抄録
Distributed energy management systems are increasingly challenged by the need to achieve low-latency decision-making, efficient coordination of heterogeneous energy resources, and secure distributed intelligence under high renewable energy penetration. This review systematically examines the synergistic integration of edge computing and machine learning as a unified technological paradigm for next-generation distributed energy management, to elucidate its underlying collaborative mechanisms, technological evolution, and future development pathways. A comprehensive analytical framework is established from three complementary dimensions—system architecture, intelligent algorithms, and collaborative operation—and recent advances are synthesized across four key research domains: edge–cloud collaborative architectures, edge-oriented machine learning prediction, edge intelligence-enabled distributed scheduling, and federated learning for privacy-preserving optimization. The analysis demonstrates that edge computing substantially enhances real-time system responsiveness by migrating computation and decision-making from centralized cloud platforms to distributed edge nodes, while machine learning significantly improves renewable generation forecasting, load prediction, and adaptive scheduling capabilities. Their deep integration enables latency-aware optimization, efficient utilization of distributed energy resources, and privacy-preserving collaborative intelligence, thereby establishing a scalable foundation for intelligent energy management in resource-constrained environments. The review further identifies four critical research bottlenecks, including heterogeneous edge resource coordination, cross-scale optimization of integrated multi-energy systems, lightweight and interpretable intelligence, and trustworthy cyber-secure operation. By integrating architectural evolution, algorithmic intelligence, and system-level collaboration into a unified analytical framework, this review clarifies the paradigm shift from cloud-centric management to edge-intelligent distributed coordination and establishes a systematic roadmap for the design, optimization, and deployment of next-generation distributed energy management systems.
