Non-regular designs for the detection of interactions between control and noise factors
by
Yingfu Li
University of Houston - Clear Lake
Coauthors: Manohar Aggarwal, University of Memphis

Robust parameter design, originally proposed by Taguchi, is an important statistical tool for reducing system variation and improving product’s quality. In order to achieve this goal, one first needs to select a design which allows to identify important interactions between control factors and noise factors so that one can choose the settings of control factors of the identified important interactions in such a way that the response variable is robust to changes in the noise factors. Russell, Lewis Dean (2004) proposed an algorithm to generate regular fractional factorial designs for such purpose. In this talk, we introduce using non-regular fractional factorial designs constructed from Hadamard matrices for the detection of interactions between control and noise factors.

PDF

Date received: July 11, 2007