Electric Vehicle Routing for Emergency Power Supply with Deep Reinforcement Learning

To maintain telecom services even during power outages, maintaining the power of the base stations is essential. Here, we consider a solution where Electric Vehicles (EVs) go around to directly supply their power to the base stations whose power is continuously decreasing. The goal is to find EV routes that minimize both total travel distance and the number of downed base stations. In this paper, we formulate this routing as a new variant of the Electric Vehicle Routing Problem (EVRP) and propose a solver that combines a rule-based vehicle selector and a reinforcement learning-based node selector. We evaluate our solver on synthetic datasets and real datasets. The results show that our solver outperforms baselines in terms of the objective value and computational time. See https://ntt-dkiku.github.io/rl-evrpeps for details (full paper, code, visualization, etc).