Product Robust Design and Process Robust Design: Are They the Same? (No.)

Taguchi's ideas of robust parameter design motivated the development of the dual-response approach, where both the mean and the variance of the quality response are modeled in terms of the design parameters and noise factors. These are then used to identify optimal settings that achieve the dual objective of optimizing the signal (the mean) and minimizing variation. While much research has been published recently with regard to how to solve the dual-response problem (DRP), relatively little attention has been given to the unique characteristics of process robust design, like the existence of systematic variation or intercorrelations among the process "controllable" variables. These properties indeed put process robust design in a category of its own (separate from product robust design). In this paper, we first expound these unique properties and develop a general formulation of the DRP as it applies to process robust design. We then report on an implementation to an industrial process in a high-tech corporation.