Playing Othello(Reversi) By Reinforcement Learning

Introduction

This is a simple application that learns to play Othello by reinforcement learning.

TD(0) is used to evaluate a policy.

Value approximation function is based on n-tuple network introduced in Wojciech's paper.

Quick Start

Run python tdl.py to learn a policy by self-play.

Edit config/config.ini to setup players and run python run.py to play Othello in command line.

Or you can try the simple web app:

• Run npm install && npm run build in web/ui.
• Install gevent and flask: pip install gevent flask
• Run python run_server.py
• Open http://localhost:44399/othello and play!

Reference

• Jaśkowski, Wojciech (2014). Systematic n-tuple networks for othello position evaluation. ICGA Journal, 37(2), 85–96.

• Sutton, R. S., & Barto, A. G. (1998). Reinforcement learning: an introduction. : MIT press Cambridge.