/RL-tic-tac-toe

Tic tac toe implementation using reinforcement learning.

Primary LanguageJava

Tic Tac Toe Implementation using RL

Tic tac toe implementation using RL espacially using expected SARSA algorithm. I tested this app in Windows 10. I used swing library in java standard library to visualize game board.

Requirements

  • maven
  • jdk 1.8 or higher

Build

  1. Go to the root directory of this project.
  2. Open terminal and type mvn package. Make sure you have target directory in this project directory.
  3. Type java -jar ./target/tic-tac-toe-1.0-SNAPSHOT.jar
  4. Make sure you can play game now.

Training of RL Agents

If you want to train your agent from the begining, you can delete state_value.bin file, and play game. You should do train agent many times to get better performance.