ainaherimam/fanorona-cpp

An implementation of the Monte Carlo Tree Search (MCTS) algorithm for the malagasy board game: Fanorona with C++ 20 ## ▶️ Run the project on colab via this Link # Fanorona on C++20 with MCTS A project that implements the Monte Carlo Tree Search (MCTS) algorithm for the traditional board game of Fanorona on C++20. --- ## 📖 About Fanorona Fanorona is a centuries-old strategy board game originating from Madagascar. It is played on a 9x5 grid (fanorona 9) or a 5x5 grid (fanorona 5) and challenges players to capture their opponent's pieces using a variety of strategic moves. The game is renowned for its depth and complexity, making it an excellent candidate for AI research. --- ## 🎯 Project - Implement a simple interface for the Fanorona game - Implement the **Monte Carlo Tree Search (MCTS)** algorithm to play Fanorona. - Develop an algorithm capable of making intelligent, adaptive decisions. --- ## 🚀 Features - **Fanorona Game Logic**: The game board is represented programmatically, supporting all valid moves and captures. - **MCTS Algorithm**: To calculate optimal moves. - **Interactive Gameplay**: Play against the AI (Still in dev) - **Learn from SelfPlay**: AlphaZero style implementation (Still in dev) --- ## 🛠️ Getting Started ## Prerequisites - **C++20 Compiler** (GCC 10+, Clang 10+, MSVC 2019+) - **CMake 3.20+** - **Libtorch** ## Installation ```sh # Clone repository git clone github.com && cd fanorona-cpp # Create build directory mkdir build && cd build # Configure and build cmake -DCMAKE_PREFIX_PATH=${LIBTORCH_PATH} .. make -j$(nproc) ``` ## 🛠️ To do - Outer loop of Self-play - Parallelization of MCTS ## 🌟 Acknowledgments - Inspired by the board game Fanorona Enjoy playing and exploring the strategies of Fanorona!