Skip to content

We compare different policies for the checkers game using reinforcement learning algorithms.

Notifications You must be signed in to change notification settings

salimandre/Monte-Carlo-Tree-Search-for-checkers-game

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

34 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Monte-Carlo Tree Search for checkers game

Installation

These scripts ran on Mac with python 3.7.6 using only Turtle and Numpy packages.

Content

  • Checkers game with graphical interface with 2 possible rules: with and without forced take (jump)
  • MCTS model using the following Upper Confidence Bound (UCB) formula:

    MCTS model can be customized with reward designs, thinking time and with option to use or not transfer of knowledge (MCTS player reuses relevant branches from former built tree according to current boardgame). It is also possible to visualize the tree with number of visits and average reward at each state.

  • 3 Different modes: MCTS vs MCTS, User vs MCTS, MCTS vs Random

MCTS vs MCTS

demo: 0.5s MCTS (black) vs 1s MCTS (red)

User vs MCTS

MCTS vs Random

demo: 0.5s MCTS (black) vs Random player (red)

Releases

No releases published

Packages

No packages published

Languages