Electric vehicle charging station pricing optimization method based on deep reinforcement learning

The invention discloses an electric vehicle charging station pricing optimization method based on deep reinforcement learning, relates to the technical field of power system planning and operation, and adopts a model-free thought to formulate the charging price of an urban public charging station in a road-electricity coupling network. Firstly, a road-electricity coupling network operation intelligent agent is constructed, and the intelligent agent and a road-electricity coupling network operation environment carry out interactive learning and optimal pricing. And then constructing an efficient interaction environment based on the deep belief network, wherein the efficient interaction environment comprises a driving path and a charging decision of an electric vehicle cluster. And then a training acceleration mechanism considering the carbon emission intensity is provided, the carbon emission of each charging station is tracked and traced to guide the optimization direction of the intelligent agent, the low-ca