Optimal scheduling for electric vehicle charging and discharging via double Q-learning-based multi-objective particle swarm optimization

With the rapid development of electric vehicles (EVs), their dispatchable energy storage characteristics are strategically valuable for the low-carbon transition of new power systems. However, the existing scheduling methods are constrained by multi-objective coupling conflicts and are challenging to optimize collaboratively. This paper proposes a multi-objective optimal scheduling method for electric vehicle charging and discharging. First, a three-objective optimization model is constructed considering user cost, load fluctuation, and carbon emission reduction, incorporating time-sharing tariffs and the uncertainties of distributed renewable energy. Second, a collaborative optimization framework combining a double Q-learning algorithm (DQL) and multi-objective particle swarm optimization (MOPSO) is proposed. This framework dynamically adjusts the MOPSO parameters through a DQL algorithm enhanced by a dual-network decoupling mechanism, drives the MOPSO algorithm to generate a multi-objective Pareto frontier solution, and selects the optimal scheduling strategy based on the TOPSIS decision-making method. Finally, simulations are conducted in typical scenarios to demonstrate the superior performance of the DQL-MOPSO algorithm in addressing this problem. The analysis covers carbon emissions, load fluctuations, and charging costs before and after scheduling, as well as the impact of the optimal scheduling strategy on the energy structure of the distribution network, validating the feasibility of the scheduling scheme.