned int Precision> class pspline3interpolant { public: int n; bool periodic; ap::template_1d_array< amp::ampf<Precision> > p; spline1d::spline1dinterpolant<Precision> x; spline1d::spline1dinterpolant<Precision> y; spline1d::spline1dinterpolant<Precision> z; };

pspline2arclength function

/************************************************************************* This function calculates arc length, i.e. length of curve between t=a and t=b. INPUT PARAMETERS: P - parametric spline interpolant A,B - parameter values corresponding to arc ends: * B>A will result in positive length returned * B<A will result in negative length returned RESULT: length of arc starting at T=A and ending at T=B. -- ALGLIB PROJECT -- Copyright 30.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> pspline2arclength(const pspline2interpolant<Precision>& p, amp::ampf<Precision> a, amp::ampf<Precision> b);

pspline2build function

/************************************************************************* This function builds non-periodic 2-dimensional parametric spline which starts at (X[0],Y[0]) and ends at (X[N-1],Y[N-1]). INPUT PARAMETERS: XY - points, array[0..N-1,0..1]. XY[I,0:1] corresponds to the Ith point. Order of points is important! N - points count, N>=5 for Akima splines, N>=2 for other types of splines. ST - spline type: * 0 Akima spline * 1 parabolically terminated Catmull-Rom spline (Tension=0) * 2 parabolically terminated cubic spline PT - parameterization type: * 0 uniform * 1 chord length * 2 centripetal OUTPUT PARAMETERS: P - parametric spline interpolant NOTES: * this function assumes that there all consequent points are distinct. I.e. (x0,y0)<>(x1,y1), (x1,y1)<>(x2,y2), (x2,y2)<>(x3,y3) and so on. However, non-consequent points may coincide, i.e. we can have (x0,y0)= =(x2,y2). -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2build(ap::template_2d_array< amp::ampf<Precision> > xy, int n, int st, int pt, pspline2interpolant<Precision>& p);

pspline2buildperiodic function

/************************************************************************* This function builds periodic 2-dimensional parametric spline which starts at (X[0],Y[0]), goes through all points to (X[N-1],Y[N-1]) and then back to (X[0],Y[0]). INPUT PARAMETERS: XY - points, array[0..N-1,0..1]. XY[I,0:1] corresponds to the Ith point. XY[N-1,0:1] must be different from XY[0,0:1]. Order of points is important! N - points count, N>=3 for other types of splines. ST - spline type: * 1 Catmull-Rom spline (Tension=0) with cyclic boundary conditions * 2 cubic spline with cyclic boundary conditions PT - parameterization type: * 0 uniform * 1 chord length * 2 centripetal OUTPUT PARAMETERS: P - parametric spline interpolant NOTES: * this function assumes that there all consequent points are distinct. I.e. (x0,y0)<>(x1,y1), (x1,y1)<>(x2,y2), (x2,y2)<>(x3,y3) and so on. However, non-consequent points may coincide, i.e. we can have (x0,y0)= =(x2,y2). * last point of sequence is NOT equal to the first point. You shouldn't make curve "explicitly periodic" by making them equal. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2buildperiodic(ap::template_2d_array< amp::ampf<Precision> > xy, int n, int st, int pt, pspline2interpolant<Precision>& p);

pspline2calc function

/************************************************************************* This function calculates the value of the parametric spline for a given value of parameter T INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-position Y - Y-position -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2calc(const pspline2interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& y);

pspline2diff function

/************************************************************************* This function calculates derivative, i.e. it returns (dX/dT,dY/dT). INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-value DX - X-derivative Y - Y-value DY - Y-derivative -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2diff(const pspline2interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& dx, amp::ampf<Precision>& y, amp::ampf<Precision>& dy);

pspline2diff2 function

/************************************************************************* This function calculates first and second derivative with respect to T. INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-value DX - derivative D2X - second derivative Y - Y-value DY - derivative D2Y - second derivative -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2diff2(const pspline2interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& dx, amp::ampf<Precision>& d2x, amp::ampf<Precision>& y, amp::ampf<Precision>& dy, amp::ampf<Precision>& d2y);

pspline2parametervalues function

/************************************************************************* This function returns vector of parameter values correspoding to points. I.e. for P created from (X[0],Y[0])...(X[N-1],Y[N-1]) and U=TValues(P) we have (X[0],Y[0]) = PSpline2Calc(P,U[0]), (X[1],Y[1]) = PSpline2Calc(P,U[1]), (X[2],Y[2]) = PSpline2Calc(P,U[2]), ... INPUT PARAMETERS: P - parametric spline interpolant OUTPUT PARAMETERS: N - array size T - array[0..N-1] NOTES: * for non-periodic splines U[0]=0, U[0]<U[1]<...<U[N-1], U[N-1]=1 * for periodic splines U[0]=0, U[0]<U[1]<...<U[N-1], U[N-1]<1 -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2parametervalues(const pspline2interpolant<Precision>& p, int& n, ap::template_1d_array< amp::ampf<Precision> >& t);

pspline2tangent function

/************************************************************************* This function calculates tangent vector for a given value of parameter T INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-component of tangent vector (normalized) Y - Y-component of tangent vector (normalized) NOTE: X^2+Y^2 is either 1 (for non-zero tangent vector) or 0. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline2tangent(const pspline2interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& y);

pspline3arclength function

/************************************************************************* This function calculates arc length, i.e. length of curve between t=a and t=b. INPUT PARAMETERS: P - parametric spline interpolant A,B - parameter values corresponding to arc ends: * B>A will result in positive length returned * B<A will result in negative length returned RESULT: length of arc starting at T=A and ending at T=B. -- ALGLIB PROJECT -- Copyright 30.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> pspline3arclength(const pspline3interpolant<Precision>& p, amp::ampf<Precision> a, amp::ampf<Precision> b);

pspline3build function

/************************************************************************* This function builds non-periodic 3-dimensional parametric spline which starts at (X[0],Y[0],Z[0]) and ends at (X[N-1],Y[N-1],Z[N-1]). Same as PSpline2Build() function, but for 3D, so we won't duplicate its description here. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3build(ap::template_2d_array< amp::ampf<Precision> > xy, int n, int st, int pt, pspline3interpolant<Precision>& p);

pspline3buildperiodic function

/************************************************************************* This function builds periodic 3-dimensional parametric spline which starts at (X[0],Y[0],Z[0]), goes through all points to (X[N-1],Y[N-1],Z[N-1]) and then back to (X[0],Y[0],Z[0]). Same as PSpline2Build() function, but for 3D, so we won't duplicate its description here. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3buildperiodic(ap::template_2d_array< amp::ampf<Precision> > xy, int n, int st, int pt, pspline3interpolant<Precision>& p);

pspline3calc function

/************************************************************************* This function calculates the value of the parametric spline for a given value of parameter T. INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-position Y - Y-position Z - Z-position -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3calc(const pspline3interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& y, amp::ampf<Precision>& z);

pspline3diff function

/************************************************************************* This function calculates derivative, i.e. it returns (dX/dT,dY/dT,dZ/dT). INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-value DX - X-derivative Y - Y-value DY - Y-derivative Z - Z-value DZ - Z-derivative -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3diff(const pspline3interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& dx, amp::ampf<Precision>& y, amp::ampf<Precision>& dy, amp::ampf<Precision>& z, amp::ampf<Precision>& dz);

pspline3diff2 function

/************************************************************************* This function calculates first and second derivative with respect to T. INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-value DX - derivative D2X - second derivative Y - Y-value DY - derivative D2Y - second derivative Z - Z-value DZ - derivative D2Z - second derivative -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3diff2(const pspline3interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& dx, amp::ampf<Precision>& d2x, amp::ampf<Precision>& y, amp::ampf<Precision>& dy, amp::ampf<Precision>& d2y, amp::ampf<Precision>& z, amp::ampf<Precision>& dz, amp::ampf<Precision>& d2z);

pspline3parametervalues function

/************************************************************************* This function returns vector of parameter values correspoding to points. Same as PSpline2ParameterValues(), but for 3D. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3parametervalues(const pspline3interpolant<Precision>& p, int& n, ap::template_1d_array< amp::ampf<Precision> >& t);

pspline3tangent function

/************************************************************************* This function calculates tangent vector for a given value of parameter T INPUT PARAMETERS: P - parametric spline interpolant T - point: * T in [0,1] corresponds to interval spanned by points * for non-periodic splines T<0 (or T>1) correspond to parts of the curve before the first (after the last) point * for periodic splines T<0 (or T>1) are projected into [0,1] by making T=T-floor(T). OUTPUT PARAMETERS: X - X-component of tangent vector (normalized) Y - Y-component of tangent vector (normalized) Z - Z-component of tangent vector (normalized) NOTE: X^2+Y^2+Z^2 is either 1 (for non-zero tangent vector) or 0. -- ALGLIB PROJECT -- Copyright 28.05.2010 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void pspline3tangent(const pspline3interpolant<Precision>& p, amp::ampf<Precision> t, amp::ampf<Precision>& x, amp::ampf<Precision>& y, amp::ampf<Precision>& z);

ratint unit

Structures

barycentricfitreport
barycentricinterpolant

Functions

barycentricbuildfloaterhormann
barycentricbuildxyw
barycentriccalc
barycentriccopy
barycentricdiff1
barycentricdiff2
barycentricfitfloaterhormann
barycentricfitfloaterhormannwc
barycentriclintransx
barycentriclintransy
barycentricserialize
barycentricunpack
barycentricunserialize

Examples

ratint_fit

barycentricfitreport structure

/************************************************************************* Barycentric fitting report: TaskRCond reciprocal of task's condition number RMSError RMS error AvgError average error AvgRelError average relative error (for non-zero Y[I]) MaxError maximum error *************************************************************************/
template<unsigned int Precision> class barycentricfitreport { public: amp::ampf<Precision> taskrcond; int dbest; amp::ampf<Precision> rmserror; amp::ampf<Precision> avgerror; amp::ampf<Precision> avgrelerror; amp::ampf<Precision> maxerror; };

barycentricinterpolant structure

/************************************************************************* Barycentric interpolant. *************************************************************************/
template<unsigned int Precision> class barycentricinterpolant { public: int n; amp::ampf<Precision> sy; ap::template_1d_array< amp::ampf<Precision> > x; ap::template_1d_array< amp::ampf<Precision> > y; ap::template_1d_array< amp::ampf<Precision> > w; };

barycentricbuildfloaterhormann function

/************************************************************************* Rational interpolant without poles The subroutine constructs the rational interpolating function without real poles (see 'Barycentric rational interpolation with no poles and high rates of approximation', Michael S. Floater. and Kai Hormann, for more information on this subject). Input parameters: X - interpolation nodes, array[0..N-1]. Y - function values, array[0..N-1]. N - number of nodes, N>0. D - order of the interpolation scheme, 0 <= D <= N-1. D<0 will cause an error. D>=N it will be replaced with D=N-1. if you don't know what D to choose, use small value about 3-5. Output parameters: B - barycentric interpolant. Note: this algorithm always succeeds and calculates the weights with close to machine precision. -- ALGLIB PROJECT -- Copyright 17.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricbuildfloaterhormann(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, int d, barycentricinterpolant<Precision>& b);

barycentricbuildxyw function

/************************************************************************* Rational interpolant from X/Y/W arrays F(t) = SUM(i=0,n-1,w[i]*f[i]/(t-x[i])) / SUM(i=0,n-1,w[i]/(t-x[i])) INPUT PARAMETERS: X - interpolation nodes, array[0..N-1] F - function values, array[0..N-1] W - barycentric weights, array[0..N-1] N - nodes count, N>0 OUTPUT PARAMETERS: B - barycentric interpolant built from (X, Y, W) -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricbuildxyw(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, int n, barycentricinterpolant<Precision>& b);

barycentriccalc function

/************************************************************************* Rational interpolation using barycentric formula F(t) = SUM(i=0,n-1,w[i]*f[i]/(t-x[i])) / SUM(i=0,n-1,w[i]/(t-x[i])) Input parameters: B - barycentric interpolant built with one of model building subroutines. T - interpolation point Result: barycentric interpolant F(t) -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> barycentriccalc(const barycentricinterpolant<Precision>& b, amp::ampf<Precision> t);

barycentriccopy function

/************************************************************************* Copying of the barycentric interpolant INPUT PARAMETERS: B - barycentric interpolant OUTPUT PARAMETERS: B2 - copy(B1) -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentriccopy(const barycentricinterpolant<Precision>& b, barycentricinterpolant<Precision>& b2);

barycentricdiff1 function

/************************************************************************* Differentiation of barycentric interpolant: first derivative. Algorithm used in this subroutine is very robust and should not fail until provided with values too close to MaxRealNumber (usually MaxRealNumber/N or greater will overflow). INPUT PARAMETERS: B - barycentric interpolant built with one of model building subroutines. T - interpolation point OUTPUT PARAMETERS: F - barycentric interpolant at T DF - first derivative NOTE -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricdiff1(const barycentricinterpolant<Precision>& b, amp::ampf<Precision> t, amp::ampf<Precision>& f, amp::ampf<Precision>& df);

barycentricdiff2 function

/************************************************************************* Differentiation of barycentric interpolant: first/second derivatives. INPUT PARAMETERS: B - barycentric interpolant built with one of model building subroutines. T - interpolation point OUTPUT PARAMETERS: F - barycentric interpolant at T DF - first derivative D2F - second derivative NOTE: this algorithm may fail due to overflow/underflor if used on data whose values are close to MaxRealNumber or MinRealNumber. Use more robust BarycentricDiff1() subroutine in such cases. -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricdiff2(const barycentricinterpolant<Precision>& b, amp::ampf<Precision> t, amp::ampf<Precision>& f, amp::ampf<Precision>& df, amp::ampf<Precision>& d2f);

barycentricfitfloaterhormann function

/************************************************************************* Rational least squares fitting, without weights and constraints. See BarycentricFitFloaterHormannWC() for more information. -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricfitfloaterhormann(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, int m, int& info, barycentricinterpolant<Precision>& b, barycentricfitreport<Precision>& rep);

Examples:   ratint_fit  

barycentricfitfloaterhormannwc function

/************************************************************************* Weghted rational least squares fitting using Floater-Hormann rational functions with optimal D chosen from [0,9], with constraints and individual weights. Equidistant grid with M node on [min(x),max(x)] is used to build basis functions. Different values of D are tried, optimal D (least WEIGHTED root mean square error) is chosen. Task is linear, so linear least squares solver is used. Complexity of this computational scheme is O(N*M^2) (mostly dominated by the least squares solver). SEE ALSO * BarycentricFitFloaterHormann(), "lightweight" fitting without invididual weights and constraints. INPUT PARAMETERS: X - points, array[0..N-1]. Y - function values, array[0..N-1]. W - weights, array[0..N-1] Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. Fill it by 1's if you don't want to solve weighted task. N - number of points, N>0. XC - points where function values/derivatives are constrained, array[0..K-1]. YC - values of constraints, array[0..K-1] DC - array[0..K-1], types of constraints: * DC[i]=0 means that S(XC[i])=YC[i] * DC[i]=1 means that S'(XC[i])=YC[i] SEE BELOW FOR IMPORTANT INFORMATION ON CONSTRAINTS K - number of constraints, 0<=K<M. K=0 means no constraints (XC/YC/DC are not used in such cases) M - number of basis functions ( = number_of_nodes), M>=2. OUTPUT PARAMETERS: Info- same format as in LSFitLinearWC() subroutine. * Info>0 task is solved * Info<=0 an error occured: -4 means inconvergence of internal SVD -3 means inconsistent constraints -1 means another errors in parameters passed (N<=0, for example) B - barycentric interpolant. Rep - report, same format as in LSFitLinearWC() subroutine. Following fields are set: * DBest best value of the D parameter * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED IMPORTANT: this subroitine doesn't calculate task's condition number for K<>0. SETTING CONSTRAINTS - DANGERS AND OPPORTUNITIES: Setting constraints can lead to undesired results, like ill-conditioned behavior, or inconsistency being detected. From the other side, it allows us to improve quality of the fit. Here we summarize our experience with constrained barycentric interpolants: * excessive constraints can be inconsistent. Floater-Hormann basis functions aren't as flexible as splines (although they are very smooth). * the more evenly constraints are spread across [min(x),max(x)], the more chances that they will be consistent * the greater is M (given fixed constraints), the more chances that constraints will be consistent * in the general case, consistency of constraints IS NOT GUARANTEED. * in the several special cases, however, we CAN guarantee consistency. * one of this cases is constraints on the function VALUES at the interval boundaries. Note that consustency of the constraints on the function DERIVATIVES is NOT guaranteed (you can use in such cases cubic splines which are more flexible). * another special case is ONE constraint on the function value (OR, but not AND, derivative) anywhere in the interval Our final recommendation is to use constraints WHEN AND ONLY WHEN you can't solve your task without them. Anything beyond special cases given above is not guaranteed and may result in inconsistency. -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricfitfloaterhormannwc(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, int n, const ap::template_1d_array< amp::ampf<Precision> >& xc, const ap::template_1d_array< amp::ampf<Precision> >& yc, const ap::template_1d_array< int >& dc, int k, int m, int& info, barycentricinterpolant<Precision>& b, barycentricfitreport<Precision>& rep);

Examples:   ratint_fit  

barycentriclintransx function

/************************************************************************* This subroutine performs linear transformation of the argument. INPUT PARAMETERS: B - rational interpolant in barycentric form CA, CB - transformation coefficients: x = CA*t + CB OUTPUT PARAMETERS: B - transformed interpolant with X replaced by T -- ALGLIB PROJECT -- Copyright 19.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentriclintransx(barycentricinterpolant<Precision>& b, amp::ampf<Precision> ca, amp::ampf<Precision> cb);

barycentriclintransy function

/************************************************************************* This subroutine performs linear transformation of the barycentric interpolant. INPUT PARAMETERS: B - rational interpolant in barycentric form CA, CB - transformation coefficients: B2(x) = CA*B(x) + CB OUTPUT PARAMETERS: B - transformed interpolant -- ALGLIB PROJECT -- Copyright 19.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentriclintransy(barycentricinterpolant<Precision>& b, amp::ampf<Precision> ca, amp::ampf<Precision> cb);

barycentricserialize function

/************************************************************************* Serialization of the barycentric interpolant INPUT PARAMETERS: B - barycentric interpolant OUTPUT PARAMETERS: RA - array of real numbers which contains interpolant, array[0..RLen-1] RLen - RA lenght -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricserialize(const barycentricinterpolant<Precision>& b, ap::template_1d_array< amp::ampf<Precision> >& ra, int& ralen);

barycentricunpack function

/************************************************************************* Extracts X/Y/W arrays from rational interpolant INPUT PARAMETERS: B - barycentric interpolant OUTPUT PARAMETERS: N - nodes count, N>0 X - interpolation nodes, array[0..N-1] F - function values, array[0..N-1] W - barycentric weights, array[0..N-1] -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricunpack(const barycentricinterpolant<Precision>& b, int& n, ap::template_1d_array< amp::ampf<Precision> >& x, ap::template_1d_array< amp::ampf<Precision> >& y, ap::template_1d_array< amp::ampf<Precision> >& w);

barycentricunserialize function

/************************************************************************* Unserialization of the barycentric interpolant INPUT PARAMETERS: RA - array of real numbers which contains interpolant, OUTPUT PARAMETERS: B - barycentric interpolant -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void barycentricunserialize(const ap::template_1d_array< amp::ampf<Precision> >& ra, barycentricinterpolant<Precision>& b);

ratint_fit example

    int m;
    int n;
    int d;
    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_1d_array< amp::ampf<Precision> > w;
    ap::template_1d_array< amp::ampf<Precision> > xc;
    ap::template_1d_array< amp::ampf<Precision> > yc;
    ap::template_1d_array< int > dc;
    ratint::barycentricfitreport<Precision> rep;
    int info;
    ratint::barycentricinterpolant<Precision> r;
    int i;
    int j;
    amp::ampf<Precision> a;
    amp::ampf<Precision> b;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;


    printf("\n\nFitting exp(2*x) at [-1,+1] by:\n1. constrained/unconstrained Floater-Hormann functions\n");
    printf("\n");
    printf("Fit type                rms.err max.err    p(0)   dp(0)  DBest\n");
    
    //
    // Prepare points
    //
    m = 5;
    a = -1;
    b = +1;
    n = 10000;
    x.setlength(n);
    y.setlength(n);
    w.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = a+(b-a)*i/(n-1);
        y(i) = amp::exp<Precision>(2*x(i));
        w(i) = amp::ampf<Precision>("1.0");
    }
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5 Floater-Hormann functions
    // c) without constraints
    //
    ratint::barycentricfitfloaterhormann<Precision>(x, y, n, m, info, r, rep);
    ratint::barycentricdiff1<Precision>(r, amp::ampf<Precision>("0.0"), v, dv);
    printf("Unconstrained FH        %7.4lf %7.4lf %7.4lf %7.4lf      %0ld\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()),
        long(rep.dbest));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5 Floater-Hormann functions
    // c) constrained: p(0)=1
    //
    xc.setlength(1);
    yc.setlength(1);
    dc.setlength(1);
    xc(0) = 0;
    yc(0) = 1;
    dc(0) = 0;
    ratint::barycentricfitfloaterhormannwc<Precision>(x, y, w, n, xc, yc, dc, 1, m, info, r, rep);
    ratint::barycentricdiff1<Precision>(r, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained FH, p(0)=1  %7.4lf %7.4lf %7.4lf %7.4lf      %0ld\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()),
        long(rep.dbest));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5 Floater-Hormann functions
    // c) constrained: dp(0)=2
    //
    xc.setlength(1);
    yc.setlength(1);
    dc.setlength(1);
    xc(0) = 0;
    yc(0) = 2;
    dc(0) = 1;
    ratint::barycentricfitfloaterhormannwc<Precision>(x, y, w, n, xc, yc, dc, 1, m, info, r, rep);
    ratint::barycentricdiff1<Precision>(r, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained FH, dp(0)=2 %7.4lf %7.4lf %7.4lf %7.4lf      %0ld\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()),
        long(rep.dbest));
    
    //
    // Fitting:
    // a) f(x)=exp(2*x) at [-1,+1]
    // b) by 5 Floater-Hormann functions
    // c) constrained: p(0)=1, dp(0)=2
    //
    xc.setlength(2);
    yc.setlength(2);
    dc.setlength(2);
    xc(0) = 0;
    yc(0) = 1;
    dc(0) = 0;
    xc(1) = 0;
    yc(1) = 2;
    dc(1) = 1;
    ratint::barycentricfitfloaterhormannwc<Precision>(x, y, w, n, xc, yc, dc, 2, m, info, r, rep);
    ratint::barycentricdiff1<Precision>(r, amp::ampf<Precision>("0.0"), v, dv);
    printf("Constrained FH, both    %7.4lf %7.4lf %7.4lf %7.4lf      %0ld\n",
        double(amp::ampf<Precision>(rep.rmserror).toDouble()),
        double(amp::ampf<Precision>(rep.maxerror).toDouble()),
        double(amp::ampf<Precision>(v).toDouble()),
        double(amp::ampf<Precision>(dv).toDouble()),
        long(rep.dbest));
    printf("\n\n");

rcond unit

Functions

cmatrixlurcond1
cmatrixlurcondinf
cmatrixrcond1
cmatrixrcondinf
cmatrixtrrcond1
cmatrixtrrcondinf
hpdmatrixcholeskyrcond
hpdmatrixrcond
rcondthreshold
rmatrixlurcond1
rmatrixlurcondinf
rmatrixrcond1
rmatrixrcondinf
rmatrixtrrcond1
rmatrixtrrcondinf
spdmatrixcholeskyrcond
spdmatrixrcond

Examples

rcond_1

cmatrixlurcond1 function

/************************************************************************* Estimate of the condition number of a matrix given by its LU decomposition (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: LUA - LU decomposition of a matrix in compact form. Output of the CMatrixLU subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixlurcond1(const ap::template_2d_array< amp::campf<Precision> >& lua, int n);

Examples:   rcond_1  

cmatrixlurcondinf function

/************************************************************************* Estimate of the condition number of a matrix given by its LU decomposition (infinity norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: LUA - LU decomposition of a matrix in compact form. Output of the CMatrixLU subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixlurcondinf(const ap::template_2d_array< amp::campf<Precision> >& lua, int n);

Examples:   rcond_1  

cmatrixrcond1 function

/************************************************************************* Estimate of a matrix condition number (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixrcond1(ap::template_2d_array< amp::campf<Precision> > a, int n);

Examples:   rcond_1  

cmatrixrcondinf function

/************************************************************************* Estimate of a matrix condition number (infinity-norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixrcondinf(ap::template_2d_array< amp::campf<Precision> > a, int n);

Examples:   rcond_1  

cmatrixtrrcond1 function

/************************************************************************* Triangular matrix: estimate of a condition number (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array[0..N-1, 0..N-1]. N - size of A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixtrrcond1(const ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, bool isunit);

cmatrixtrrcondinf function

/************************************************************************* Triangular matrix: estimate of a matrix condition number (infinity-norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> cmatrixtrrcondinf(const ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper, bool isunit);

hpdmatrixcholeskyrcond function

/************************************************************************* Condition number estimate of a Hermitian positive definite matrix given by Cholesky decomposition. The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: CD - Cholesky decomposition of matrix A, output of SMatrixCholesky subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> hpdmatrixcholeskyrcond(const ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper);

Examples:   rcond_1  

hpdmatrixrcond function

/************************************************************************* Condition number estimate of a Hermitian positive definite matrix. The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm of condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: A - Hermitian positive definite matrix which is given by its upper or lower triangle depending on the value of IsUpper. Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. Result: 1/LowerBound(cond(A)), if matrix A is positive definite, -1, if matrix A is not positive definite, and its condition number could not be found by this algorithm. NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> hpdmatrixrcond(ap::template_2d_array< amp::campf<Precision> > a, int n, bool isupper);

Examples:   rcond_1  

rcondthreshold function

/************************************************************************* Threshold for rcond: matrices with condition number beyond this threshold are considered singular. Threshold must be far enough from underflow, at least Sqr(Threshold) must be greater than underflow. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rcondthreshold();

rmatrixlurcond1 function

/************************************************************************* Estimate of the condition number of a matrix given by its LU decomposition (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: LUA - LU decomposition of a matrix in compact form. Output of the RMatrixLU subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixlurcond1(const ap::template_2d_array< amp::ampf<Precision> >& lua, int n);

Examples:   rcond_1  

rmatrixlurcondinf function

/************************************************************************* Estimate of the condition number of a matrix given by its LU decomposition (infinity norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: LUA - LU decomposition of a matrix in compact form. Output of the RMatrixLU subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixlurcondinf(const ap::template_2d_array< amp::ampf<Precision> >& lua, int n);

Examples:   rcond_1  

rmatrixrcond1 function

/************************************************************************* Estimate of a matrix condition number (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixrcond1(ap::template_2d_array< amp::ampf<Precision> > a, int n);

Examples:   rcond_1  

rmatrixrcondinf function

/************************************************************************* Estimate of a matrix condition number (infinity-norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixrcondinf(ap::template_2d_array< amp::ampf<Precision> > a, int n);

Examples:   rcond_1  

rmatrixtrrcond1 function

/************************************************************************* Triangular matrix: estimate of a condition number (1-norm) The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array[0..N-1, 0..N-1]. N - size of A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixtrrcond1(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, bool isunit);

rmatrixtrrcondinf function

/************************************************************************* Triangular matrix: estimate of a matrix condition number (infinity-norm). The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). Input parameters: A - matrix. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - True, if the matrix is upper triangular. IsUnit - True, if the matrix has a unit diagonal. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> rmatrixtrrcondinf(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper, bool isunit);

spdmatrixcholeskyrcond function

/************************************************************************* Condition number estimate of a symmetric positive definite matrix given by Cholesky decomposition. The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: CD - Cholesky decomposition of matrix A, output of SMatrixCholesky subroutine. N - size of matrix A. Result: 1/LowerBound(cond(A)) NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spdmatrixcholeskyrcond(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper);

Examples:   rcond_1  

spdmatrixrcond function

/************************************************************************* Condition number estimate of a symmetric positive definite matrix. The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm of condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: A - symmetric positive definite matrix which is given by its upper or lower triangle depending on the value of IsUpper. Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. Result: 1/LowerBound(cond(A)), if matrix A is positive definite, -1, if matrix A is not positive definite, and its condition number could not be found by this algorithm. NOTE: if k(A) is very large, then matrix is assumed degenerate, k(A)=INF, 0.0 is returned in such cases. *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spdmatrixrcond(ap::template_2d_array< amp::ampf<Precision> > a, int n, bool isupper);

Examples:   rcond_1  

rcond_1 example

    int n;
    int i;
    int j;
    amp::ampf<Precision> c1;
    amp::ampf<Precision> x;
    ap::template_2d_array< amp::ampf<Precision> > a;


    printf("                 CONDITION NUMBERS\n");
    printf("OF VANDERMONDE AND CHEBYSHEV INTERPOLATION MATRICES\n\n");
    printf("    VANDERMONDE   CHEBYSHEV\n");
    printf("  N      1-norm      1-norm\n");
    for(n=2; n<=14; n++)
    {
        a.setlength(n, n);
        printf("%3ld",
            long(n));
        
        //
        // Vandermone matrix
        //
        for(i=0; i<=n-1; i++)
        {
            x = amp::ampf<Precision>(2*i)/(amp::ampf<Precision>(n-1))-1;
            a(i,0) = 1;
            for(j=1; j<=n-1; j++)
            {
                a(i,j) = a(i,j-1)*x;
            }
        }
        c1 = 1/rcond::rmatrixrcond1<Precision>(a, n);
        printf(" %11.1lf",
            double(amp::ampf<Precision>(c1).toDouble()));
        
        //
        // Chebyshev interpolation matrix
        //
        for(i=0; i<=n-1; i++)
        {
            x = amp::ampf<Precision>(2*i)/(amp::ampf<Precision>(n-1))-1;
            a(i,0) = 1;
            if( n>=2 )
            {
                a(i,1) = x;
            }
            for(j=2; j<=n-1; j++)
            {
                a(i,j) = 2*x*a(i,j-1)-a(i,j-2);
            }
        }
        c1 = 1/rcond::rmatrixrcond1<Precision>(a, n);
        printf(" %11.1lf\n",
            double(amp::ampf<Precision>(c1).toDouble()));
    }

schur unit

Functions

rmatrixschur

rmatrixschur function

/************************************************************************* Subroutine performing the Schur decomposition of a general matrix by using the QR algorithm with multiple shifts. The source matrix A is represented as S'*A*S = T, where S is an orthogonal matrix (Schur vectors), T - upper quasi-triangular matrix (with blocks of sizes 1x1 and 2x2 on the main diagonal). Input parameters: A - matrix to be decomposed. Array whose indexes range within [0..N-1, 0..N-1]. N - size of A, N>=0. Output parameters: A - contains matrix T. Array whose indexes range within [0..N-1, 0..N-1]. S - contains Schur vectors. Array whose indexes range within [0..N-1, 0..N-1]. Note 1: The block structure of matrix T can be easily recognized: since all the elements below the blocks are zeros, the elements a[i+1,i] which are equal to 0 show the block border. Note 2: The algorithm performance depends on the value of the internal parameter NS of the InternalSchurDecomposition subroutine which defines the number of shifts in the QR algorithm (similarly to the block width in block-matrix algorithms in linear algebra). If you require maximum performance on your machine, it is recommended to adjust this parameter manually. Result: True, if the algorithm has converged and parameters A and S contain the result. False, if the algorithm has not converged. Algorithm implemented on the basis of the DHSEQR subroutine (LAPACK 3.0 library). *************************************************************************/
template<unsigned int Precision> bool rmatrixschur(ap::template_2d_array< amp::ampf<Precision> >& a, int n, ap::template_2d_array< amp::ampf<Precision> >& s);

sdet unit

Functions

smatrixdet
smatrixldltdet

smatrixdet function

/************************************************************************* Determinant calculation of the symmetric matrix Input parameters: A - matrix. Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - if IsUpper = True, then symmetric matrix A is given by its upper triangle, and the lower triangle isn’t used by subroutine. Similarly, if IsUpper = False, then A is given by its lower triangle. Result: determinant of matrix A. -- ALGLIB -- Copyright 2005-2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> smatrixdet(ap::template_2d_array< amp::ampf<Precision> > a, int n, bool isupper);

smatrixldltdet function

/************************************************************************* Determinant calculation of the matrix given by LDLT decomposition. Input parameters: A - LDLT-decomposition of the matrix, output of subroutine SMatrixLDLT. Pivots - table of permutations which were made during LDLT decomposition, output of subroutine SMatrixLDLT. N - size of matrix A. IsUpper - matrix storage format. The value is equal to the input parameter of subroutine SMatrixLDLT. Result: matrix determinant. -- ALGLIB -- Copyright 2005-2008 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> smatrixldltdet(const ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n, bool isupper);

sinverse unit

Functions

smatrixinverse
smatrixldltinverse

smatrixinverse function

/************************************************************************* Inversion of a symmetric indefinite matrix Given a lower or upper triangle of matrix A, the algorithm generates matrix A^-1 and saves the lower or upper triangle depending on the input. Input parameters: A - matrix to be inverted (upper or lower triangle). Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then the upper triangle of matrix A is given, otherwise the lower triangle is given. Output parameters: A - inverse of matrix A. Array with elements [0..N-1, 0..N-1]. If IsUpper = True, then A contains the upper triangle of matrix A^-1, and the elements below the main diagonal are not used nor changed. The same applies if IsUpper = False. Result: True, if the matrix is not singular. False, if the matrix is singular and could not be inverted. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University March 31, 1993 *************************************************************************/
template<unsigned int Precision> bool smatrixinverse(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper);

smatrixldltinverse function

/************************************************************************* Inversion of a symmetric indefinite matrix The algorithm gets an LDLT-decomposition as an input, generates matrix A^-1 and saves the lower or upper triangle of an inverse matrix depending on the input (U*D*U' or L*D*L'). Input parameters: A - LDLT-decomposition of the matrix, Output of subroutine SMatrixLDLT. N - size of matrix A. IsUpper - storage format. If IsUpper = True, then the symmetric matrix is given as decomposition A = U*D*U' and this decomposition is stored in the upper triangle of matrix A and on the main diagonal, and the lower triangle of matrix A is not used. Pivots - a table of permutations, output of subroutine SMatrixLDLT. Output parameters: A - inverse of the matrix, whose LDLT-decomposition was stored in matrix A as a subroutine input. Array with elements [0..N-1, 0..N-1]. If IsUpper = True, then A contains the upper triangle of matrix A^-1, and the elements below the main diagonal are not used nor changed. The same applies if IsUpper = False. Result: True, if the matrix is not singular. False, if the matrix is singular and could not be inverted. -- LAPACK routine (version 3.0) -- Univ. of Tennessee, Univ. of California Berkeley, NAG Ltd., Courant Institute, Argonne National Lab, and Rice University March 31, 1993 *************************************************************************/
template<unsigned int Precision> bool smatrixldltinverse(ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_1d_array< int >& pivots, int n, bool isupper);

spdgevd unit

Functions

smatrixgevd
smatrixgevdreduce

smatrixgevd function

/************************************************************************* Algorithm for solving the following generalized symmetric positive-definite eigenproblem: A*x = lambda*B*x (1) or A*B*x = lambda*x (2) or B*A*x = lambda*x (3). where A is a symmetric matrix, B - symmetric positive-definite matrix. The problem is solved by reducing it to an ordinary symmetric eigenvalue problem. Input parameters: A - symmetric matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrices A and B. IsUpperA - storage format of matrix A. B - symmetric positive-definite matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. IsUpperB - storage format of matrix B. ZNeeded - if ZNeeded is equal to: * 0, the eigenvectors are not returned; * 1, the eigenvectors are returned. ProblemType - if ProblemType is equal to: * 1, the following problem is solved: A*x = lambda*B*x; * 2, the following problem is solved: A*B*x = lambda*x; * 3, the following problem is solved: B*A*x = lambda*x. Output parameters: D - eigenvalues in ascending order. Array whose index ranges within [0..N-1]. Z - if ZNeeded is equal to: * 0, Z hasn’t changed; * 1, Z contains eigenvectors. Array whose indexes range within [0..N-1, 0..N-1]. The eigenvectors are stored in matrix columns. It should be noted that the eigenvectors in such problems do not form an orthogonal system. Result: True, if the problem was solved successfully. False, if the error occurred during the Cholesky decomposition of matrix B (the matrix isn’t positive-definite) or during the work of the iterative algorithm for solving the symmetric eigenproblem. See also the GeneralizedSymmetricDefiniteEVDReduce subroutine. -- ALGLIB -- Copyright 1.28.2006 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixgevd(ap::template_2d_array< amp::ampf<Precision> > a, int n, bool isuppera, const ap::template_2d_array< amp::ampf<Precision> >& b, bool isupperb, int zneeded, int problemtype, ap::template_1d_array< amp::ampf<Precision> >& d, ap::template_2d_array< amp::ampf<Precision> >& z);

smatrixgevdreduce function

/************************************************************************* Algorithm for reduction of the following generalized symmetric positive- definite eigenvalue problem: A*x = lambda*B*x (1) or A*B*x = lambda*x (2) or B*A*x = lambda*x (3) to the symmetric eigenvalues problem C*y = lambda*y (eigenvalues of this and the given problems are the same, and the eigenvectors of the given problem could be obtained by multiplying the obtained eigenvectors by the transformation matrix x = R*y). Here A is a symmetric matrix, B - symmetric positive-definite matrix. Input parameters: A - symmetric matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. N - size of matrices A and B. IsUpperA - storage format of matrix A. B - symmetric positive-definite matrix which is given by its upper or lower triangular part. Array whose indexes range within [0..N-1, 0..N-1]. IsUpperB - storage format of matrix B. ProblemType - if ProblemType is equal to: * 1, the following problem is solved: A*x = lambda*B*x; * 2, the following problem is solved: A*B*x = lambda*x; * 3, the following problem is solved: B*A*x = lambda*x. Output parameters: A - symmetric matrix which is given by its upper or lower triangle depending on IsUpperA. Contains matrix C. Array whose indexes range within [0..N-1, 0..N-1]. R - upper triangular or low triangular transformation matrix which is used to obtain the eigenvectors of a given problem as the product of eigenvectors of C (from the right) and matrix R (from the left). If the matrix is upper triangular, the elements below the main diagonal are equal to 0 (and vice versa). Thus, we can perform the multiplication without taking into account the internal structure (which is an easier though less effective way). Array whose indexes range within [0..N-1, 0..N-1]. IsUpperR - type of matrix R (upper or lower triangular). Result: True, if the problem was reduced successfully. False, if the error occurred during the Cholesky decomposition of matrix B (the matrix is not positive-definite). -- ALGLIB -- Copyright 1.28.2006 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixgevdreduce(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isuppera, const ap::template_2d_array< amp::ampf<Precision> >& b, bool isupperb, int problemtype, ap::template_2d_array< amp::ampf<Precision> >& r, bool& isupperr);

spline1d unit

Structures

spline1dfitreport
spline1dinterpolant

Functions

spline1dbuildakima
spline1dbuildcatmullrom
spline1dbuildcubic
spline1dbuildhermite
spline1dbuildlinear
spline1dcalc
spline1dcopy
spline1ddiff
spline1dfitcubic
spline1dfitcubicwc
spline1dfithermite
spline1dfithermitewc
spline1dintegrate
spline1dlintransx
spline1dlintransy
spline1dunpack

Examples

spline1d_calc
spline1d_cubic
spline1d_fit
spline1d_fitc
spline1d_hermite
spline1d_linear

spline1dfitreport structure

/************************************************************************* Spline fitting report: TaskRCond reciprocal of task's condition number RMSError RMS error AvgError average error AvgRelError average relative error (for non-zero Y[I]) MaxError maximum error *************************************************************************/
template<unsigned int Precision> class spline1dfitreport { public: amp::ampf<Precision> taskrcond; amp::ampf<Precision> rmserror; amp::ampf<Precision> avgerror; amp::ampf<Precision> avgrelerror; amp::ampf<Precision> maxerror; };

spline1dinterpolant structure

/************************************************************************* 1-dimensional spline inteprolant *************************************************************************/
template<unsigned int Precision> class spline1dinterpolant { public: bool periodic; int n; int k; ap::template_1d_array< amp::ampf<Precision> > x; ap::template_1d_array< amp::ampf<Precision> > c; };

spline1dbuildakima function

/************************************************************************* This subroutine builds Akima spline interpolant INPUT PARAMETERS: X - spline nodes, array[0..N-1] Y - function values, array[0..N-1] N - points count, N>=5 OUTPUT PARAMETERS: C - spline interpolant ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. -- ALGLIB PROJECT -- Copyright 24.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dbuildakima(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, int n, spline1dinterpolant<Precision>& c);

spline1dbuildcatmullrom function

/************************************************************************* This subroutine builds Catmull-Rom spline interpolant. INPUT PARAMETERS: X - spline nodes, array[0..N-1]. Y - function values, array[0..N-1]. N - points count, N>=2 BoundType - boundary condition type: * -1 for periodic boundary condition * 0 for parabolically terminated spline Tension - tension parameter: * tension=0 corresponds to classic Catmull-Rom spline * 0<tension<1 corresponds to more general form - cardinal spline OUTPUT PARAMETERS: C - spline interpolant ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. PROBLEMS WITH PERIODIC BOUNDARY CONDITIONS: Problems with periodic boundary conditions have Y[first_point]=Y[last_point]. However, this subroutine doesn't require you to specify equal values for the first and last points - it automatically forces them to be equal. -- ALGLIB PROJECT -- Copyright 23.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dbuildcatmullrom(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, int n, int boundtype, amp::ampf<Precision> tension, spline1dinterpolant<Precision>& c);

spline1dbuildcubic function

/************************************************************************* This subroutine builds cubic spline interpolant. INPUT PARAMETERS: X - spline nodes, array[0..N-1]. Y - function values, array[0..N-1]. N - points count, N>=2 BoundLType - boundary condition type for the left boundary BoundL - left boundary condition (first or second derivative, depending on the BoundLType) BoundRType - boundary condition type for the right boundary BoundR - right boundary condition (first or second derivative, depending on the BoundRType) OUTPUT PARAMETERS: C - spline interpolant ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. SETTING BOUNDARY VALUES: The BoundLType/BoundRType parameters can have the following values: * -1, which corresonds to the periodic (cyclic) boundary conditions. In this case: * both BoundLType and BoundRType must be equal to -1. * BoundL/BoundR are ignored * Y[last] is ignored (it is assumed to be equal to Y[first]). * 0, which corresponds to the parabolically terminated spline (BoundL and/or BoundR are ignored). * 1, which corresponds to the first derivative boundary condition * 2, which corresponds to the second derivative boundary condition PROBLEMS WITH PERIODIC BOUNDARY CONDITIONS: Problems with periodic boundary conditions have Y[first_point]=Y[last_point]. However, this subroutine doesn't require you to specify equal values for the first and last points - it automatically forces them to be equal. -- ALGLIB PROJECT -- Copyright 23.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dbuildcubic(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, int n, int boundltype, amp::ampf<Precision> boundl, int boundrtype, amp::ampf<Precision> boundr, spline1dinterpolant<Precision>& c);

Examples:   spline1d_calc  spline1d_cubic  

spline1dbuildhermite function

/************************************************************************* This subroutine builds Hermite spline interpolant. INPUT PARAMETERS: X - spline nodes, array[0..N-1] Y - function values, array[0..N-1] D - derivatives, array[0..N-1] N - points count, N>=2 OUTPUT PARAMETERS: C - spline interpolant. ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. -- ALGLIB PROJECT -- Copyright 23.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dbuildhermite(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, ap::template_1d_array< amp::ampf<Precision> > d, int n, spline1dinterpolant<Precision>& c);

Examples:   spline1d_hermite  

spline1dbuildlinear function

/************************************************************************* This subroutine builds linear spline interpolant INPUT PARAMETERS: X - spline nodes, array[0..N-1] Y - function values, array[0..N-1] N - points count, N>=2 OUTPUT PARAMETERS: C - spline interpolant ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. -- ALGLIB PROJECT -- Copyright 24.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dbuildlinear(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, int n, spline1dinterpolant<Precision>& c);

Examples:   spline1d_linear  

spline1dcalc function

/************************************************************************* This subroutine calculates the value of the spline at the given point X. INPUT PARAMETERS: C - spline interpolant X - point Result: S(x) -- ALGLIB PROJECT -- Copyright 23.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spline1dcalc(const spline1dinterpolant<Precision>& c, amp::ampf<Precision> x);

Examples:   spline1d_calc  spline1d_cubic  spline1d_hermite  spline1d_linear  

spline1dcopy function

/************************************************************************* This subroutine makes the copy of the spline. INPUT PARAMETERS: C - spline interpolant. Result: CC - spline copy -- ALGLIB PROJECT -- Copyright 29.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dcopy(const spline1dinterpolant<Precision>& c, spline1dinterpolant<Precision>& cc);

spline1ddiff function

/************************************************************************* This subroutine differentiates the spline. INPUT PARAMETERS: C - spline interpolant. X - point Result: S - S(x) DS - S'(x) D2S - S''(x) -- ALGLIB PROJECT -- Copyright 24.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1ddiff(const spline1dinterpolant<Precision>& c, amp::ampf<Precision> x, amp::ampf<Precision>& s, amp::ampf<Precision>& ds, amp::ampf<Precision>& d2s);

Examples:   spline1d_calc  

spline1dfitcubic function

/************************************************************************* Least squares fitting by cubic spline. This subroutine is "lightweight" alternative for more complex and feature- rich Spline1DFitCubicWC(). See Spline1DFitCubicWC() for more information about subroutine parameters (we don't duplicate it here because of length) -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dfitcubic(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, int m, int& info, spline1dinterpolant<Precision>& s, spline1dfitreport<Precision>& rep);

Examples:   spline1d_fit  

spline1dfitcubicwc function

/************************************************************************* Weighted fitting by cubic spline, with constraints on function values or derivatives. Equidistant grid with M-2 nodes on [min(x,xc),max(x,xc)] is used to build basis functions. Basis functions are cubic splines with continuous second derivatives and non-fixed first derivatives at interval ends. Small regularizing term is used when solving constrained tasks (to improve stability). Task is linear, so linear least squares solver is used. Complexity of this computational scheme is O(N*M^2), mostly dominated by least squares solver SEE ALSO Spline1DFitHermiteWC() - fitting by Hermite splines (more flexible, less smooth) Spline1DFitCubic() - "lightweight" fitting by cubic splines, without invididual weights and constraints INPUT PARAMETERS: X - points, array[0..N-1]. Y - function values, array[0..N-1]. W - weights, array[0..N-1] Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. Fill it by 1's if you don't want to solve weighted task. N - number of points, N>0. XC - points where spline values/derivatives are constrained, array[0..K-1]. YC - values of constraints, array[0..K-1] DC - array[0..K-1], types of constraints: * DC[i]=0 means that S(XC[i])=YC[i] * DC[i]=1 means that S'(XC[i])=YC[i] SEE BELOW FOR IMPORTANT INFORMATION ON CONSTRAINTS K - number of constraints, 0<=K<M. K=0 means no constraints (XC/YC/DC are not used in such cases) M - number of basis functions ( = number_of_nodes+2), M>=4. OUTPUT PARAMETERS: Info- same format as in LSFitLinearWC() subroutine. * Info>0 task is solved * Info<=0 an error occured: -4 means inconvergence of internal SVD -3 means inconsistent constraints -1 means another errors in parameters passed (N<=0, for example) S - spline interpolant. Rep - report, same format as in LSFitLinearWC() subroutine. Following fields are set: * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED IMPORTANT: this subroitine doesn't calculate task's condition number for K<>0. ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. SETTING CONSTRAINTS - DANGERS AND OPPORTUNITIES: Setting constraints can lead to undesired results, like ill-conditioned behavior, or inconsistency being detected. From the other side, it allows us to improve quality of the fit. Here we summarize our experience with constrained regression splines: * excessive constraints can be inconsistent. Splines are piecewise cubic functions, and it is easy to create an example, where large number of constraints concentrated in small area will result in inconsistency. Just because spline is not flexible enough to satisfy all of them. And same constraints spread across the [min(x),max(x)] will be perfectly consistent. * the more evenly constraints are spread across [min(x),max(x)], the more chances that they will be consistent * the greater is M (given fixed constraints), the more chances that constraints will be consistent * in the general case, consistency of constraints IS NOT GUARANTEED. * in the several special cases, however, we CAN guarantee consistency. * one of this cases is constraints on the function values AND/OR its derivatives at the interval boundaries. * another special case is ONE constraint on the function value (OR, but not AND, derivative) anywhere in the interval Our final recommendation is to use constraints WHEN AND ONLY WHEN you can't solve your task without them. Anything beyond special cases given above is not guaranteed and may result in inconsistency. -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dfitcubicwc(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, int n, const ap::template_1d_array< amp::ampf<Precision> >& xc, const ap::template_1d_array< amp::ampf<Precision> >& yc, const ap::template_1d_array< int >& dc, int k, int m, int& info, spline1dinterpolant<Precision>& s, spline1dfitreport<Precision>& rep);

Examples:   spline1d_fitc  

spline1dfithermite function

/************************************************************************* Least squares fitting by Hermite spline. This subroutine is "lightweight" alternative for more complex and feature- rich Spline1DFitHermiteWC(). See Spline1DFitHermiteWC() description for more information about subroutine parameters (we don't duplicate it here because of length). -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dfithermite(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, int n, int m, int& info, spline1dinterpolant<Precision>& s, spline1dfitreport<Precision>& rep);

Examples:   spline1d_fit  

spline1dfithermitewc function

/************************************************************************* Weighted fitting by Hermite spline, with constraints on function values or first derivatives. Equidistant grid with M nodes on [min(x,xc),max(x,xc)] is used to build basis functions. Basis functions are Hermite splines. Small regularizing term is used when solving constrained tasks (to improve stability). Task is linear, so linear least squares solver is used. Complexity of this computational scheme is O(N*M^2), mostly dominated by least squares solver SEE ALSO Spline1DFitCubicWC() - fitting by Cubic splines (less flexible, more smooth) Spline1DFitHermite() - "lightweight" Hermite fitting, without invididual weights and constraints INPUT PARAMETERS: X - points, array[0..N-1]. Y - function values, array[0..N-1]. W - weights, array[0..N-1] Each summand in square sum of approximation deviations from given values is multiplied by the square of corresponding weight. Fill it by 1's if you don't want to solve weighted task. N - number of points, N>0. XC - points where spline values/derivatives are constrained, array[0..K-1]. YC - values of constraints, array[0..K-1] DC - array[0..K-1], types of constraints: * DC[i]=0 means that S(XC[i])=YC[i] * DC[i]=1 means that S'(XC[i])=YC[i] SEE BELOW FOR IMPORTANT INFORMATION ON CONSTRAINTS K - number of constraints, 0<=K<M. K=0 means no constraints (XC/YC/DC are not used in such cases) M - number of basis functions (= 2 * number of nodes), M>=4, M IS EVEN! OUTPUT PARAMETERS: Info- same format as in LSFitLinearW() subroutine: * Info>0 task is solved * Info<=0 an error occured: -4 means inconvergence of internal SVD -3 means inconsistent constraints -2 means odd M was passed (which is not supported) -1 means another errors in parameters passed (N<=0, for example) S - spline interpolant. Rep - report, same format as in LSFitLinearW() subroutine. Following fields are set: * RMSError rms error on the (X,Y). * AvgError average error on the (X,Y). * AvgRelError average relative error on the non-zero Y * MaxError maximum error NON-WEIGHTED ERRORS ARE CALCULATED IMPORTANT: this subroitine doesn't calculate task's condition number for K<>0. IMPORTANT: this subroitine supports only even M's ORDER OF POINTS Subroutine automatically sorts points, so caller may pass unsorted array. SETTING CONSTRAINTS - DANGERS AND OPPORTUNITIES: Setting constraints can lead to undesired results, like ill-conditioned behavior, or inconsistency being detected. From the other side, it allows us to improve quality of the fit. Here we summarize our experience with constrained regression splines: * excessive constraints can be inconsistent. Splines are piecewise cubic functions, and it is easy to create an example, where large number of constraints concentrated in small area will result in inconsistency. Just because spline is not flexible enough to satisfy all of them. And same constraints spread across the [min(x),max(x)] will be perfectly consistent. * the more evenly constraints are spread across [min(x),max(x)], the more chances that they will be consistent * the greater is M (given fixed constraints), the more chances that constraints will be consistent * in the general case, consistency of constraints is NOT GUARANTEED. * in the several special cases, however, we can guarantee consistency. * one of this cases is M>=4 and constraints on the function value (AND/OR its derivative) at the interval boundaries. * another special case is M>=4 and ONE constraint on the function value (OR, BUT NOT AND, derivative) anywhere in [min(x),max(x)] Our final recommendation is to use constraints WHEN AND ONLY when you can't solve your task without them. Anything beyond special cases given above is not guaranteed and may result in inconsistency. -- ALGLIB PROJECT -- Copyright 18.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dfithermitewc(const ap::template_1d_array< amp::ampf<Precision> >& x, const ap::template_1d_array< amp::ampf<Precision> >& y, const ap::template_1d_array< amp::ampf<Precision> >& w, int n, const ap::template_1d_array< amp::ampf<Precision> >& xc, const ap::template_1d_array< amp::ampf<Precision> >& yc, const ap::template_1d_array< int >& dc, int k, int m, int& info, spline1dinterpolant<Precision>& s, spline1dfitreport<Precision>& rep);

Examples:   spline1d_fitc  

spline1dintegrate function

/************************************************************************* This subroutine integrates the spline. INPUT PARAMETERS: C - spline interpolant. X - right bound of the integration interval [a, x], here 'a' denotes min(x[]) Result: integral(S(t)dt,a,x) -- ALGLIB PROJECT -- Copyright 23.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spline1dintegrate(const spline1dinterpolant<Precision>& c, amp::ampf<Precision> x);

Examples:   spline1d_calc  

spline1dlintransx function

/************************************************************************* This subroutine performs linear transformation of the spline argument. INPUT PARAMETERS: C - spline interpolant. A, B- transformation coefficients: x = A*t + B Result: C - transformed spline -- ALGLIB PROJECT -- Copyright 30.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dlintransx(spline1dinterpolant<Precision>& c, amp::ampf<Precision> a, amp::ampf<Precision> b);

spline1dlintransy function

/************************************************************************* This subroutine performs linear transformation of the spline. INPUT PARAMETERS: C - spline interpolant. A, B- transformation coefficients: S2(x) = A*S(x) + B Result: C - transformed spline -- ALGLIB PROJECT -- Copyright 30.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dlintransy(spline1dinterpolant<Precision>& c, amp::ampf<Precision> a, amp::ampf<Precision> b);

spline1dunpack function

/************************************************************************* This subroutine unpacks the spline into the coefficients table. INPUT PARAMETERS: C - spline interpolant. X - point Result: Tbl - coefficients table, unpacked format, array[0..N-2, 0..5]. For I = 0...N-2: Tbl[I,0] = X[i] Tbl[I,1] = X[i+1] Tbl[I,2] = C0 Tbl[I,3] = C1 Tbl[I,4] = C2 Tbl[I,5] = C3 On [x[i], x[i+1]] spline is equals to: S(x) = C0 + C1*t + C2*t^2 + C3*t^3 t = x-x[i] -- ALGLIB PROJECT -- Copyright 29.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline1dunpack(const spline1dinterpolant<Precision>& c, int& n, ap::template_2d_array< amp::ampf<Precision> >& tbl);

spline1d_calc example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> v;
    amp::ampf<Precision> dv;
    amp::ampf<Precision> d2v;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Demonstration of Spline1DCalc(), Spline1DDiff(), Spline1DIntegrate()
    //
    printf("DEMONSTRATION OF Spline1DCalc(), Spline1DDiff(), Spline1DIntegrate()\n\n");
    printf("F(x)=sin(x), [0, pi]\n");
    printf("Natural cubic spline with 3 nodes is used\n\n");
    
    //
    // Create spline
    //
    n = 3;
    x.setlength(n);
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
    }
    spline1d::spline1dbuildcubic<Precision>(x, y, n, 2, amp::ampf<Precision>("0.0"), 2, amp::ampf<Precision>("0.0"), s);
    
    //
    // Output results
    //
    spline1d::spline1ddiff<Precision>(s, amp::ampf<Precision>(0), v, dv, d2v);
    printf("                 S(x)    F(x) \n");
    printf("function       %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, amp::ampf<Precision>(0))).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("d/dx(0)        %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(dv).toDouble()),
        double(amp::ampf<Precision>(1).toDouble()));
    printf("d2/dx2(0)      %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(d2v).toDouble()),
        double(amp::ampf<Precision>(0).toDouble()));
    printf("integral(0,pi) %6.3lf  %6.3lf \n",
        double(amp::ampf<Precision>(spline1d::spline1dintegrate<Precision>(s, amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(2).toDouble()));
    printf("\n\n");

spline1d_cubic example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Interpolation by natural Cubic spline.
    //
    printf("INTERPOLATION BY NATURAL CUBIC SPLINE\n\n");
    printf("F(x)=sin(x), [0, pi], 3 nodes\n\n");
    printf("     x   F(x)   S(x)  Error\n");
    
    //
    // Create spline
    //
    n = 3;
    x.setlength(n);
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
    }
    spline1d::spline1dbuildcubic<Precision>(x, y, n, 1, amp::ampf<Precision>(+1), 1, amp::ampf<Precision>(-1), s);
    
    //
    // Output results
    //
    t = 0;
    maxerr = 0;
    while( t<amp::ampf<Precision>("0.999999")*amp::pi<Precision>() )
    {
        err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t));
        maxerr = amp::maximum<Precision>(err, maxerr);
        printf("%6.3lf %6.3lf %6.3lf %6.3lf\n",
            double(amp::ampf<Precision>(t).toDouble()),
            double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
            double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
            double(amp::ampf<Precision>(err).toDouble()));
        t = amp::minimum<Precision>(amp::pi<Precision>(), t+amp::ampf<Precision>("0.25"));
    }
    err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())-amp::sin<Precision>(amp::pi<Precision>()));
    maxerr = amp::maximum<Precision>(err, maxerr);
    printf("%6.3lf %6.3lf %6.3lf %6.3lf\n\n",
        double(amp::ampf<Precision>(amp::pi<Precision>()).toDouble()),
        double(amp::ampf<Precision>(amp::sin<Precision>(amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(err).toDouble()));
    printf("max|error| = %0.3lf\n",
        double(amp::ampf<Precision>(maxerr).toDouble()));
    printf("Try other demos (spline1d_linear, spline1d_hermite) and compare errors...\n\n\n");

spline1d_fit example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    int info;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> t;
    spline1d::spline1dfitreport<Precision> rep;


    
    //
    // Fitting by unconstrained natural cubic spline
    //
    printf("FITTING BY UNCONSTRAINED NATURAL CUBIC SPLINE\n\n");
    printf("F(x)=sin(x)      function being fitted\n");
    printf("[0, pi]          interval\n");
    printf("M=4              number of basis functions to use\n");
    printf("N=100            number of points to fit\n");
    
    //
    // Create and fit
    //
    n = 100;
    x.setlength(n);
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
    }
    spline1d::spline1dfitcubic<Precision>(x, y, n, 4, info, s, rep);
    
    //
    // Output results
    //
    if( info>0 )
    {
        printf("\nOK, we have finished\n\n");
        printf("     x   F(x)   S(x)  Error\n");
        t = 0;
        while( t<amp::ampf<Precision>("0.999999")*amp::pi<Precision>() )
        {
            printf("%6.3lf %6.3lf %6.3lf %6.3lf\n",
                double(amp::ampf<Precision>(t).toDouble()),
                double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
                double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
                double(amp::ampf<Precision>(amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t))).toDouble()));
            t = amp::minimum<Precision>(amp::pi<Precision>(), t+amp::ampf<Precision>("0.25"));
        }
        printf("%6.3lf %6.3lf %6.3lf %6.3lf\n\n",
            double(amp::ampf<Precision>(t).toDouble()),
            double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
            double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
            double(amp::ampf<Precision>(amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t))).toDouble()));
        printf("rms error is %6.3lf\n",
            double(amp::ampf<Precision>(rep.rmserror).toDouble()));
        printf("max error is %6.3lf\n",
            double(amp::ampf<Precision>(rep.maxerror).toDouble()));
    }
    else
    {
        printf("\nSomething wrong, Info=%0ld",
            long(info));
    }

spline1d_fitc example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_1d_array< amp::ampf<Precision> > w;
    ap::template_1d_array< amp::ampf<Precision> > xc;
    ap::template_1d_array< amp::ampf<Precision> > yc;
    ap::template_1d_array< int > dc;
    int n;
    int i;
    int info;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> t;
    spline1d::spline1dfitreport<Precision> rep;


    
    //
    // Fitting by constrained Hermite spline
    //
    printf("FITTING BY CONSTRAINED HERMITE SPLINE\n\n");
    printf("F(x)=sin(x)      function being fitted\n");
    printf("[0, pi]          interval\n");
    printf("M=6              number of basis functions to use\n");
    printf("S(0)=0           first constraint\n");
    printf("S(pi)=0          second constraint\n");
    printf("N=100            number of points to fit\n");
    
    //
    // Create and fit:
    // * X  contains points
    // * Y  contains values
    // * W  contains weights
    // * XC contains constraints locations
    // * YC contains constraints values
    // * DC contains derivative indexes (0 = constrained function value)
    //
    n = 100;
    x.setlength(n);
    y.setlength(n);
    w.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
        w(i) = 1;
    }
    xc.setlength(2);
    yc.setlength(2);
    dc.setlength(2);
    xc(0) = 0;
    yc(0) = 0;
    dc(0) = 0;
    xc(0) = amp::pi<Precision>();
    yc(0) = 0;
    dc(0) = 0;
    spline1d::spline1dfithermitewc<Precision>(x, y, w, n, xc, yc, dc, 2, 6, info, s, rep);
    
    //
    // Output results
    //
    if( info>0 )
    {
        printf("\nOK, we have finished\n\n");
        printf("     x   F(x)   S(x)  Error\n");
        t = 0;
        while( t<amp::ampf<Precision>("0.999999")*amp::pi<Precision>() )
        {
            printf("%6.3lf %6.3lf %6.3lf %6.3lf\n",
                double(amp::ampf<Precision>(t).toDouble()),
                double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
                double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
                double(amp::ampf<Precision>(amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t))).toDouble()));
            t = amp::minimum<Precision>(amp::pi<Precision>(), t+amp::ampf<Precision>("0.25"));
        }
        printf("%6.3lf %6.3lf %6.3lf %6.3lf\n\n",
            double(amp::ampf<Precision>(t).toDouble()),
            double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
            double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
            double(amp::ampf<Precision>(amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t))).toDouble()));
        printf("rms error is %6.3lf\n",
            double(amp::ampf<Precision>(rep.rmserror).toDouble()));
        printf("max error is %6.3lf\n",
            double(amp::ampf<Precision>(rep.maxerror).toDouble()));
        printf("S(0) = S(pi) = 0 (exactly)\n\n");
    }
    else
    {
        printf("\nSomething wrong, Info=%0ld",
            long(info));
    }

spline1d_hermite example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    ap::template_1d_array< amp::ampf<Precision> > d;
    int n;
    int i;
    amp::ampf<Precision> t;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Interpolation by natural Cubic spline.
    //
    printf("INTERPOLATION BY HERMITE SPLINE\n\n");
    printf("F(x)=sin(x), [0, pi], 3 nodes\n\n");
    printf("     x   F(x)   S(x)  Error\n");
    
    //
    // Create spline
    //
    n = 3;
    x.setlength(n);
    y.setlength(n);
    d.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
        d(i) = amp::cos<Precision>(x(i));
    }
    spline1d::spline1dbuildhermite<Precision>(x, y, d, n, s);
    
    //
    // Output results
    //
    t = 0;
    maxerr = 0;
    while( t<amp::ampf<Precision>("0.999999")*amp::pi<Precision>() )
    {
        err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t));
        maxerr = amp::maximum<Precision>(err, maxerr);
        printf("%6.3lf %6.3lf %6.3lf %6.3lf\n",
            double(amp::ampf<Precision>(t).toDouble()),
            double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
            double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
            double(amp::ampf<Precision>(err).toDouble()));
        t = amp::minimum<Precision>(amp::pi<Precision>(), t+amp::ampf<Precision>("0.25"));
    }
    err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())-amp::sin<Precision>(amp::pi<Precision>()));
    maxerr = amp::maximum<Precision>(err, maxerr);
    printf("%6.3lf %6.3lf %6.3lf %6.3lf\n\n",
        double(amp::ampf<Precision>(amp::pi<Precision>()).toDouble()),
        double(amp::ampf<Precision>(amp::sin<Precision>(amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(err).toDouble()));
    printf("max|error| = %0.3lf\n",
        double(amp::ampf<Precision>(maxerr).toDouble()));
    printf("Try other demos (spline1d_linear, spline1d_cubic) and compare errors...\n\n\n");

spline1d_linear example

    ap::template_1d_array< amp::ampf<Precision> > x;
    ap::template_1d_array< amp::ampf<Precision> > y;
    int n;
    int i;
    amp::ampf<Precision> t;
    spline1d::spline1dinterpolant<Precision> s;
    amp::ampf<Precision> err;
    amp::ampf<Precision> maxerr;


    
    //
    // Interpolation by linear spline.
    //
    printf("INTERPOLATION BY LINEAR SPLINE\n\n");
    printf("F(x)=sin(x), [0, pi], 3 nodes\n\n");
    printf("     x   F(x)   S(x)  Error\n");
    
    //
    // Create spline
    //
    n = 3;
    x.setlength(n);
    y.setlength(n);
    for(i=0; i<=n-1; i++)
    {
        x(i) = amp::pi<Precision>()*i/(n-1);
        y(i) = amp::sin<Precision>(x(i));
    }
    spline1d::spline1dbuildlinear<Precision>(x, y, n, s);
    
    //
    // Output results
    //
    t = 0;
    maxerr = 0;
    while( t<amp::ampf<Precision>("0.999999")*amp::pi<Precision>() )
    {
        err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, t)-amp::sin<Precision>(t));
        maxerr = amp::maximum<Precision>(err, maxerr);
        printf("%6.3lf %6.3lf %6.3lf %6.3lf\n",
            double(amp::ampf<Precision>(t).toDouble()),
            double(amp::ampf<Precision>(amp::sin<Precision>(t)).toDouble()),
            double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, t)).toDouble()),
            double(amp::ampf<Precision>(err).toDouble()));
        t = amp::minimum<Precision>(amp::pi<Precision>(), t+amp::ampf<Precision>("0.25"));
    }
    err = amp::abs<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())-amp::sin<Precision>(amp::pi<Precision>()));
    maxerr = amp::maximum<Precision>(err, maxerr);
    printf("%6.3lf %6.3lf %6.3lf %6.3lf\n\n",
        double(amp::ampf<Precision>(amp::pi<Precision>()).toDouble()),
        double(amp::ampf<Precision>(amp::sin<Precision>(amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(spline1d::spline1dcalc<Precision>(s, amp::pi<Precision>())).toDouble()),
        double(amp::ampf<Precision>(err).toDouble()));
    printf("max|error| = %0.3lf\n",
        double(amp::ampf<Precision>(maxerr).toDouble()));
    printf("Try other demos (spline1d_hermite, spline1d_cubic) and compare errors...\n\n\n");

spline2d unit

Structures

spline2dinterpolant

Functions

spline2dbuildbicubic
spline2dbuildbilinear
spline2dcalc
spline2dcopy
spline2ddiff
spline2dlintransf
spline2dlintransxy
spline2dresamplebicubic
spline2dresamplebilinear
spline2dserialize
spline2dunpack
spline2dunserialize

spline2dinterpolant structure

/************************************************************************* 2-dimensional spline inteprolant *************************************************************************/
template<unsigned int Precision> class spline2dinterpolant { public: int k; ap::template_1d_array< amp::ampf<Precision> > c; };

spline2dbuildbicubic function

/************************************************************************* This subroutine builds bicubic spline coefficients table. Input parameters: X - spline abscissas, array[0..N-1] Y - spline ordinates, array[0..M-1] F - function values, array[0..M-1,0..N-1] M,N - grid size, M>=2, N>=2 Output parameters: C - spline interpolant -- ALGLIB PROJECT -- Copyright 05.07.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dbuildbicubic(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, ap::template_2d_array< amp::ampf<Precision> > f, int m, int n, spline2dinterpolant<Precision>& c);

spline2dbuildbilinear function

/************************************************************************* This subroutine builds bilinear spline coefficients table. Input parameters: X - spline abscissas, array[0..N-1] Y - spline ordinates, array[0..M-1] F - function values, array[0..M-1,0..N-1] M,N - grid size, M>=2, N>=2 Output parameters: C - spline interpolant -- ALGLIB PROJECT -- Copyright 05.07.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dbuildbilinear(ap::template_1d_array< amp::ampf<Precision> > x, ap::template_1d_array< amp::ampf<Precision> > y, ap::template_2d_array< amp::ampf<Precision> > f, int m, int n, spline2dinterpolant<Precision>& c);

spline2dcalc function

/************************************************************************* This subroutine calculates the value of the bilinear or bicubic spline at the given point X. Input parameters: C - coefficients table. Built by BuildBilinearSpline or BuildBicubicSpline. X, Y- point Result: S(x,y) -- ALGLIB PROJECT -- Copyright 05.07.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> spline2dcalc(const spline2dinterpolant<Precision>& c, amp::ampf<Precision> x, amp::ampf<Precision> y);

spline2dcopy function

/************************************************************************* This subroutine makes the copy of the spline model. Input parameters: C - spline interpolant Output parameters: CC - spline copy -- ALGLIB PROJECT -- Copyright 29.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dcopy(const spline2dinterpolant<Precision>& c, spline2dinterpolant<Precision>& cc);

spline2ddiff function

/************************************************************************* This subroutine calculates the value of the bilinear or bicubic spline at the given point X and its derivatives. Input parameters: C - spline interpolant. X, Y- point Output parameters: F - S(x,y) FX - dS(x,y)/dX FY - dS(x,y)/dY FXY - d2S(x,y)/dXdY -- ALGLIB PROJECT -- Copyright 05.07.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2ddiff(const spline2dinterpolant<Precision>& c, amp::ampf<Precision> x, amp::ampf<Precision> y, amp::ampf<Precision>& f, amp::ampf<Precision>& fx, amp::ampf<Precision>& fy, amp::ampf<Precision>& fxy);

spline2dlintransf function

/************************************************************************* This subroutine performs linear transformation of the spline. Input parameters: C - spline interpolant. A, B- transformation coefficients: S2(x,y) = A*S(x,y) + B Output parameters: C - transformed spline -- ALGLIB PROJECT -- Copyright 30.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dlintransf(spline2dinterpolant<Precision>& c, amp::ampf<Precision> a, amp::ampf<Precision> b);

spline2dlintransxy function

/************************************************************************* This subroutine performs linear transformation of the spline argument. Input parameters: C - spline interpolant AX, BX - transformation coefficients: x = A*t + B AY, BY - transformation coefficients: y = A*u + B Result: C - transformed spline -- ALGLIB PROJECT -- Copyright 30.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dlintransxy(spline2dinterpolant<Precision>& c, amp::ampf<Precision> ax, amp::ampf<Precision> bx, amp::ampf<Precision> ay, amp::ampf<Precision> by);

spline2dresamplebicubic function

/************************************************************************* Bicubic spline resampling Input parameters: A - function values at the old grid, array[0..OldHeight-1, 0..OldWidth-1] OldHeight - old grid height, OldHeight>1 OldWidth - old grid width, OldWidth>1 NewHeight - new grid height, NewHeight>1 NewWidth - new grid width, NewWidth>1 Output parameters: B - function values at the new grid, array[0..NewHeight-1, 0..NewWidth-1] -- ALGLIB routine -- 15 May, 2007 Copyright by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dresamplebicubic(const ap::template_2d_array< amp::ampf<Precision> >& a, int oldheight, int oldwidth, ap::template_2d_array< amp::ampf<Precision> >& b, int newheight, int newwidth);

spline2dresamplebilinear function

/************************************************************************* Bilinear spline resampling Input parameters: A - function values at the old grid, array[0..OldHeight-1, 0..OldWidth-1] OldHeight - old grid height, OldHeight>1 OldWidth - old grid width, OldWidth>1 NewHeight - new grid height, NewHeight>1 NewWidth - new grid width, NewWidth>1 Output parameters: B - function values at the new grid, array[0..NewHeight-1, 0..NewWidth-1] -- ALGLIB routine -- 09.07.2007 Copyright by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dresamplebilinear(const ap::template_2d_array< amp::ampf<Precision> >& a, int oldheight, int oldwidth, ap::template_2d_array< amp::ampf<Precision> >& b, int newheight, int newwidth);

spline2dserialize function

/************************************************************************* Serialization of the spline interpolant INPUT PARAMETERS: B - spline interpolant OUTPUT PARAMETERS: RA - array of real numbers which contains interpolant, array[0..RLen-1] RLen - RA lenght -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dserialize(const spline2dinterpolant<Precision>& c, ap::template_1d_array< amp::ampf<Precision> >& ra, int& ralen);

spline2dunpack function

/************************************************************************* This subroutine unpacks two-dimensional spline into the coefficients table Input parameters: C - spline interpolant. Result: M, N- grid size (x-axis and y-axis) Tbl - coefficients table, unpacked format, [0..(N-1)*(M-1)-1, 0..19]. For I = 0...M-2, J=0..N-2: K = I*(N-1)+J Tbl[K,0] = X[j] Tbl[K,1] = X[j+1] Tbl[K,2] = Y[i] Tbl[K,3] = Y[i+1] Tbl[K,4] = C00 Tbl[K,5] = C01 Tbl[K,6] = C02 Tbl[K,7] = C03 Tbl[K,8] = C10 Tbl[K,9] = C11 ... Tbl[K,19] = C33 On each grid square spline is equals to: S(x) = SUM(c[i,j]*(x^i)*(y^j), i=0..3, j=0..3) t = x-x[j] u = y-y[i] -- ALGLIB PROJECT -- Copyright 29.06.2007 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dunpack(const spline2dinterpolant<Precision>& c, int& m, int& n, ap::template_2d_array< amp::ampf<Precision> >& tbl);

spline2dunserialize function

/************************************************************************* Unserialization of the spline interpolant INPUT PARAMETERS: RA - array of real numbers which contains interpolant, OUTPUT PARAMETERS: B - spline interpolant -- ALGLIB -- Copyright 17.08.2009 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void spline2dunserialize(const ap::template_1d_array< amp::ampf<Precision> >& ra, spline2dinterpolant<Precision>& c);

srcond unit

Functions

smatrixldltrcond
smatrixrcond

smatrixldltrcond function

/************************************************************************* Condition number estimate of a matrix given by LDLT-decomposition The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: L - LDLT-decomposition of matrix A given by the upper or lower triangle depending on IsUpper. Output of SMatrixLDLT subroutine. Pivots - table of permutations which were made during LDLT-decomposition, Output of SMatrixLDLT subroutine. N - size of matrix A. IsUpper - storage format. Result: 1/LowerBound(cond(A)) *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> smatrixldltrcond(const ap::template_2d_array< amp::ampf<Precision> >& l, const ap::template_1d_array< int >& pivots, int n, bool isupper);

smatrixrcond function

/************************************************************************* Condition number estimate of a symmetric matrix The algorithm calculates a lower bound of the condition number. In this case, the algorithm does not return a lower bound of the condition number, but an inverse number (to avoid an overflow in case of a singular matrix). It should be noted that 1-norm and inf-norm condition numbers of symmetric matrices are equal, so the algorithm doesn't take into account the differences between these types of norms. Input parameters: A - symmetric definite matrix which is given by its upper or lower triangle depending on IsUpper. Array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - storage format. Result: 1/LowerBound(cond(A)) *************************************************************************/
template<unsigned int Precision> amp::ampf<Precision> smatrixrcond(const ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper);

ssolve unit

Functions

smatrixldltsolve
smatrixsolve

smatrixldltsolve function

/************************************************************************* Solving a system of linear equations with a system matrix given by its LDLT decomposition The algorithm solves systems with a square matrix only. Input parameters: A - LDLT decomposition of the matrix (the result of the SMatrixLDLT subroutine). Pivots - row permutation table (the result of the SMatrixLDLT subroutine). B - right side of a system. Array whose index ranges within [0..N-1]. N - size of matrix A. IsUpper - points to the triangle of matrix A in which the LDLT decomposition is stored. If IsUpper=True, the decomposition has the form of U*D*U', matrix U is stored in the upper triangle of matrix A (in that case, the lower triangle isn't used and isn't changed by the subroutine). Similarly, if IsUpper=False, the decomposition has the form of L*D*L' and the lower triangle stores matrix L. Output parameters: X - solution of a system. Array whose index ranges within [0..N-1]. Result: True, if the matrix is not singular. X contains the solution. False, if the matrix is singular (the determinant of matrix D is equal to 0). In this case, X doesn't contain a solution. *************************************************************************/
template<unsigned int Precision> bool smatrixldltsolve(const ap::template_2d_array< amp::ampf<Precision> >& a, const ap::template_1d_array< int >& pivots, ap::template_1d_array< amp::ampf<Precision> > b, int n, bool isupper, ap::template_1d_array< amp::ampf<Precision> >& x);

smatrixsolve function

/************************************************************************* Solving a system of linear equations with a symmetric system matrix Input parameters: A - system matrix (upper or lower triangle). Array whose indexes range within [0..N-1, 0..N-1]. B - right side of a system. Array whose index ranges within [0..N-1]. N - size of matrix A. IsUpper - If IsUpper = True, A contains the upper triangle, otherwise A contains the lower triangle. Output parameters: X - solution of a system. Array whose index ranges within [0..N-1]. Result: True, if the matrix is not singular. X contains the solution. False, if the matrix is singular (the determinant of the matrix is equal to 0). In this case, X doesn't contain a solution. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool smatrixsolve(ap::template_2d_array< amp::ampf<Precision> > a, const ap::template_1d_array< amp::ampf<Precision> >& b, int n, bool isupper, ap::template_1d_array< amp::ampf<Precision> >& x);

svd unit

Functions

rmatrixsvd

rmatrixsvd function

/************************************************************************* Singular value decomposition of a rectangular matrix. The algorithm calculates the singular value decomposition of a matrix of size MxN: A = U * S * V^T The algorithm finds the singular values and, optionally, matrices U and V^T. The algorithm can find both first min(M,N) columns of matrix U and rows of matrix V^T (singular vectors), and matrices U and V^T wholly (of sizes MxM and NxN respectively). Take into account that the subroutine does not return matrix V but V^T. Input parameters: A - matrix to be decomposed. Array whose indexes range within [0..M-1, 0..N-1]. M - number of rows in matrix A. N - number of columns in matrix A. UNeeded - 0, 1 or 2. See the description of the parameter U. VTNeeded - 0, 1 or 2. See the description of the parameter VT. AdditionalMemory - If the parameter: * equals 0, the algorithm doesn’t use additional memory (lower requirements, lower performance). * equals 1, the algorithm uses additional memory of size min(M,N)*min(M,N) of real numbers. It often speeds up the algorithm. * equals 2, the algorithm uses additional memory of size M*min(M,N) of real numbers. It allows to get a maximum performance. The recommended value of the parameter is 2. Output parameters: W - contains singular values in descending order. U - if UNeeded=0, U isn't changed, the left singular vectors are not calculated. if Uneeded=1, U contains left singular vectors (first min(M,N) columns of matrix U). Array whose indexes range within [0..M-1, 0..Min(M,N)-1]. if UNeeded=2, U contains matrix U wholly. Array whose indexes range within [0..M-1, 0..M-1]. VT - if VTNeeded=0, VT isn’t changed, the right singular vectors are not calculated. if VTNeeded=1, VT contains right singular vectors (first min(M,N) rows of matrix V^T). Array whose indexes range within [0..min(M,N)-1, 0..N-1]. if VTNeeded=2, VT contains matrix V^T wholly. Array whose indexes range within [0..N-1, 0..N-1]. -- ALGLIB -- Copyright 2005 by Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool rmatrixsvd(ap::template_2d_array< amp::ampf<Precision> > a, int m, int n, int uneeded, int vtneeded, int additionalmemory, ap::template_1d_array< amp::ampf<Precision> >& w, ap::template_2d_array< amp::ampf<Precision> >& u, ap::template_2d_array< amp::ampf<Precision> >& vt);

trfac unit

Functions

cmatrixlu
hpdmatrixcholesky
rmatrixlu
spdmatrixcholesky

cmatrixlu function

/************************************************************************* LU decomposition of a general complex matrix with row pivoting A is represented as A = P*L*U, where: * L is lower unitriangular matrix * U is upper triangular matrix * P = P0*P1*...*PK, K=min(M,N)-1, Pi - permutation matrix for I and Pivots[I] This is cache-oblivous implementation of LU decomposition. It is optimized for square matrices. As for rectangular matrices: * best case - M>>N * worst case - N>>M, small M, large N, matrix does not fit in CPU cache INPUT PARAMETERS: A - array[0..M-1, 0..N-1]. M - number of rows in matrix A. N - number of columns in matrix A. OUTPUT PARAMETERS: A - matrices L and U in compact form: * L is stored under main diagonal * U is stored on and above main diagonal Pivots - permutation matrix in compact form. array[0..Min(M-1,N-1)]. -- ALGLIB routine -- 10.01.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void cmatrixlu(ap::template_2d_array< amp::campf<Precision> >& a, int m, int n, ap::template_1d_array< int >& pivots);

hpdmatrixcholesky function

/************************************************************************* Cache-oblivious Cholesky decomposition The algorithm computes Cholesky decomposition of a Hermitian positive- definite matrix. The result of an algorithm is a representation of A as A=U'*U or A=L*L' (here X' detones conj(X^T)). INPUT PARAMETERS: A - upper or lower triangle of a factorized matrix. array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - if IsUpper=True, then A contains an upper triangle of a symmetric matrix, otherwise A contains a lower one. OUTPUT PARAMETERS: A - the result of factorization. If IsUpper=True, then the upper triangle contains matrix U, so that A = U'*U, and the elements below the main diagonal are not modified. Similarly, if IsUpper = False. RESULT: If the matrix is positive-definite, the function returns True. Otherwise, the function returns False. Contents of A is not determined in such case. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool hpdmatrixcholesky(ap::template_2d_array< amp::campf<Precision> >& a, int n, bool isupper);

rmatrixlu function

/************************************************************************* LU decomposition of a general real matrix with row pivoting A is represented as A = P*L*U, where: * L is lower unitriangular matrix * U is upper triangular matrix * P = P0*P1*...*PK, K=min(M,N)-1, Pi - permutation matrix for I and Pivots[I] This is cache-oblivous implementation of LU decomposition. It is optimized for square matrices. As for rectangular matrices: * best case - M>>N * worst case - N>>M, small M, large N, matrix does not fit in CPU cache INPUT PARAMETERS: A - array[0..M-1, 0..N-1]. M - number of rows in matrix A. N - number of columns in matrix A. OUTPUT PARAMETERS: A - matrices L and U in compact form: * L is stored under main diagonal * U is stored on and above main diagonal Pivots - permutation matrix in compact form. array[0..Min(M-1,N-1)]. -- ALGLIB routine -- 10.01.2010 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> void rmatrixlu(ap::template_2d_array< amp::ampf<Precision> >& a, int m, int n, ap::template_1d_array< int >& pivots);

spdmatrixcholesky function

/************************************************************************* Cache-oblivious Cholesky decomposition The algorithm computes Cholesky decomposition of a symmetric positive- definite matrix. The result of an algorithm is a representation of A as A=U^T*U or A=L*L^T INPUT PARAMETERS: A - upper or lower triangle of a factorized matrix. array with elements [0..N-1, 0..N-1]. N - size of matrix A. IsUpper - if IsUpper=True, then A contains an upper triangle of a symmetric matrix, otherwise A contains a lower one. OUTPUT PARAMETERS: A - the result of factorization. If IsUpper=True, then the upper triangle contains matrix U, so that A = U^T*U, and the elements below the main diagonal are not modified. Similarly, if IsUpper = False. RESULT: If the matrix is positive-definite, the function returns True. Otherwise, the function returns False. Contents of A is not determined in such case. -- ALGLIB routine -- 15.12.2009 Bochkanov Sergey *************************************************************************/
template<unsigned int Precision> bool spdmatrixcholesky(ap::template_2d_array< amp::ampf<Precision> >& a, int n, bool isupper);