Gomoku AI Project

Overview

This repository contains a project on Gomoku (Five in a Row), a strategic board game often used to explore Artificial Intelligence (AI) techniques.
The project includes a playable version of Gomoku in Python, as well as documentation and analysis resources related to the game and its AI implementations.

Files

1. gomoku.py

• Python implementation of the Gomoku board game on a 15×15 grid.
• Implements basic rules of Gomoku, including victory conditions (five stones in a row horizontally, vertically, or diagonally).
• Supports:
  • Human vs AI gameplay
  • Special rule Long Pro variant, which balances the advantage of the first player (Black).
  • Player input and text-based board display in the console.
• Includes placeholders for Minimax and Alpha-Beta pruning functions to extend the AI decision-making.

2. gomoku_analysis.pdf

• A research-based analysis of Gomoku strategies and AI approaches.
• Key contents:
  • Game patterns (victories, attacks, threats)
  • Classical AI approaches (Minimax, Alpha-Beta pruning)
  • Advanced methods such as Monte Carlo Tree Search and Deep Learning
  • Discussion of "Wine", a strong AI Gomoku player from the Gomocup tournament
• This document serves as a theoretical foundation to better understand how to design and improve an AI player for Gomoku.

3. IA-Affrontement_entre_IA.pdf

• A project assignment document (in French) describing the rules, requirements, and objectives of implementing a Gomoku AI:
  • Game rules and Long Pro variant
  • Requirements for the AI (must use Minimax with Alpha-Beta pruning)
  • AI should be able to play against a human within a 5-second response time
  • Gameplay displayed in a text-based format
  • Final evaluation based on the AI’s correct functionality and ability to play multiple games without errors
• Also includes instructions for hosting and testing on Google Colab.

Getting Started

Requirements

Only Python 3.x needed.

How to Play

1. Run the Python script:

   python gomoku.py

2. Choose your color:
   • B for Black (goes first, forced to start at H7 in Long Pro variant)
   • W for White
3. Enter moves in the format:

   A0  (row A, column 0)
   H7  (row H, column 7)
   

4. The game ends when:
   • A player places five stones in a row
   • Or the board is full (draw).

References

• Gomoku analysis paper (arXiv) – theoretical foundation for AI design.
• Assignment guidelines in IA-Affrontement_entre_IA.pdf.