Motor vehicles typically exhibit a “speed-varying range” (SVR) characteristic. For battery-powered electric vehicles (BEVs), the range diminishes at higher speed. This characteristic greatly impacts BEV operation for demanding commercial uses like express delivery, given their limited range and long recharge times. In view of the above, this article examines a new electric vehicle routing problem (VRP) that explicitly models BEVs’ SVR and considers the joint planning of BEV route, speed, and charging under stochastic traffic conditions. A deep reinforcement learning (DRL) approach that exploits the interdependence among the above three decision aspects is then developed to generate real-time policies. Experiments on hypothetical and real-world instances showcase that the proposed approach can efficiently find high-quality policies that effectively accommodate BEVs’ SVR.
Real-Time Planning of Route, Speed, and Charging for Electric Delivery Vehicles: A Deep Reinforcement Learning Approach
IEEE Transactions on Transportation Electrification ; 11 , 2 ; 7066-7082
2025-04-01
2105577 byte
Article (Journal)
Electronic Resource
English
Deep reinforcement learning based trajectory real-time planning for hypersonic gliding vehicles
SAGE Publications | 2024
|Route and charging planning for electric vehicles: a multi-objective approach
Taylor & Francis Verlag | 2025
|A Real Time Route Optimisation Model for Electric Delivery Vehicles
British Library Conference Proceedings | 1998
|