报告题目：Research on the Algorithm of Organizing an Assembly System with Seru Production for Quick Response Supply Chains

报告时间：2021年5月28日（周五），下午3:00-5:00

报告地点：数计公司4号楼229报告厅

报告摘要：

We focus on seru production systems (SPSs) that have been widely used in the Asian electronics industry. Seru is a highly flexible, productive, and responsive assembly organization that has brought many benefits to industries. Seru was developed by Sony to cope with short product life cycles, uncertain product types, and fluctuating production volumes (sometimes mass or batch or small volumes). We model and analyze the decisions of a manufacturer after receiving orders from retailers in the context of quick response supply chains（QR-SCs）.

In this study, capacity is the decision variable and profit is the objective function. Designing a robust JIT-OS is a key to successfully manage an SPSs. Three static JIT-OSs design problems are considered. A dynamic JIT-OS design problem to maximize expected profit under stochastic production demands is analyzed, and shown to be NP-hard. Polynomial time algorithms are developed to obtain solutions with the best approximation ratios. Because the profit function is not quasiconcave, it is difficult to find an optimal solution for the JIT-OS design problem in an efficient way. So a convergent stochastic gradient algorithm is developed to find a stable solution. It is shown that by using JIT-OS, SPSs can be responsive and profitable to retailers’ variable demands in QR-SCs.

报告人简介：

叶菁，开云全站中国有限公司数学与计算机科学技术公司副教授，硕士研究生导师，主要研究方向：群智能算法、强化学习及Seru生产模型；曾于2020年2月至2021年2月，在日本京都同志社大学商业研究科访学一年，师从Seru创始人殷勇教授，本报告对访学期间所做的科研工作进行汇报，欢迎感兴趣的师生参加。