Skip to content

Ultimate Tic-Tac-Toe game and agents. The following agents are implemented: random, Monte-Carlo, Min-Max, DQN

Notifications You must be signed in to change notification settings

JosselinSomervilleRoberts/UltimateTicTacToe-RL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

69 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

UltimateTicTacToe-RL

For more details about the project, please visit: https://josselinsomervilleroberts.github.io/

image

The goal is to design an AI using RL to play Ultimate Tic Tac Toe.

See the paper to learn the rules.

Requirements

pip install pygame

pip install numpy

pip install time

pip install gym

pip install torchvision

How to launch

You need to launch from the root of the folder the scripts in play_modes.

You can use agent_in_single_player_env.py to make 2 agent fight each other (one of them can be you).

The best agent is MinimaxPruningAgentSeveralRewards so try to beat him!


WARNING: You may experience some path issues. To solve this simply add the absolute path of this folder in the script like so

import sys
sys.path.append("C:\\Users\\Marie\\Organisation_Marie\\X\\3A\\INF 581 - Advanced machine learning\\Project\\UltimateTicTacToe-RL")

Enjoy !

How to recreate the results presented in the paper

You can launch play_modes/stats.py to make each agent fight against each other for several games in order to get statistics. However, this process takes dozen of hours so you need to be patient. To then visualize the figure presented in the paper, you can then use display_results.py with the values printed by the previous script.

About

Ultimate Tic-Tac-Toe game and agents. The following agents are implemented: random, Monte-Carlo, Min-Max, DQN

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages