Improved Version of the Pong Game Using RL TD3

An improved version of the pong game.

This program uses the TD3 RL technique to train an agnet for the pong game.

It is the same as the original one developed by Johanna Pingel (a MATLAB staff and blogger) with the following changes:

1. Another state is added to the observation-state vector called q8 with the equation: q8 = abs(ball_x-paddle_x)

2. TD3 is used instead of DDPG

3. rng is added to use the same random noise at each time the agent is trained (for comparison)

Original code:

https://blogs.mathworks.com/deep-learning/2021/03/12/playing-pong-using-reinforcement-learning/