IGMS: Optimizing Supply Networks and Re-Entrant Manufacturing Systems in the Semiconductor Industry

PI将在INTEL公司的技术和制造集团居住期间对供应网络和可重入制造系统的调度进行研究。 在将非线性控制的理论成果和范例转化为大规模工业应用的同时，一个主要目标是深入了解工业运营的实际需求，约束和目标，足以与工业，商业和工业工程学院的代表进行有效的未来合作。并为未来的本科生和研究生做好最佳准备。技术重点是半导体行业直接感兴趣的问题。作为一个大约1000亿美元的业务，仅在美国，这一行业的特点是资本密集型制造工厂和产品生命周期短。再加上需求和价格的快速变化，这一框架有望为优化运营带来特别高的投资回报。第一个问题是优化半导体行业全球供应网络的管理。这包括逐步引进/淘汰不断变化的产品，使用分布在地球仪各地的设施，避免生产过剩，并对需求和生产的波动进行对冲。 一个关键的目标是连接常用的财务与最优控制。第二个问题是可重入制造系统的调度优化问题。一个主要的理论问题是制定一个类似的动态规划范例，这样的系统没有一个明确的方向的信号flow.This项目将立即影响许多课程教授和研讨会系列组织的PI。他还将迅速将他的见解传达给数学控制和本科生教育领域的更大社区。 该IGMS项目由MPS多学科活动办公室（OMA）和数学科学部（DMS）共同支持。

The PI will perform research on supply networks and scheduling of re-entrant manufacturing systems while in residence in the Technology and Manufacturing Group at INTEL Corporation. While transferring theoretical results and paradigms of nonlinear control into large scale industrial applications, a major objective is to develop in-depth knowledge of the practical needs, constraints, and objectives of industrial operations, sufficient for effective future collaborations with representatives from industry, business, and industrial engineering faculty, and to optimally prepare future undergraduate and graduate students.The technical focus is on problems of immediate interest to the semi-conductor industry. As a roughly $100B business in the US alone, this industry is characterized by capital intensive manufacturing plants andshort product life-cycles. Together with rapidly changing demands and prices, this framework promises a particularly high return on investments to optimize operations. The first problem is to optimize the management of the global supply networks of the semiconductor industry. This includes phasing in/out changing products, using facilities that are distributed around the globe, avoiding overproduction and hedging against fluctuations in demand and production. A key objective is to connect commonly used financial with optimal control. The second problem is to optimize the scheduling of re-entrant manufacturing systems. A major theoretical problem is to formulate an analogue of the dynamic programming paradigm for such systems without a distinguished direction of signal flow.This project will immediately impact the many courses taught and seminar series organized by the PI. He will also quickly communicate his insights to the larger communities in mathematical control and undergraduateeducation. This IGMS project is jointly supported by the MPS Office of Multidisciplinary Activities (OMA) and the Division of Mathematical Sciences (DMS).