MPC4RL -- A Software Package for Reinforcement Learning based on Model Predictive Control

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Reinhardt, Dirk, Baumgärnter, Katrin, Frey, Jonathan, Diehl, Moritz, Gros, Sebastien
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915123595051008
author Reinhardt, Dirk
Baumgärnter, Katrin
Frey, Jonathan
Diehl, Moritz
Gros, Sebastien
author_facet Reinhardt, Dirk
Baumgärnter, Katrin
Frey, Jonathan
Diehl, Moritz
Gros, Sebastien
contents In this paper, we present an early software integrating Reinforcement Learning (RL) with Model Predictive Control (MPC). Our aim is to make recent theoretical contributions from the literature more accessible to both the RL and MPC communities. We combine standard software tools developed by the RL community, such as Gymnasium, stable-baselines3, or CleanRL with the acados toolbox, a widely-used software package for efficient MPC algorithms. Our core contribution is MPC4RL, an open-source Python package that supports learning-enhanced MPC schemes for existing acados implementations. The package is designed to be modular, extensible, and user-friendly, facilitating the tuning of MPC algorithms for a broad range of control problems. It is available on GitHub.
format Preprint
id arxiv_https___arxiv_org_abs_2501_15897
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MPC4RL -- A Software Package for Reinforcement Learning based on Model Predictive Control
Reinhardt, Dirk
Baumgärnter, Katrin
Frey, Jonathan
Diehl, Moritz
Gros, Sebastien
Systems and Control
In this paper, we present an early software integrating Reinforcement Learning (RL) with Model Predictive Control (MPC). Our aim is to make recent theoretical contributions from the literature more accessible to both the RL and MPC communities. We combine standard software tools developed by the RL community, such as Gymnasium, stable-baselines3, or CleanRL with the acados toolbox, a widely-used software package for efficient MPC algorithms. Our core contribution is MPC4RL, an open-source Python package that supports learning-enhanced MPC schemes for existing acados implementations. The package is designed to be modular, extensible, and user-friendly, facilitating the tuning of MPC algorithms for a broad range of control problems. It is available on GitHub.
title MPC4RL -- A Software Package for Reinforcement Learning based on Model Predictive Control
topic Systems and Control
url https://arxiv.org/abs/2501.15897