DynaPlex is a software library for solving Markov Decision Problems and similar models (POMDP, HMM) written primarily in C++20 with bindings for python. It supports deep reinforcement learning, approximate dynamic programming, classical parameterized policies, and exact methods based on policy and value iteration. Models in DynaPlex are written in C++, and exposed via a generic and vectorized interface.

DynaPlex focuses on solving problems arising in Operations Management: Supply Chain, Transportation and Logistics, Manufacturing, etc.

Cloning the Repository with Dependencies¶

When cloning the repository, it’s essential to also download the required submodules:

`bash git clone --recurse-submodules https://github.com/DynaPlex/DynaPlex.git `

if you did not recurse submodules, or if you use other tools for cloning repos, please somehow ensure that submodules (especially googletest) are available.

Contents¶