Optimal scheduling of electric vehicle aggregators based on sac reinforcement learning

Abstract

Aiming at the optimal scheduling problem of electric vehicle aggregators, a charge change load optimal scheduling strategy based on sac deep reinforcement learning is proposed. The strategy fully considers the users, power system and power market in the process of load regulation, and can realize the friendly interaction between electric vehicle and power system. Based on the establishment of the joint optimal scheduling framework of charging pile and exchange station, considering the multiple temporal and spatial characteristics of optimal scheduling of charging and exchange station, the optimal scheduling model of adjustable charging and exchange load in different scenarios is constructed, and the real-time scheduling scheme of grid connected charging and exchange load is solved based on sac algorithm. An example is given to verify the economy and efficiency of sac algorithm applied to the real-time optimal scheduling of charging and changing load of large-scale electric vehicles.