A variance-based estimation of the resilience indices in the preliminary design optimisation of engineering systems under epistemic uncertainty

This paper presents novel heuristics for the fast conservative approximation of resilience indices in the preliminary design optimisation of engineering systems under uncertainty. Since the uncertain in the early phases of the design process is mainly of an epistemic nature, Dempster–Shafer theory of evidence is proposed as the reasoning framework. The heuristics proposed in this paper are used to partition the uncertainty space in a collection of subsets that is smaller than the full set of focal elements but still provides a good approximation of Belief and Plausibility. Under suitable assumptions, this methodology renders the approximation of the Belief and Plausibility curves cost-effective for large-scale evidence-based models. Its application to the preliminary-design sizing of a small spacecraft solar array under epistemic uncertainty will be demonstrated.