Statistical Experimentation Methods for Achieving Affordable Concurrent Systems Design

Abstract

We describe an affordable method for designing in the conceptual stage using relatively high-fidelity concurrent systems analysis. Our method is rooted in three domains, namely, the design of experiments, the response surface methodology, and the compromise decision support problem. A sequential experimentation strategy as well as the heuristic rules for creating high-order response surface models are introduced as a cost-effective approach to applying the statistical experimentation methods in design of complex systems. A high-speed civil transport aircraft design is used as an example to illustrate the potential of our approach.