A Preference-Based Robust Design Metric
Abstract Robust optimal design can be studied as a problem in decision-making requiring tradeoffs between mean and variance attributes. In this context, this paper views Taguchi’s philosophy based design metrics using signal-to-noise (SN) ratios as empirical applications of decision-making under uncertainty with a priori sets of attribute tradeoff values. Alternatively, this paper presents a more rigorous preference-based design metric using concepts from utility theory to accurately capture designer’s intent and preferences. The use of this design metric as the robust optimal design criterion in a modified TRED (Tradeoffs in Robust Engineering Design) method with an innovative response-surface based iterative design space reduction strategy is presented. The effectiveness of the overall design procedure and the performance of the preference-based design metric are tested with the aid of demonstrative case studies and the results are discussed.