Abstract

In many industries, quality characteristics are measured as ordered categorical responses due to measurement difficulties, or for cost-effectiveness. Although various approaches have been proposed to find robust levels of product and process design parameters when the response variable is continuous, the literature presents a limited number of studies with their own drawbacks for the ordered categorical response case. Motivated by this, we propose an alternative approach that models location and dispersion effects separately as functions of design parameters and finds settings of the parameters by considering both effects simultaneously and incorporating the decision maker’s preferences on the tradeoffs between them.

摘要

在许多行业中，由于测量困难或成本效益，质量特性被测量为有序的分类响应。尽管已经提出了各种方法来在响应变量连续时找到产品和过程设计参数的稳健水平，但文献提出的研究数量有限，对于有序分类响应案例有其自身的缺点。受此启发，我们提出了一种替代方法，将位置和分散效应分别建模为设计参数的函数，并通过同时考虑这两种效应并结合决策者对它们之间权衡的偏好来找到参数的设置。