Diffusion Model Predictive Control

Guangyao Zhou, Sivaramakrishnan Swaminathan, Rajkumar Vasudeva Raju, J. Swaroop Guntupalli, Wolfgang Lehrach +4 more
10/7/2024
cs.LGcs.AI

No official code found. Best implementation: hrh6666/Flexible-Locomotion-Learning-with-Diffusion-Model-Predictive-Control (24 stars, MIT, last commit 2026-04-18).

Abstract

We propose Diffusion Model Predictive Control (D-MPC), a novel MPC approach that learns a multi-step action proposal and a multi-step dynamics model, both using diffusion models, and combines them for use in online MPC. On the popular D4RL benchmark, we show performance that is significantly better than existing model-based offline planning methods using MPC (e.g. MBOP) and competitive with state-of-the-art (SOTA) model-based and model-free reinforcement learning methods. We additionally illustrate D-MPC's ability to optimize novel reward functions at run time and adapt to novel dynamics, and highlight its advantages compared to existing diffusion-based planning baselines.

View on arXivView PDF

Code Implementations(1)

hrh6666/Flexible-Locomotion-Learning-with-Diffusion-Model-Predictive-ControlCommunity64%
MIT

Official Implementation of Flexible Locomotion Learning with Diffusion Model Predictive Control

241Shell, PythonOct 3, 20256 months agoMIT

Similar Papers

Diffusion Predictive Control with Constraints

86% match1 repo632024

Diffusion Modulation via Environment Mechanism Modeling for Planning

86% match2026

Diffusion-Based Forecasting for Uncertainty-Aware Model Predictive Control

86% match2025

What Makes a Good Diffusion Planner for Decision Making?

84% match1 repo842025

Diffusion Controller: Framework, Algorithms and Parameterization

84% match2026

Diffusion Model is an Effective Planner and Data Synthesizer for Multi-Task Reinforcement Learning

84% match1 repo642023

Cite this paper

@article{zhou2024diffusion,
  title  = {Diffusion Model Predictive Control},
  author = {Guangyao Zhou and Sivaramakrishnan Swaminathan and Rajkumar Vasudeva Raju and J. Swaroop Guntupalli and Wolfgang Lehrach and Joseph Ortiz and Antoine Dedieu and Miguel Lázaro-Gredilla and Kevin Murphy},
  year   = {2024},
  eprint = {2410.05364},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2410.05364v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2410.05364.svg)](https://codeofpaper.com/paper/2410.05364)

Discussion