Process Control in Batch Manufacturing: Dynamics and Coordination

本文研究了批量生产系统中的质量控制问题。其目的是开发技术，以确定适用于质量控制问题的政策和控制措施。这些技术被描述为动态的，在这个意义上说，他们不限于参数设计，但也考虑到政策和控制的发现。重点是顺序技术。顺序技术利用测试和过程监控产生的动态信息。 过程控制将与批量生产中的其他相关物流问题一起研究。具体而言，研究将沿着以下方向进行：（1）能力约束下的阶段间协调，（2）生产，检查和工艺修订之间的协调，（3）子装配操作和可替代部件之间的协调，以及（4）计算，近似和聚合技术。研究的目的还在于识别某些最优顺序过程并揭示其结构特性，特别是具有阈值结构的最优顺序策略。研究重点的基本方法是马尔可夫决策规划结合使用随机比较技术，包括那些基于随机凸性和随机子模块化的概念。 这项研究如果成功，将产生几个好处。它将促进批量生产质量控制的知识状态。特别是，顺序控制政策的有效性，质量动态的重要性，以及过程控制与其他生产物流的协调代表了关键的贡献。 此外，将有新的工业技术的研究成果的基础上。最后，还有一些教育方面的收获。本研究将为动态过程控制的工程课程提供新的教材。通过这种方式，下一代工程师将从研究中受益，将从课程中获得的知识带到他们的各种雇佣组织。

This research is focused on the problem of quality control in a batch manufacturing system. Its aim is to develop techniques for identifying policies and control measures to apply to quality control problems. The techniques are described as dynamic in the sense that they are not limited to parametric designs but also consider the finding of policies and controls. The focus is on sequential techniques. Sequential techniques make use of the dynamic information that result from testing and process monitoring. Process control will be studied with other related logistics issues in batch manufacturing. In particular, the research will proceed along the following directions: (1) inter-stage coordination under capacity constraint, (2) coordination among production, inspection, and process revision, (3) coordination among subassembly operations and substitutable components, and (4) computation, approximation, and aggregation techniques. The research also aims at identifying certain optimal sequential procedures and revealing their structural properties, particularly, optimal sequential policie s with threshold structures. The basic methodology underlying the research emphasis is Markovian decision programming incorporated with the use of stochastic comparison techniques, including those based on the notion of stochastic convexity and stochastic submodularity. This research will yield several benefits if successful. It will advance the state of knowledge in quality control in batch manufacturing. In particular, the effectiveness of the sequential contrml policies, the importance of quality dynamics, and the coordination of process control with other production logistics represent key contributions. In addition, there will be new industrial technology findings based on the research. Finally, there are some educational gains as well. The research will present new teaching materials for an engineering course in dynamic process control. This way, next generation engineers will benefit from the research by taking the knowledge gained from the course to their various employing organizations.