A neat-python implementation on the simple pong game
Training simulation on a population of 50 for 2 generations
pip install neat-python
pip install pygame
For a detailed concept read the paper
In my approach there is a Player class which represents either a player and the right or on the left.
For each genome in the main evaluation loop there are lists of same length containing a complete gamestate.
If a genome does well in it's gamestate (in this case if it pongs the ball), it's fitness increases.
The NN is given the y-position of the player and the y-position of the ball and outputs a sigmoid activation,
which then decides wether the player should move up, down or stay where it is.