Skip to content

PyMDPs

PyMDPs is a Python library for working with Markov Decision Processes (MDPs) and other related models like Partially Observable MDPs (POMDPs). It is designed to be easy to use and flexible, with a focus on providing a clean and intuitive API for defining and solving MDPs.

Features

  • Simple API: PyMDPs provides a simple and intuitive API for defining MDPs and solving them using a variety of algorithms.
  • Modular Design: PyMDPs is designed to be modular and extensible, making it easy to add new algorithms and features.
  • Well-Documented: PyMDPs comes with comprehensive documentation and examples to help you get started quickly.
  • Open Source: PyMDPs is open source and released under the MIT license, so you are free to use it for any purpose.

Getting Started

For an introduction to PyMDPs and a guide on how to get started, check out the Getting Started page.

To quickly install PyMDPs, you can use pip:

pip install pymdps

Documentation

The documentation is divided into the following sections:

  • Getting Started: A guide to getting started with PyMDPs. This section covers installation instructions and a simple example of creating and solving an MDP.
  • API Reference: The complete API reference for PyMDPs, including detailed descriptions of all classes and functions.
  • Reference: This contains detailed information about the problems, algorithms, models, and other concepts used in PyMDPs.