An application of taguchi's methods reconsidered |
| |
Authors: | Veronica Czitrom |
| |
Institution: | The University of Texas at San Antonio , San Antonio, Texas, 78285 |
| |
Abstract: | Two aspects of Taguchi's methods for analyzing parameter design experiments that can be improved upon are considered. It is shown how using interaction graphs instead of marginal graphs, and how using the sample variance instead of a signal-to-noise ratio, can lead to product designs that are more robust to variation. The advantages of the alternative analysis will be illustrated by reanalyzing a case study considered by Barker (1986). |
| |
Keywords: | quality assurance parameter design interaction signal to noise ratio robust |
|
|