felicialiu / RL-Tic-tac-toe

Reinforcement Learning for Tic-tac-toe

Training agent/running code

Use qlearninggame.py to train a Q-learning agent; sarsagame.py for a SARSA agent. Use rungame.py to play as a human against a computer. Command line arguments are explained inside the respective files. Requires dill to run.

Testing

Unit tests are provided in tictactoe_test.py and qlearn_test.py.