A Two-Stage Preference-Aware Model for Electric Vehicle Spatiotemporal Charging Scheduling

Abstract

The popularity of electric vehicles (EVs) is driving the electrification of transportation. However, the high penetration of EVs may lead to road congestion and overload of the distribution network (DN). With the overall goal of optimising the spatiotemporal distribution of EV charging load and reducing the charging navigation cost, a two-stage spatiotemporal scheduling method considering different user preferences is proposed. In Stage I, a multi-agent cooperative framework and an improved Dijkstra algorithm are proposed. To effectively perceive user preferences, this method incorporates dynamic traffic flow, DN information and the value preferences of EV users to improve the charging station (CS) utility function, and sets the search area to reduce the time complexity of the Dijkstra algorithm. In Stage II, variable time-scale charging scheduling is performed for EVs with different charging requirements to optimise the temporal distribution of EV loads. The simulation case verifies the effectiveness of the two-stage EV spatiotemporal scheduling model. The results show that this method provides a low-cost and fast-response charging navigation solution, effectively alleviating traffic congestion and the peak-to-valley difference of DN is reduced by 23.4%.