This C++ header-only library offers a framework for simulating a variety of board games, including Connect4 and Chinese Checkers. It is designed for reinforcement learning applications, enabling numerous playthroughs while emphasizing template programming and minimizing heap allocations.
The library introduces a unified representation of games via states and actions. In each state, the current player chooses an action from a set of legal moves, resulting in a transition to a new state. Upon reaching a terminal state, a reward is assigned based on the outcome of the playthrough.
For ease of integration with common machine learning frameworks, Python bindings are provided in a separate repository.
To integrate this library in your project, the recommended way is CMake's FetchContent module.
Add the following to your CMakeLists.txt file:
include(FetchContent)
FetchContent_Declare(
game
GIT_REPOSITORY https://github.com/jojolebarjos/board-game-simulator.git
GIT_TAG main
GIT_SHALLOW 1
)
FetchContent_MakeAvailable(game)The src and tests folders contain usage examples, which are not included when using FetchContent.
To compile and run them, use CMake directly on this repository:
git clone https://github.com/jojolebarjos/board-game-simulator.git
cd board-game-simulator
For single-config generators (e.g. MinGW Makefile):
cmake -S . -B build -G "MinGW Makefiles" -DCMAKE_BUILD_TYPE=Release
cmake --build build
ctest --test-dir build
For multi-config generators (e.g. Visual Studio):
cmake -S . -B build -G "Visual Studio 17 2022" -A x64 -DCMAKE_CONFIGURATION_TYPES=Release
cmake --build build --config Release
ctest --test-dir build --build-config Release