SGER: Joint Optimization of Resource Allocation and Operational Decisions for Improving Random Yield

这项探索性研究小额资助（SGER）将重点关注集成电路（IC）制造和航空旅行提供，作为此类制造和服务业的核心例子。目标是开发数学模型，为确定如何在各种可能的方法中分配稀缺资源以提高产量提供基础。 其中要解决的关键问题是：（1）不同的产量改变干预措施如何影响产量分布的参数？ (2)收益率分布的变化如何影响经营决策？ (3)在收益率分布的哪些随机和/或参数顺序下，公司在经济上受益？ (4)如何量化这一效益的大小？ 这些问题将使用运筹学的技术来回答，包括马尔可夫决策过程，随机/可变性排序和随机规划。 随着生产过程的日益复杂以及服务行业创造和管理大量细分市场的能力，产量管理对制造商和服务提供商的重要性日益提高。有大量的文献涉及制造业背景下进行生产库存决策所需的数学模型，或者类似地，以决定在服务业的背景下，在特定票价等级中提供多少容量，或者超额预订多少。 这项研究将通过开发共同的模型来评估产量改变干预措施，从而在两个不同的文献机构之间架起一座桥梁。

This Small Grant for Exploratory Research (SGER) will focus on the manufacture of integrated circuits (ICs) and the provision of air travel as the central examples of such manufacturing and service industries. The goal is to develop mathematical models that provide a basis for determining how to allocate scarce resources among a variety of possible methods available to improve yield. Among the key questions to be addressed are: (1) How do different yield-altering interventions affect the parameters of the yield distribution? (2) How are operational decisions affected by changes in yield distributions? (3) Under which stochastic and/or parametric orders of yield distributions does a company benefit economically? (4) How can the magnitude of this benefit be quantified? These questions will be answered using techniques from operations research, including Markov decision processes, stochastic/variability orderings, and stochastic programming. The increasing complexity of production processes and the ability to create and manage a large number of market segments in service industries have served to elevate the importance of yield management to manufacturers and service providers alike.There is an extensive body of literature dealing with mathematical models necessary to make production-inventory decisions in the manufacturing context, or analogously, to decide how much capacity to make available in a particular fare-class, or how much to overbook, in the service industry context. This research will provide a bridge between two disparate bodies of literature by developing common models to evaluate yield-altering interventions.