Robust Designs for Quality Improvement

Taguchi's parameter design method allows one to make products or processes robust to uncontrollable noise factors. Recent research has focused on combining control and noise factors into a single array to model the response directly rather than relying on a performance measure such as the signal-to-noise ratio. Examination of control-by-noise interaction effects reveals control factor settings that dampen the effects of the noise factors. The objective of this paper is to propose modified cross-product designs for robustness improvement that enable one to reduce the experimentation costs while maximizing the amount of useful information.