主题：Leadtime Flexibility in an Assemble-to-Order Manufacturer

时间：2021年10月14日（周四）15:00-16:30

腾讯会议 ID：203 541 541

会议密码：6688

主讲人：邓天虎副教授（清华大学）

主持人：卿前恺副教授

主讲人简介：

邓天虎，清华大学工业工程系副教授。2008年毕业于清华大学工业工程系，获学士学位；2013年毕业于美国加州大学伯克利分校，获博士学位。主要研究智慧供应链的方法论框架和企业解决方案。2018年获国家自然科学基金优秀青年科学基金项目资助。负责执行的中石油天然气管网优化项目入围INFORMS设立的管理科学应用界最高奖项弗兰茨·厄德曼奖 （Franz Edelman Award）2018年决赛。目前研究成果已在Operations Research、Manufacturing & Service Operations Management, Informs Journal on Computing以及Interfaces等国际学术期刊发表。

摘要：

    Motivated by the widespread industrial practice of leadtime flexibility, that is, flexibly expediting outstanding orders, we study leadtime flexibility in an assemble-to-order (ATO) system. Although existing inventory planning engines do a good job for exogenous leadtimes, they come up short for flexible leadtime. We model deterministic/random leadtime flexibility with/without adjustment costs in the inventory planning stage as Markov Decision Process (MDP) models. State-reduction techniques are employed to simplify the models, and empirical data are tested to quantify the value of leadtime flexibility. First, we prove that the multidimensional MDP models of costless leadtime flexibility are equivalent to multiple unidimensional MDP models. Second, we numerically show that even a simple form of leadtime flexibility can achieve notable cost savings under a wide range of costs, demand variation, and flexibility randomness. Our analysis fills in the gap between the planned leadtime and flexibly executed leadtime, underscoring the importance of optimizing when and how to push suppliers to expedite.