Research on Orderly Charging Method of Electric Vehicles Based on Deep Reinforcement Learning

Abstract

Electric vehicle charging stations are vital for the integration of electric vehicles (EV) with the future energy internet. To minimize daily operating costs in the face of uncertain time-of-use pricing and user behavior, this paper proposes an ordered charging optimization strategy from the perspective of load aggregators. By modeling the optimization problem as a finite-horizon Markov Decision Process (MDP) and using a reward-penalty feedback function to extract high-quality experience, the optimal operating strategy can be obtained. Results show that the proposed strategy reduces daily operating costs, protects batteries, and meets the charging.