PyMCTS: Monte Carlo Tree Search for Python

What is PyMCTS?

PyMCTS is a Python implementation of Monte Carlo Tree Search, which is a heuristic technique for discrete decision processes, most notably in game AI and in planning.

PyMCTS aims for a simple, clean framework for MCTS so that particular variants, such as UCT (Upper Confidence Trees) can be easily worked in. PyMCTS will also have a basic toolkit for evaluating quality of search trees, and visualization tools to aid the user.

PyMCTS is in early development. Features are still in flux, and are being implemented. See 'Development status' below.

Requirements

PyMCTS requires Python 3.5.1 or later to run. Users are strongly urged to consider Continuum Analytics' Anaconda Python distribution, as it contains packages likely to be of interest to PyMCTS users and contributors.

Quick start

You can install the latest version from git:

$ pip3 install git+git://github.com/smallnamespace/pymcts.git

Development status and Issue Tracker

PyMCTS is an early work in progress; the API is still in flux and additional features are being discussed and implemented.

Please see open issues for development progress and upcoming features.

Contributing

Any help would be appreciated. Please reach out directly, or send pull requests.

License

PyMCTS is licensed under the Apache v2 License (see LICENSE).