Dynamic Sensing and Collaborative Optimization Algorithm for Adjustable Resources on the Power User Side Combining Edge Computing and Internet of Things

Abstract

To address the decentralized and dynamic characteristics of adjustable resources on the power user side, as well as the high sensing delay and low collaborative efficiency of traditional centralized architectures, this paper proposes a dynamic sensing and collaborative optimization algorithm combining edge computing with the Energy Internet of Things. An edge-terminal two-level sensing architecture is first constructed. Smart terminals collect electrical quantity, state quantity, and behavioral feature data on the user side, while edge nodes perform local real-time processing and preliminary data analysis. A dynamic sensing model based on two-dimensional clustering is then designed by integrating K-medoids and DBSCAN to jointly identify user behaviors and resource regulation capabilities. Finally, a multi-objective collaborative optimization model is established, and an improved particle swarm optimization algorithm is adopted to solve the Pareto optimal solution considering economy, stability, and environmental protection. Simulation results show that the algorithm reduces sensing delay to less than 55 ms for user scales up to 500 households, decreases the peak-valley load difference rate by 28.6%, and increases the renewable energy consumption rate by 11.3%, improving real-time regulation accuracy in distributed electromagnetic energy systems.