iCharging: A Game-Based Decentralized Nash Equilibrium Approach for Shared EV Charging

Shared Electric Vehicle (EV) charging emerges as a mutually beneficial paradigm where EV charging hosts can earn income by renting out their home EV chargers with autonomous pricing, and EV users gain access to a broader range of charging options. However, existing works on EV charging scheduling either propose global optimal solutions that may conflict with users' individual interests, or fail to fully consider the diverse demands of EV users and the varying characteristics of EV hosts. In this paper, we propose iCharging, an efficient decentralized Nash equilibrium approach for shared EV charging using weighted congestion game. In particular, each EV user only follows the strategy to minimize its user-specific cost while accounting for the congestion incurred by other users. By proving that the formulated game possesses at least one pure Nash equilibrium (PNE), we devise an optimized parallel batch-based best response dynamics algorithm to steer users' strategy profiles towards a PNE in linear time complexity for each individual variable. Through extensive evaluations using real-world datasets from OpenStreetMap and ACN-Data, we demonstrate that iCharging can converge to a PNE time-efficiently in EV host-user matching while integrating users' specific preferences and dynamic congestion levels, with only a 3.34 % increase in the average total cost compared to the global optimal solution.