Last-mile delivery optimization: Leveraging electric vehicles and parcel lockers for prime customer service

Abstract

In last-mile delivery, offering home delivery services is one of the most costly and time-consuming approaches for a delivery company. Parcel lockers (PLs) are an alternative to mitigate operational costs, enabling self-service parcel collection at post offices, transportation hubs, and campuses. Additionally, companies like Amazon often prioritize specific customers, such as ‘Amazon Prime Members,’ by offering them same-delivery options. This study explores last-mile delivery with home-attended service via electric vehicles (EVs) and self-collection at PLs while serving high-priority prime customers via EV or PL. The problem is termed the selective electric vehicle routing problem with PLs (SEVRP-PL), finding EV routes for prime customers and delivering parcels to designated PLs for customer pickup while minimizing EVs’ route, usage, and PLs’ opening costs and maximizing the collected prizes. A novel mixed-integer linear programming model is developed. Also, an efficient problem-tailored large neighborhood search heuristic with simulated annealing criterion (solution acceptance/rejection with a probability) is proposed to tackle large instances. The sensitivity analysis and performance comparison of the methods are presented. Significant reductions in route and usage costs for EVs are achieved through the proposed multi-modal delivery while prioritizing prime customers; savings of 24% in EV routing cost and 22.50% in EV usage cost on average for various instances. These results indicate that the addressed multi-modal delivery system effectively utilizes the resources, e.g., EVs and PLs, mainly when fast service for prime customers is desired.