Real-time recommendation algorithm of battery swapping stations for electric taxis

This paper proposes a real-time algorithm that recommends battery swapping stations (BSSs) to electric taxis (ETs) that need their batteries swapped. The algorithm takes into consideration available batteries at BSSs, driving ranges of ETs and the current traffic conditions, etc, in order to avoid long queues at BSSs. We consider a basic model that assumes past decisions are perfectly executed, and formulate an optimal ET-to-BSS assignment problem. This problem is an integer program and therefore hard to scale. For real-time implementation, we approximate the optimal assignment problem by a repeated potential game and propose an iterative best response algorithm to compute ET-to-BSS assignments. Preliminary numerical results suggest that our heuristic algorithm solves the optimal assignment problem approximately.