Electric vehicle charging planning in complex systems considering HARL dynamic pricing

Abstract

Abstract To address the challenges of spatio-temporal coordination and economic optimization for electric vehicles (EVs) charging infrastructure planning in energy-transportation coupled networks. we propose an innovative planning framework that integrates the HARL dynamic time-of-use (TOU) control mechanism, the small-world network model, and NSGA-III algorithm. Closed-loop model of user response and electricity price regulation is constructed based on an electricity price elasticity matrix to quantify the spatio-temporal displacement effects of peak, off-peak, and valley-hour electricity prices on charging demand. Real-time dynamic TOU regulation mechanism depending on HARL is designed. Coordinating the coupling relationship among electricity price adjustments, traffic flow and charging demand ensures real-time matching between electricity price signals and the dynamic characteristics of the energy-transportation system; By exploiting the local clustering and long-range connectivity characteristics of small-world networks, the spatial layout of charging stations is optimized to balance accessibility and network resilience; Multi-objective coordinated optimization is achieved via the NSGA-III algorithm, aiming to maximize station utility and operating revenue and minimize total travel time and investment costs. Simulation results demonstrate that the proposed framework significantly enhances the reliability and operational efficiency of the charging network. Compared with traditional fixed TOU strategies, it performs better in load leveling, revenue improvement and user cost control, thus validating its feasibility in integrated energy-transportation systems.