A choice-based optimization framework for crowdsourced last-mile delivery
本文提出一个集成仿真与优化的框架，同时考虑包裹服务点选址、临时快递员任务接受行为和专业车队路径规划，利用哥本哈根真实数据验证了该框架能显著提高快递员参与率、减少专业车队工作量并提升系统效率。
Crowdsourced last-mile delivery is an emerging paradigm in urban logistics, offering a flexible approach to mitigating operational costs and urban congestion. However, its effectiveness depends on the interplay between three key factors: the strategic placement of parcel service points, the task acceptance behavior of occasional couriers, and the operation of a professional fleet alongside them. This paper addresses these factors simultaneously by developing an integrated simulation-optimization framework that links strategic planning and operational behavior. A key methodological contribution lies in combining a cost-aware facility location model for parcel placement, a behaviorally choice model for courier task acceptance, and a vehicle routing solver for the professional fleet. Using real parcel data and simulated passenger trips from Copenhagen, the results show that the coordinated framework substantially increases courier participation, reduces professional fleet routing workload, and improves overall system efficiency compared with realistic benchmark strategies. The analysis highlights how behavioral modeling and spatial optimization jointly enable cost-effective and scalable collaboration between crowds and fleets in urban delivery networks.