Abstract: In order to efficiently manage the charging process of electric vehicles and ensure the stability of the power grid, a research on collaborative optimization scheduling of electric vehicle charging stations based on reinforcement learning algorithm is proposed. By analyzing the charging mode of electric vehicles at charging stations, it can be concluded that the operation of charging stations is closely related to the cost of electricity consumption. Considering the uncertainty of the operating mode and charging demand of electric vehicles, construct an operating model for electric vehicle charging stations. Based on the above charging station operation model, combined with the Markov decision process in reinforcement learning algorithm, a collaborative optimization scheduling design is carried out. At the same time, through deep reinforcement learning algorithm, the robustness model and opportunity model are solved, and a more reasonable scheduling strategy is formulated, which helps to reduce charging costs. The experimental results show that the collaborative optimization scheduling method for electric vehicle charging stations based on reinforcement learning algorithms can improve economic benefits and renewable energy utilization, verifying the feasibility and effectiveness of reinforcement learning algorithms in charging station scheduling.

Key words: collaborative scheduling; electric vehicles; collaborative optimization; car charging station; reinforcement learning algorithm

摘要： 为高效管理电动汽车的充电过程，保证电网的稳定性，提出基于强化学习算法的电动 汽车充电站协同优化调度研究。通过对充电站的电动汽车充电模式进行分析，可知充电站的 运行与用电成本息息相关。考虑电动汽车的运行模式和充电需求的不确定性，构建电动汽车 充电站的运行模型。基于上述充电站运行模型，结合强化学习算法中的马尔可夫决策过程， 对其进行协同优化调度设计，同时，通过深度强化学习算法，对鲁棒性模型和机遇性模型进 行求解，进而制定出更加合理的调度策略，有助于降低充电成本。实验结果表明，基于强化 学习算法的电动汽车充电站协同优化调度方法能够提升经济效益和可再生能源利用率，验证 了强化学习算法在充电站调度中的可行性和有效性。

关键词: 协同调度；电动汽车；协同优化；汽车充电站；强化学习算法