Real-time Locally Optimal Schedule for Electric Vehicle Load via Diversity-maximization NSGA-II

Abstract

Electric vehicles (EVs), as the distributed energy storage equipment, show great strengths to serve the grid system. Meanwhile, reasonable optimization for the charging time of EVs can also cut down the expense of customers. Thus, if considering the above aspects, the schedule of EV load refers to the multi-objective optimization. This paper presents a diversity-maximization NSGA-II to perform the multi-objective optimization by considering the grid load profile, charging cost of customers and battery degradation. Furthermore, in order to make this strategy more practical, a real-time locally optimal schedule is adopted by utilizing a flexible time scale. Case study illustrates that proposed DM-NSGA-II can effectively avoid the solutions being trapped in a relatively limited region so as to diversify the optimal results and provide trade-off solutions to decision-makers. Analysis of the locally optimal schedule shows that the variable time-scale can continuously involve the currently arrived EVs into the real-time optimization rather than relying on the forecasting data, so this strategy makes the schedule of EV load more practical without loss of the accuracy.