Safe-AutoSAC: AutoML-enhanced safe deep reinforcement learning for integrated energy system scheduling with multi-channel informer forecasting and electric vehicle demand response

Yang Li, Bingsong Zhao, Yuanzheng Li, Chao Long, Sen Li, Zhaoyang Dong and Mohammad Shahidehpour

Applied Energy, 2025, vol. 399, issue C, No S0306261925011985

Abstract: To coordinate renewable uncertainties and electric vehicle (EV) demand response, this paper proposes a safe deep reinforcement learning-based optimal scheduling approach for integrated energy systems (IESs). An optimal scheduling model for IESs is established to maximize net profit through an integrated demand response (IDR) strategy driven by a dynamic pricing mechanism. This strategy comprehensively coordinates conventional flexible loads and EV demand response, thereby unlocking the potential of flexible resources. The scheduling model is formulated as a Markov decision process and solved using a safety-assured adaptive delay policy network and automated machine learning-enhanced soft actor-critic (Safe-AutoSAC) algorithm, ensuring that the generated scheduling decisions comply with IES power flow constraints. To improve renewable and load prediction accuracy, a novel multi-channel Informer forecasting model is proposed with cross-channel attention and multi-scale temporal embeddings for capturing spatiotemporal dependencies and periodic patterns. Experimental results demonstrate that our approach is superior to the traditional soft actor-critic algorithm, achieving a 7.06 % improvement in net profit while maintaining operational safety. Our forecasting method achieves high accuracy and outperforms alternative methods. Moreover, the implementation of the proposed IDR strategies enhances the operational flexibility of the IES and results in a 44.35 % increase in net profit compared to the baseline scenario without demand response.

Keywords: Integrated energy system scheduling; Electric vehicle demand response; Multi-channel informer forecasting; Safe deep reinforcement learning; Integrated demand response; Dynamic pricing mechanism (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261925011985
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:399:y:2025:i:c:s0306261925011985

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2025.126468

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().