UltimateTicTacToe

Play Ultimate Tic-Tac-Toe against an AI agent. The project utilizes Monte Carlo Tree Search (MCTS).

Demo

Features

✅ AI-powered Gameplay: Uses MCTS to make intelligent moves.
✅ Java & Python Integration: Communicates between Java and Python via Py4J.
✅ Fast Search & Decision Making: AI runs simulations to improve move selection.
✅ Customizable AI Levels: Adjustable difficulty by modifying tree search depth.
✅ Unit Tested: Covered with JUnit 5 tests.

MCTS Process:

1️⃣ Selection: Traverse the tree using Upper Confidence Bound (UCT).
2️⃣ Expansion: Add a new node if a promising move is found.
3️⃣ Simulation: Play random moves from the new node to estimate the outcome.
4️⃣ Backpropagation: Update node values based on the result.

Formula for UCT (used in MCTS decision-making):

$$UCT = (winScore / visits) + C * sqrt(log(totalVisits) / visits)$$

Installation & Dependencies

This project is built with Java, Maven, Python and JavaScript. To use it, clone the repository and build with Maven:

Requirements:

• Java 11+
• Maven
• Python 3.x

For Linux:

sudo apt-get install build-essential libgl1-mesa-dev
pip3 install pyqt5 pyqtwebengine
pip3 install pywebview
pip3 install pywebview[qt]

Clone and Build:

git clone https://github.com/T-Lak/UltimateTicTacToe.git

cd mcts_gui/

python -m venv .venv
source .venv/bin/activate  # Linux/macOS
.venv\Scripts\activate     # Windows

pip install -r requirements.txt

Usage

# Start the Java MCTS AI server
mvn compile exec:java -Dexec.mainClass="Main"

# Run the Python UI (if applicable)
cd mcts_gui/
python api.py

Project Structure

/src
 ├── main/java/
 │     ├── api/         # Handles external communication (Py4J, etc.)
 │     ├── board/       # Game board logic (GlobalBoard, MCTSBoard)
 │     ├── common/      # Utility classes (Move, Player)
 │     ├── mcts/        # Monte Carlo Tree Search implementation
 │          ├── tree/   # MCTS Node and Tree structure
 │     ├── Main.java    # Entry point for AI server
 ├── mcts_gui/          # (If applicable) Frontend & Python API
 │     ├── pages/       # UI components
 │     ├── static/      # Static assets
 │     ├── api.py       # Python-Java communication (Py4J)
 ├── test/java/         # Unit tests
 │     ├── board/       # Board-related tests
 │     ├── utilities/   # Helper tests

Testing

Run unit tests using Maven:

mvn test

Contributing

If you find any issues or optimizations, feel free to open an issue or submit a pull request!

License

This project is licensed under the MIT License.