Saved in:
Bibliographic Details
Main Authors: Zhao, Yunyi, Zhang, Wei, Xiang, Cheng, Du, Hongyang, Niyato, Dusit, Gao, Shuhua
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.16867
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909701327814656
author Zhao, Yunyi
Zhang, Wei
Xiang, Cheng
Du, Hongyang
Niyato, Dusit
Gao, Shuhua
author_facet Zhao, Yunyi
Zhang, Wei
Xiang, Cheng
Du, Hongyang
Niyato, Dusit
Gao, Shuhua
contents This paper introduces DiffCarl, a diffusion-modeled carbon- and risk-aware reinforcement learning algorithm for intelligent operation of multi-microgrid systems. With the growing integration of renewables and increasing system complexity, microgrid communities face significant challenges in real-time energy scheduling and optimization under uncertainty. DiffCarl integrates a diffusion model into a deep reinforcement learning (DRL) framework to enable adaptive energy scheduling under uncertainty and explicitly account for carbon emissions and operational risk. By learning action distributions through a denoising generation process, DiffCarl enhances DRL policy expressiveness and enables carbon- and risk-aware scheduling in dynamic and uncertain microgrid environments. Extensive experimental studies demonstrate that it outperforms classic algorithms and state-of-the-art DRL solutions, with 2.3-30.1% lower operational cost. It also achieves 28.7% lower carbon emissions than those of its carbon-unaware variant and reduces performance variability. These results highlight DiffCarl as a practical and forward-looking solution. Its flexible design allows efficient adaptation to different system configurations and objectives to support real-world deployment in evolving energy systems.
format Preprint
id arxiv_https___arxiv_org_abs_2507_16867
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Diffusion-Modeled Reinforcement Learning for Carbon and Risk-Aware Microgrid Optimization
Zhao, Yunyi
Zhang, Wei
Xiang, Cheng
Du, Hongyang
Niyato, Dusit
Gao, Shuhua
Machine Learning
Artificial Intelligence
This paper introduces DiffCarl, a diffusion-modeled carbon- and risk-aware reinforcement learning algorithm for intelligent operation of multi-microgrid systems. With the growing integration of renewables and increasing system complexity, microgrid communities face significant challenges in real-time energy scheduling and optimization under uncertainty. DiffCarl integrates a diffusion model into a deep reinforcement learning (DRL) framework to enable adaptive energy scheduling under uncertainty and explicitly account for carbon emissions and operational risk. By learning action distributions through a denoising generation process, DiffCarl enhances DRL policy expressiveness and enables carbon- and risk-aware scheduling in dynamic and uncertain microgrid environments. Extensive experimental studies demonstrate that it outperforms classic algorithms and state-of-the-art DRL solutions, with 2.3-30.1% lower operational cost. It also achieves 28.7% lower carbon emissions than those of its carbon-unaware variant and reduces performance variability. These results highlight DiffCarl as a practical and forward-looking solution. Its flexible design allows efficient adaptation to different system configurations and objectives to support real-world deployment in evolving energy systems.
title Diffusion-Modeled Reinforcement Learning for Carbon and Risk-Aware Microgrid Optimization
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2507.16867