This is the code for the MCTS learning agent written in python using Seoul AI gym
You cann find a copy of the code at: https://github.com/chroadhouse/checkerAIDiss/edit/main/README.md
To do this you must install the seoul ai gym libarary from github: https://github.com/seoulai/gym
Or install via pip
pip3 install seoulai-gym
Once install simply
cd checkerAI
python3 Main.py
To play the game you have 3 options
- 1: To play against the agent yourself
- 2: watch the agent play against a random agent
- 3: Watch the agent play against a version of itself.
Once in you will be asked:
- How many games you want to play
- How big the search space of the agent will be
- Whether you want to increase the search space in the next game
- Whether you want to play some of the game at random (up to 170 moves)
When playing the game against the opponent use the numbered board (board map.png file)below:
which is also included in the in the file to make a move.
The input must be in the format of X,Y
for both the start and end position of the piece. If you do not enter a valid move then you will be asked to move again. If you try and move an opponent piece, one of your pieces will be moved randomly