Order fulfillment in omni-channel (OC) drug retailing necessitates the integrated optimization of order allocation and delivery routing. This integration is particularly challenging, as it requires collaboration among multiple pharmacies to fulfill online orders. To address this issue, we propose a novel variant of the pickup and delivery problem, namely the (UM-PDPTW). This problem is distinguished by the incorporation of item-level allocation decisions that determine the sourcing pharmacy for each item in an order into vehicle routing decisions, together with item-flow decisions under multiple pickups per order. The UM-PDPTW is computationally intractable using existing algorithms or solvers, as it incorporates three interrelated NP-hard sub-problems: item allocation for multi-item orders, multi-commodity flow, and delivery routing. To solve this complex problem, we develop a decomposition-iteration-based algorithm, the (IAR-CFM). IAR-CFM leverages the partial decoupling between item allocation and delivery routing decisions by duplicating the multi-commodity flow variables. It iteratively solves the sub-problems using tailored algorithms and coordinates them into a feasible solution to the primal problem. This coordinated solution is then further improved via an improved variable neighborhood search algorithm. Computational results indicate that IAR-CFM achieves optimal solutions for small-scale instances and near-optimal solutions for large-scale cases, surpassing the upper bounds provided by CPLEX by up to 41.96%. A comprehensive sensitivity analysis based on real-world omni-channel drug retailing scenarios yields managerial insights on delivery time commitments, capacity constraints, assortment compositions, and stocking modes. These findings offer practical guidance for optimizing operations in drug retailing and other similar fulfillment and delivery operations, especially those involving multiple fulfillment centers and synchronized visits.Unpaired Multi-Pickup and Delivery Problem with Time WindowsIterative Allocation and Routing algorithm with Commodity Flow Matching
<- Back to list
Collaborative order fulfillment in omni-channel retailing: Unpaired multi-pickup and delivery problem with time windows
Annals of Operations Research
operationsoperations/supply-chainanalyticsanalytics/optimizationmethod
摘要
本文研究了全渠道药品零售中的协作订单履行问题，提出了一个新颖的未配对多取货与带时间窗配送问题（UM-PDPTW）。为解决这一NP难问题，开发了基于分解迭代的商品流匹配迭代分配与路径算法（IAR-CFM）。计算结果表明，该算法在小规模实例中可获得最优解，在大规模案例中可获得接近最优解，性能优于CPLEX求解器。基于实际场景的敏感性分析为配送时间承诺、能力约束、品类组合和库存模式提供了管理启示。