“送取一体”模式下电商配送的路径与装载联合优化研究

Joint Optimization of Distribution Routing and Cargo Loading for E-Commerce Delivery in the “Pickup and Delivery” Mode

摘要: 同时送取货的车辆路线问题(VRPSDP)随着逆向物流的发展，已成为物流运输领域的高频研究方向。已有多项研究取得相关进展，现有文献缺乏对复杂需求客户三维装载方案的灵活性与配送路径优化的协同考量。为增强同时送取问题的装载与路径的协同作用，本文通过对客户的聚类分析与车辆空间的有效利用，达到了对确定需求场景下，车辆装载及路线规划的综合优化，提升了送取一体化问题的可用性，具有重要的现实意义。本文提出了对客户聚类、路径规划及装载方案优化的三阶段框架(CLR-VRPSDP)，以生成最佳的装载及配送路线，高效完成对多类型需求客户的需求满足。本文利用随机生成的中小规模测试集，进行了一系列实验，验证了本文所提方案的有效性。实验结果表明，该方法在有效性及资源利用上明显优于其他方法。

Abstract: With the development of reverse logistics, the Vehicle Routing Problem with Simultaneous Pickup and Delivery (VRPSDP) has become a frequent research focus in the field of logistics transportation. Although many studies have made relevant progress, existing literature lacks a synergistic consideration between the flexibility of three-dimensional loading schemes for customers with complex demands and the optimization of delivery routes. To enhance the synergy between loading and routing in the simultaneous pickup and delivery problem, this paper achieves comprehensive optimization of vehicle loading and route planning under deterministic demand scenarios through customer clustering analysis and effective utilization of vehicle space, thereby improving the applicability of the integrated pickup and delivery problem, which holds significant practical implications. This paper proposes a three-stage framework (CLR-VRPSDP) that integrates customer clustering, route planning, and loading scheme optimization to generate optimal loading and delivery routes, efficiently satisfying the demands of customers with multiple types of requirements. A series of experiments is conducted using randomly generated small- and medium-scale test sets to validate the effectiveness of the proposed approach. Experimental results demonstrate that this method significantly outperforms other approaches in terms of effectiveness and resource utilization.