The Value of Representing Epistemic Uncertainty in Engineering Design

Engineering design decisions inherently are made under uncertainty. In this paper, we recognize a difference between aleatory uncertainty (due to inherent randomness) and epistemic uncertainty (due to lack of knowledge). Our hypothesis is that, in engineering design decisions, it is valuable to explicitly represent epistemic uncertainty distinctly from aleatory uncertainty. In this paper, we support this hypothesis with a computational experiment in which a pressure vessel is designed using two approaches, both variations of utility-based decision making. In the first approach, designers use a purely probabilistic, best-fit normal distribution to represent uncertainty. In the second approach, designers explicitly express epistemic uncertainty distinctly from aleatory uncertainty using a probability box, or p-box. When the epistemic uncertainty is large, this latter approach results on average in designs with expected utilities that are greater than those for designs created with the purely probabilistic approach. In the context of decision theory, this suggests that there are design problems for which it is valuable to explicitly represent epistemic uncertainty distinctly from aleatory uncertainty. INTRODUCTION Of the many challenges in engineering design, one of the greatest is uncertainty. During the design process, engineers must make decisions without being certain of the outcomes of these decisions. Prior to making a decision, engineers can remove some of this uncertainty by expending resources to acquire more knowledge, for example, by modeling the product