A High-Performance Scheduling Model For Electric Vehicle Battery Charging Using Multi-Objective Optimization

Abstract

The widespread uptake of electric vehicles (EVs) creates great opportunities for carbon emission savings but also brings new challenges to power grid stability, particularly the peak demand periods. EV charging scheduling is basically a multi-objective optimization problem which requires that the charging cost be minimized, the peak grid load be lessened, and user satisfaction be achieved by meeting target State of Charge (SOC) requirements. This article introduces a new Hybrid Memetic Adaptive Surrogate-Assisted NSGA-III (HMAS-NSGA-III) algorithm to mitigate the computational and scalability issues of large-scale real-time electric vehicle charging optimization. The method combines memetic local search for precise exploitation, surrogate-assisted modeling for minimum computational overhead, and adaptive NSGA-III for ensuring solution diversity over high-dimensional Pareto fronts. The suggested approach was compared with benchmark algorithms such as MOPSO and NSGA-II in realistic dynamic pricing and fleet scenarios. Experimental outcomes verify that HMAS-NSGA-III results in the minimum operational cost (8.47 USD), minimum peak load (142.78 kW), and least SOC deviation and takes only a runtime of 5.92 seconds. The algorithm has better convergence and solution quality compared to traditional approaches and is a feasible and scalable solution for intelligent EV charging management in smart grid scenarios.