Here are some codes of Principles of Artificial Intelligence in SJTU
A Chinese Checker game.
We use an updated greedy algorithm when starting and also to end one game.
In order to play Chinese Checker, run python runGame.py.
Our agent is in agent.py.
Well it can only take the basic credits.
Much more to be done.
An environment of a grid world with punishment for traps, rewards for certain points and an exit. Implement RL to train the agent to try its best to avoid the punishment, get high scores as it can and then get out.
Run main.py to start training.
The result is shown in demo.png.
It is a famous game called Pacman in the env of gym. Actually we did not succeed in training the little pacman effectively. Hope we can complete the game some day.