Low-Carbon Multi-Objective Optimization of Ordered Electric Vehicle Charging with Photovoltaic Integration viaAdaptive Dynamic Hierarchical Decoupling Planning

Abstract The coordination of electric vehicle (EV) charging schedules faces multiple challenges, including variations in grid load, unpredictable demand patterns, and the instability of renewable energy supplies. This study develops a multi-criteria optimization framework that considers the fluctuating characteristics of photovoltaic (PV) generation, EV travel patterns, energy storage system properties, and carbon emission constraints. A data-driven Adaptive Dynamic Hierarchical Decoupling Planning (ADHDP) algorithm is utilized to optimize the ordered charging strategy for EV. First, to address the stochasticity and seasonal variations of PV generation, an improved Long Short-Term Memory (LSTM) network is constructed for PV power prediction across multiple meteorological scenarios throughout the year. Compared to traditional methods such as Support Vector Machine Regression and Autoregressive Integrated Moving Average models, the proposed model significantly enhances forecasting accuracy and stability, achieving a reduction of approximately 12.7% in terms of Mean Absolute Percentage Error across all seasons. Second, a multi-objective optimization strategy based on ADHDP is designed by jointly considering charging demand uncertainty and PV output fluctuations, aiming to effectively control system carbon emissions, enhance energy utilization efficiency, and reduce user waiting times. Simulation results demonstrate that the ADHDP-based method exhibits superior performance in terms of economic efficiency, environmental friendliness, and user satisfaction, thereby demonstrating the viability and performance advantages of the proposed optimization strategy.