On Comparison of Multiple Non-linear Profiles for Process Selection

Abstract

A product can be produced by using alternative processes. The processes that can produce the product having consistent quality performance as the existing process are qualified alternative processes. Having several qualified alternative processes can relax the conditions for production scheduling and increase production capacity. This research focuses on analyzing the quality characteristic which is nonlinear profile. The quality characteristic, which is the response variable, has a non-linear functional relationship with the explanatory variable. The objective of this research is to select a subset of qualified alternative processes among I alternative processes, I > 1. A Step-up Test Procedure is proposed to compare the non-linear profile of each process with the existing process. Polynomial regression is used to estimate the non-linear profile models before applying the step-up sequential tests. The proposed method can effectively test the significance of the differences among processes, while controlling the overall error rate of testing I processes below α.