PAL should be able to learn how to optimally control a given system with an unknown and also slowly adapting second controller.
Optimal Control, Differential Games, Identification, Neural Networks, DDPG
What it does (when executing main):
- Simulates the training of two controllers
- Plots the result (You can make more information like the control variable u visible in the plot window)
- Prints some interesting information about the run
- Saves the result into a file