Comparative analysis of Taguchi's crossed array approach vs combined array approach to robust parameter design: A study based on apparel industry

Abstract

This research compares Taguchi's crossed array approach with the combined array approach (an alternative proposed by statisticians) using an experiment designed to optimize the sewing conditions to minimize the variability of a high end women's garment. The optimum settings given by the two approaches were similar, although the combined array approach provided more information about the process with a slightly better parameter setting. The findings imply that the benefit an apparel organization gets in conducting a statistically sophisticated analysis needs to be carefully weighed against the statistical simplicity Taguchi's crossed array approach offers in robust parameter designs.