Policy-Prior-Guided Generative Coordination for Resilient Energy Management in Cyber-Physical Distributed Systems
IEEE Transactions on Industrial Cyber-Physical Systems · 2026 · 4 · 455-469
DOI: 10.1109/ticps.2026.3693245 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Industrial Cyber-Physical Systems, 2026. Verfasst von Man Zhou und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 20 Literaturangaben. Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
The dynamic network of heterogeneous and mobile energy nodes, such as drones, electric vehicles, and ground energy stations, forms the basis of air–ground collaborative cyber-physical energy systems. It is difficult to optimize system energy efficiency and economic operation while maintaining resilience under information and physical disturbances. The problem becomes more complex under strong spatiotemporal coupling, limited local information, and non-ideal communication. This paper proposes a strategy-prior-guided generative collaborative energy management framework, named PG-TSCD. The framework adopts a two-stage structure-aware decision process. At the system level, a spatiotemporal graph-based conditional diffusion model captures the dynamic coupling among nodes and global safety constraints. It describes the feasible scheduling space under complex constraints and generates a coordinated discharge and recharge power prior sequence. At the node level, each agent combines the global prior with incomplete local observations through a lightweight adaptive reconstruction module. It then produces executable scheduling actions that remain robust under disturbances. In this way, distributed agents achieve system objectives through coordinated decision making. Theoretical analysis shows that PG-TSCD approximates centralized optimal scheduling with a bounded performance gap and ensures stable distributed learning. Experiments show that the model converges in about 85 training rounds, reaches a peak energy efficiency of 3.5, and maintains a task completion rate above 93% under high load. The performance remains stable under 300 ms communication delay and 12% packet loss.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Zhou, M., & Che, X. (2026). Policy-Prior-Guided Generative Coordination for Resilient Energy Management in Cyber-Physical Distributed Systems. IEEE Transactions on Industrial Cyber-Physical Systems, 4, 455-469. https://doi.org/10.1109/ticps.2026.3693245
Vancouver
Zhou M, Che X. Policy-Prior-Guided Generative Coordination for Resilient Energy Management in Cyber-Physical Distributed Systems. IEEE Transactions on Industrial Cyber-Physical Systems. 2026;4:455-469. doi:10.1109/ticps.2026.3693245
BibTeX
@article{Zhou2026,
author = {Zhou, Man and Che, Xin},
title = {Policy-Prior-Guided Generative Coordination for Resilient Energy Management in Cyber-Physical Distributed Systems},
journal = {IEEE Transactions on Industrial Cyber-Physical Systems},
year = {2026},
volume = {4},
pages = {455--469},
doi = {10.1109/ticps.2026.3693245}
}
EUVIMED