Containerized framework for building control performance comparisons: Model predictive control vs deep reinforcement learning control

Yangyang Fu, Shichao Xu, Qi Zhu, Zheng O'Neill

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Scopus citations

Abstract

While both model predictive control (MPC) and deep reinforcement learning control (DRL) have shown significant energy cost savings for building systems, there is a lack of in-depth quantitative study on the comparison between the two. One major obstacle is the lack of a holistic evaluation environment for the building community and the control community to integrate their expertise in studying both model-based and learning-based control methods. To address this challenge, this paper presents a scalable containerized software framework for building control performance comparisons, with a special focus on enabling both optimal model-based control and deep learning-based control. The framework provides a standardized building environment for the control community to benchmark different advanced control strategies, and a flexible software architecture for the building community to standardize their own customized building environments. A preliminary performance comparison of MPC and DRL on a single zone building is performed in the case study. Both MPC and DRL can outperform the rule-based baseline controllers in terms of reducing energy cost and maintaining thermal discomfort. DRL can outperform MPC after a long training time with a predefined reward policy.

Original languageEnglish (US)
Title of host publicationBuildSys 2021 - Proceedings of the 2021 ACM International Conference on Systems for Energy-Efficient Built Environments
PublisherAssociation for Computing Machinery, Inc
Pages276-280
Number of pages5
ISBN (Electronic)9781450391146
DOIs
StatePublished - Nov 17 2021
Event8th ACM International Conference on Systems for Energy-Efficient Built Environments, BuildSys 2021 - Virtual, Online, Portugal
Duration: Nov 17 2021Nov 18 2021

Publication series

NameBuildSys 2021 - Proceedings of the 2021 ACM International Conference on Systems for Energy-Efficient Built Environments

Conference

Conference8th ACM International Conference on Systems for Energy-Efficient Built Environments, BuildSys 2021
Country/TerritoryPortugal
CityVirtual, Online
Period11/17/2111/18/21

Funding

The research reported in this paper was supported by the Building Technologies Office at the U.S. Department of Energy through the Emerging Technologies program under award number DE-EE0009150.

Keywords

  • OpenAI-Gym
  • building energy and control system
  • deep reinforcement learning
  • model predictive control

ASJC Scopus subject areas

  • Information Systems
  • Renewable Energy, Sustainability and the Environment
  • Electrical and Electronic Engineering
  • Building and Construction
  • Computer Networks and Communications
  • Architecture

Fingerprint

Dive into the research topics of 'Containerized framework for building control performance comparisons: Model predictive control vs deep reinforcement learning control'. Together they form a unique fingerprint.

Cite this