Electric vehicle charging navigation strategy in coupled smart grid and transportation network: A hierarchical reinforcement learning approach

• DOI: 10.1016/j.ijepes.2024.109823
• Publication Date: 2024-01-31T23:54:57Z
• Authors: Changxu Jiang; Longcan Zhou; J.H. Zheng; Zhenguo Shao
• Keywords: Charging station

Most of the existing electric vehicle (EV) charging navigation methods do not simultaneously take into account destination optimization and path planning. Moreover, they are unable to provide online real-time decision-making under a variety uncertain factors. To address these problems, this paper first establishes bilevel stochastic model for EV considering various uncertainties, then proposes an method based on hierarchical enhanced deep Q network (HEDQN) solve above in real-time. The proposed HEDQN contains two networks, which utilized optimize route EVs, respectively. Finally, is simulated validated urban transportation networks. simulation results demonstrate that compared with Dijkstra shortest algorithm, single-layer reinforcement learning traditional algorithm can effectively reduce total cost vehicles realize vehicles, shows excellent generalization ability scalability.