Optimization model and algorithm for coordinated delivery of autonomous delivery vehicles and electric vehicles with battery swapping consideration

• Published in: Computers & industrial engineering Vol. 207; p. 111265
• Main Authors: Mu, Lifeng, Zhang, Wenmin, Sugumaran, Vijayan, Li, Yongxin, Liu, Qigang, Xu, Wei
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2025
• Subjects: Battery swapping infrastructure; Collaborative delivery; Hybrid metaheuristics; Last-mile logistics; Two-echelon vehicle routing problem; Last-mile logistics; Hybrid metaheuristics; Battery swapping infrastructure; Collaborative delivery; Two-echelon vehicle routing problem
• ISSN: 0360-8352
• DOI: 10.1016/j.cie.2025.111265

•Novel last-mile distribution system with electric vehicles (EVs) & autonomous delivery vehicles (ADVs) for parcel deliveries.•Battery swapping stations (BSSs) strategically deployed in the distribution network to increase the driving range of EVs.•Problem formulated as a MILP and solved using a novel hybrid metaheuristic algorithm with fluctuating cooling schedule.•Sensitivity analyses have been performed for ADV speed and battery capacity, which reveal interesting insights.•Experiments across various instance sizes demonstrate algorithm's competitive performance in solving the proposed problem. This paper investigates a novel last-mile delivery framework involving electric vehicles (EVs) and autonomous delivery vehicles (ADVs) operating in a coordinated manner. The proposed model allows customers located in pedestrian-friendly zones to be served either directly by EVs or indirectly via ADVs dispatched from nearby delivery stations. Battery swapping stations (BSSs), located at predefined sites, offer a time-efficient alternative to conventional charging and are embedded into the EV routing decisions. A mixed-integer programming (MIP) model is developed to jointly optimize vehicle routes, customer–vehicle assignments, and EV battery swapping decisions, subject to time window, capacity, and cruising range constraints. To solve large-scale instances, a Hybrid Metaheuristic with Fluctuating Cooling Schedule (HM-FCS) is proposed, incorporating problem-specific operators and a novel fluctuating cooling mechanism to improve search robustness and solution quality. Computational experiments on problem-specific instances of varying scales validate the effectiveness and scalability of the proposed algorithm. Sensitivity analyses further examine how ADV operational parameters and customer spatial distributions affect overall system cost-efficiency and service feasibility.