##OBJECTIVE OF THE EXPERIMENTAL ANALYSYS##

The experimental analysis that is based on the two enclosed Python script is
intended to validate the two main results of the paper:

1. The proposal of the Local SpatioTemporal Propagation Algorithm -
counterpart of Backpropation in time

2. The Hamiltonian Sign Flip Strategy


**PYTHON SCRIPTS

LSTP.py                 simulation on the described problem of tacking
                        in optimal control
			
PlotResults.py          plot of the results

They are two simple scripts that make use of numpy only. 


**LSTP.py**

It is a script which has been used for testing LSPT for tracking
artificially generated signals. You can choose the signal for the simulation
by setting the variable choice in the main. The variables used in the main
should be easily interpreted.

Variables which define the signal involved:

horizon = 21600
point_T = 43200
period=10
frequency = 50
n_points = point_T * period  # set the number of points

Variables which define  the initial values: weights, bias, state and the network architecture

n_neurons = 5
w_init = 5
b_init = 5
x_init = 1

Variables that define the tracking task

accuracy = 100
reg_w=0.1
reg_0w=1
reg_0x =1
lq = [w_init, b_init, x_init, accuracy, reg_w, reg_0w, reg_0x]


It is also possible to define an external input but in the presented tracking problem this
is not relevant and, therefore, we choose the zero signal
#
# Select the input
#
choice = 'zero'
#choice = 'constant'
#choice='sinusoindal'
#choice='g-sinusoindal'
#choice='gg-sinusoindal'
#choice='cosinusoidal'
#choice='square'
#choice='cos-constant'
#choice='grad-sim'
#choice: 'sin-constant'
#choice = 'hybrid'

Here you can choose the type of target

#
# Select the target
#
# choice = 'zero'
#choice = 'constant'
choice='sinusoindal'
#choice='g-sinusoindal'
#choice='gg-sinusoindal'
#choice='cosinusoidal'
#choice='square'
#choice='cos-constant'
#choice='grad-sim'
#choice: 'sin-constant'
#choice = 'hybrid'

In order to plot the results the file as saved on files that are visualized
by the PlotResults.py script


**PlotResults.py**

This is a simple script to visualize the results of LSTP.py on the basis of
the saved files.

