Optimal policy for managing stochastic cash flows in a financial supply chain

• 成果类型: Article; Early Access
• 署名作者: Kang, Keumseok; Gupta, Sushil; Hur, Inkyoung; Jun, Sungbum
• 署名单位: Korea Advanced Institute of Science & Technology (KAIST); State University System of Florida; Florida International University; State University System of Florida; Florida Atlantic University; Dongguk University
• 刊物名称: PRODUCTION AND OPERATIONS MANAGEMENT
• ISSN/ISSBN: 1059-1478
• DOI: 10.1177/10591478261460124
• 发表日期: 2026
• 关键词: trade credit; queuing-systems; IMPACT; jobs; assignment

摘要

Billions of dollars are exchanged between companies through accounts payable in today's business landscape. Given its immense scale and critical role in business operations, effectively managing accounts payable and working capital is essential for organizations. In this study, we examine the financial supply chain problem, where a company seeks to minimize total payments toward accounts payable received from its upstream partners (e.g., suppliers) while leveraging cash inflows from downstream partners (e.g., distributors, wholesalers, retailers, and customers). This is accomplished by optimizing payment decisions based on payment terms and capitalizing on interest gains from cash on hand over time. Unlike prior studies, we investigate this problem in a more realistic setting where information about incoming invoices and cash inflows is uncertain. We formulate the problem as a stochastic dynamic program and identify the structural properties of an optimal policy. The optimal policy reveals payment priorities among invoices and establishes thresholds for maintaining cash on hand. We further find that payment priorities can be deterministic or stochastic, depending on the problem's state and random parameters. Additionally, we identify all instances where payment priorities are deterministic. Our study provides valuable managerial insights and practical implications derived from the structural properties of the optimal policy. Notably, some of these insights challenge well-known heuristics and seemingly intuitive practices. Lastly, we develop a simple heuristic based on the identified structural properties and demonstrate that it outperforms other widely used methods for solving large-scale practical problems.