Some of my reinforcement learning experimentations.
- GridWorld - the agent tries to find the fastest way to the goal
- GridWorld_ApproximateQLearning - the agent also tries to find the fastest way to the goal of a more complex map but the approach uses features to decide on the next action instead of the whole map layout.
- PacMan - the pacman agent tries to eat all the cheese on the map. You can load different maps and draw your own.
- ReinforcementLearning - the (fast and poorly written) library i wrote, just to be able to quickly experiment with reinforcement.