Generalized Minimum Aberration and Design Efficiency For Nonregular Fractional Factorial Designs

Abstract

Generalized Minimum Aberration (GMA) criterion was recently proposed by Deng and Tang for both regular and nonregular designs. It is a natural extension to the Minimum Aberration (MA) criterion for regular factorial designs. In this talk, we establish a relationship between the GMA criterion and some design efficiency criteria. It is shown that the GMA criterion is supported by these design efficiency criteria. Several advantages of using the GMA criterion are also discussed. An extensive evaluation of designs will be presented to show that the GMA criterion can classify and can also give a complete ranking on these designs. The empirical study also demonstrates that the GMA ranking is consistent with the rankings by these model dependent criteria.