This is a project aiming to explore the potential of reinforcement learning techniques combined with genetic algorithms as explained in the paper about the Neuro Evolution of Augmenting Topologies (link: https://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf) when applied to "simple" goal based classic games, namely, Snake.
This project includes:
- A full implementation of the Snake game, using PyGame.
- A setup of the learning algorithm, helped with the neat-python library's implementation of it.
- A training mode and a testing mode.
Preview of the training phase: