Armijo–Goldstein. More...
#include <FgArmijoLineSearch.h>
Public Member Functions | |
ArmijoLineSearch () | |
Default constructor. More... | |
void | Linesearch (NonlinearObjectiveFunction< Scalar > &objFunc, Scalar &fx, af::array &x, af::array &grad, Scalar &step, const af::array &drt, const af::array &xp) override |
Linesearches. More... | |
~ArmijoLineSearch () | |
Destructor. More... | |
virtual void | Linesearch (NonlinearObjectiveFunction< Scalar > &objFunc, Scalar &fx, af::array &x, af::array &grad, Scalar &step, const af::array &drt, const af::array &xp)=0 |
Linesearches. More... | |
Armijo–Goldstein.
Armijo–Goldstein condition, is a line search method to determine the maximum amount to move along a given search direction. It involves starting with a relatively large estimate of the step size for movement along the search direction, and iteratively shrinking the step size (i.e., "backtracking") until a decrease of the objective function is observed that adequately corresponds to the decrease that is expected, based on the local gradient of the objective function.
HmetalT, 10/06/2019.
Definition at line 45 of file FgArmijoLineSearch.h.
NeuralEngine::MachineLearning::ArmijoLineSearch< Scalar >::ArmijoLineSearch | ( | ) |
Default constructor.
Hmetal T, 11/06/2019.
|
inline |
|
overridevirtual |
Linesearches.
Hmetal T, 10/06/2019.
x | The af::array to process. |
searchDirection | The search direction. |
prob | [in,out] The prob. |
alpha0 | The alpha 0. |
Implements NeuralEngine::MachineLearning::ILineSearch< Scalar >.