E-Commerce Order Fulfillment Problem with Limited Time Window

• 成果类型: Article; Early Access
• 署名作者: Zhou, Quan; Gumus, Mehmet; Miao, Sentao
• 署名单位: McGill University; University of Colorado System; University of Colorado Boulder
• 刊物名称: OPERATIONS RESEARCH
• ISSN/ISSBN: 0030-364X
• DOI: 10.1287/opre.2023.0453
• 发表日期: 2025
• 关键词: decisions; bounds

摘要

We explore the optimization of the middle-mile fulfillment process in the context of e-commerce. In collaboration with a prominent e-commerce retailer in North America specializing in electronics and computer products, we develop a stochastic optimization problem to demonstrate how an efficient middle mile can alleviate strain on the critical last mile, leading to cost reduction and improved performance. First, we prove that the optimal policy is of a state-dependent threshold type. However, computing such thresholds is notoriously difficult because of the curse of dimensionality. We then introduce a Lagrangian relaxation-based policy (referred to as threshold Lagrangian relaxation (tLR)) as a heuristic approach for fulfillment decisions and prove its performance guarantee. To validate our findings, we utilize both synthetically generated data sets and data provided by our partner e-commerce retailer, and we conduct two numerical studies comparing the tLR policy with benchmark policies. These studies validate our findings on the performance guarantees of the proposed heuristic approach, highlighting the benefits of multiperiod fulfillment windows and the cost-reducing capabilities of the tLR policy. We conclude by emphasizing the importance of dynamic fulfillment strategies and the considerations that e-commerce companies should take into account when selecting their fulfillment policies.