Optimization Under Uncertainty: Rethinking Margins at the Conceptual Design

摘要

In this paper we introduce two novel processes that aim to provide quantitative and perhaps visual understanding of how both uncertainty level and desired probability of constraint satisfaction affect the feasible design space and consequently the objectives space. The purpose of the "Price of Reliability" process is to quantitatively assess the inherent impact of targeting higher levels of reliability enabling the decision-maker to trade-off risk, cost and performance. The "Cost of Uncertainty" process on the other hand aims to provide quantitative information of how the uncertain parameters variability affects the feasible design space enabling the decision-maker to decide whether it is worth investing in diminishing uncertainties and if so, which uncertainties are more important for a particular project.