Dynamic Multi-Objective Optimization for Smart Grid Stability Under Electric Vehicles Charging Demand Using EvoGrid-Optimizer

Abstract

Electric vehicles (EVs) are increasingly considered a significant challenge to the stability of smart grids as they are integrated into urban distribution systems. Stochastic load variations are introduced by uncoordinated EV charging, leading to voltage distortion, transformer overloading, and increased power losses. Existing optimization methods, such as Particle Swarm Optimization (PSO) and Model Predictive Control (MPC), are characterized by limited real-time adaptability, high computational cost, and insufficient scalability under uncertain operating conditions. To address these limitations, the EvoGrid-Optimizer is proposed as a novel evolutionary algorithm for the dynamic multi-objective optimization of EV charging schedules in smart grids. The proposed framework simultaneously optimizes operational cost, voltage deviation, and peak load demand using a weighted fitness formulation, while grid operational constraints are satisfied, and is evaluated on a simulated IEEE 33-bus distribution network with stochastic EVs load and renewable energy sources. The simulation results demonstrate that superior performance is achieved by the EvoGrid-Optimizer across all key performance indicators when compared to baseline uncoordinated charging, PSO, and MPC. Specifically, a reduction in operational cost of 22.4%, a voltage deviation of 0.037 p.u., a peak load of 1480 kW, and a load balancing efficiency of 93.8% are achieved.