Article ID: | iaor1994940 |
Country: | United Kingdom |
Volume: | 42 |
Issue: | 2 |
Start Page Number: | 271 |
End Page Number: | 281 |
Publication Date: | Apr 1993 |
Journal: | Applied Statistics |
Authors: | Peddada Shyamal Das, Kirmani S.N.U.A. |
Keywords: | Taguchi method |
The notion of stochastic ordering is used to obtain performance criteria for the parameter design problem when the target value for the quality characteristic is fixed. This approach provides a rigorous alternative to Taguchi’s signal-to-noise ratios. Sample analogues of the performance criteria are introduced. A procedure for obtaining optimum factor levels is developed and illustrated. While retaining the good features of the Taguchi methodology, the present procedure overcomes some of the objections to his applications of the analysis of variance.