多元质量特性的诊断与控制是现代统计过程控制技术研究的热点和难点,如何分析多元过程的能力及如何计算其多元过程能力指数还没有公认的方法.本文详细列举了4种主要思路下构建多元过程能力指数的模型,简单介绍了几种特别的计算和度量多元过程能力指数的方法,如Bootstrap法、几何距离法等,最后对各种思路和模型进行了简单对比.
Diagnosing and controlling the multivariate characteristics are hot and difficult topics in the area of modern statistical process control. There is no any effective way to analyze the multivariate process capability, nor is to compute its indices. Four ideas of computing multivariate process capability indices(MPCIs) and the corresponding models are described; some special methods for their computation are also briefly introduced, such as the bootstrap method, and the geometric-distance approach. Finally, a simple comparison and conclusion of these models and methods are made.