A0323
Title: Three-orthogonal designs robust to nonnegligible two-factor interactions
Authors: Jinyu Yang - Nankai University (China) [presenting]
Abstract: Considering the robustness of nonnegligible two-factor interactions, two-level orthogonal arrays with clear two-factor interactions are popular choices. However, if some significant quadratic effects exist, two-level designs do not work to detect them. There is a need to consider designs with three levels. Existing three-level designs with clear effects have a large number of runs, which are not economical choices. The concept of clear two-factor interactions is generalized to three-orthogonal designs, especially three-level ones. The constructions of this new class of robust designs are explored, and their desirable properties are illustrated. Specifically, some of the obtained designs can accommodate a large number of factors $(m = n/4 + 1)$. Due to the appealing structure, the obtained designs make themselves attractive choices for robust designs.
