Sustainable EVs Intelligent Management Framework with Multi-Objective Differential Parrot Optimization

Abstract

The increasing penetration of Electric Vehicles (EVs) and rooftop photovoltaic (PV) systems introduces substantial operational complexities into radial distribution networks, including voltage instability, harmonic distortion, and uneven power flows. Existing methodologies often treat planning and control separately, lacking real-time adaptability under stochastic demand-generation patterns. To bridge this critical gap, this study proposes a novel Multi-Objective Differential Parrot Optimization (MODPO) algorithm integrated with a Hierarchical Reinforced Predictive Load Control (HRPLC) framework. The MODPO algorithm ensures optimal siting and sizing of EV charging stations and PV units by minimizing power loss, voltage deviation, and operational cost. Simultaneously, HRPLC coordinates dynamic G2V/V2G operations using a deep deterministic policy gradient (DDPG) agent embedded within a predictive MPC layer. Extensive simulations on an IEEE 69-bus system show a 60.38% cost reduction, 9.84 kW loss minimization, and voltage deviation contained to 4.06%, while 100% PV utilization is achieved under harmonic and voltage unbalance constraints. This unified optimization-control framework provides a scalable pathway for reliable, cost-efficient, and quality-assured operation of future distribution networks with high DER proliferation.