EAGER: Turning Uncertainty into Advantage: Novel Policies that Exploit Randomness of Lead Times in Supply Chains

由于环境、政治和商业风险促使公司分散供应来源，技术进步使分布式供应商网络具有竞争力，因此交货时间的随机性仍然是库存管理中一个重要而具有挑战性的问题。与普遍认为可变性会降低库存管理策略绩效的观点相反，非常早期的证据表明，在某些情况下，交货时间的可变性可以用于制定降低库存总成本的策略。这个早期概念探索性研究（EAGER）项目将提供进一步的实验证据，以了解是否以及如何利用交货时间的不确定性来降低库存成本。该项目将调查交货时间和需求模式的分布信息和可变性如何对库存管理的适应性方法产生新的见解，这些方法可以动态调整交货时间的变化。该项目将深入研究可变性可以有效地以自适应方式开发数据驱动算法的前提，以确定降低总体库存成本的策略。这与绝大多数库存文献形成鲜明对比，这些文献侧重于明确的分析方法，需要对交货时间和需求的（平稳）分布进行假设。这些假设在实际的库存管理场景中通常不受支持。本项目审查现实的和一般的模型设置，目的是确定影响新政策绩效的关键因素并对这种绩效进行量化。虽然这个项目仍然是非常初步和实验性的，但预期结果将导致可检验的假设，这些假设可能对库存理论和应用产生重大影响，从而产生广泛的实际影响。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

As environmental, political and business risks motivate companies to diversify their supply sources, and technology progress makes distributed networks of suppliers competitive, randomness in lead times remains an important and challenging issue in inventory management. Contrary to prevailing wisdom that variability degrades the performance of inventory management strategies, very early evidence suggests that lead-time variability can, in certain cases, be used to advantage to produce strategies that reduce overall cost of inventory. This EArly-concept Grant for Exploratory Research (EAGER) project will provide further experimental evidence for understanding whether and how lead time uncertainty can be exploited to reduce inventory costs. The project will investigate how distributional information and variability in lead-time and demand patterns may yield new insights into adaptive approaches to inventory management that can dynamically adjust to changes in lead times. This project will thoroughly investigate the premise that variability can be effectively exploited in an adaptive fashion to develop data-driven algorithms that identify strategies to reduce overall cost of inventory. This stands in contrast with the vast majority of inventory literature, which is focused on explicit analytical methods that require assumptions on the (stationary) distribution of lead time and demand. These assumptions are often not supported in realistic inventory management scenarios. This project examines realistic and general model settings, with the goal of identifying the key factors affecting the performance of new policies and quantifying this performance. While still very preliminary and experimental, the results of this project are expected to lead to testable hypotheses that could potentially have great implications for inventory theory and application and therefore broad practical effect.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.