Selection of non-regular fractional factorial designs when some two-factor interactions are important
Metadata[+] Show full item record
Introduction: Non-regular two-level fractional factorial designs, such as Plackett–Burman designs, are becoming popular choices in many areas of scientific investigation due to their run size economy and flexibility. The run size of non-regular two-level factorial designs is a multiple of 4. They fill the gaps left by the regular two-level fractional factorial designs whose run size is always a power of 2 (4, 8, 16, 32, …). In non-regular factorial designs each main effect is partially confounded with all the two-factor interactions not involving itself. Because of this complex aliasing structure, non-regular factorial designs had not received sufficient attention until recently. ... In practical applications of non-regular designs, it is often in the case that some of the two-factor interactions are important and need to be estimated in addition to the main effects. In this article, we consider how to select non-regular two-level fractional factorial designs when some of the two-factor interactions are presumably important. We propose and study a method to select the optimal non-regular two-level fractional factorial designs in the situation that some of the two-factor interactions are potentially important. We then discuss how to search for the best designs according to this method and present some results for the Plackett-Burman design of 12 runs.
Journal of Modern Applied Statistical Methods, May, 2008, Vol. 7, No. 1, 94-100.