Constructing minimum aberration split-plot designs via complementary designs when the subplot factors are more important

Abstract

Robust parameter design (RPD) is a statistical engineering method for reducing the sensitivity of a system (i.e., a product or process) to noise variations by selecting a treatment combination of control factors, with the aim of reducing performance fluctuations of the system. Then the control factors are more important in RPD. The fractional factorial split-plot (FFSP) designs can be used when the noise and control factors are considered as the whole plot (WP) factors and subplot (SP) factors respectively in the design. Based on the above situation, this paper considers the FFSP design when the SP factors are more important. A criterion called minimum aberration of type SP (SP-MA) is proposed and the SP-MA FFSP designs are constructed by using complementary designs. The superiority of the SP-MA criterion in seeking FFSP designs when the SP factors are more important than the WP factors is illustrated by comparison.