Hierarchical propagation of probabilistic and non-probabilistic uncertainty in the parameters of a risk model

We consider a model for the risk-based design of a flood protection dike, and use probability distributions to represent aleatory uncertainty and possibility distributions to describe the epistemic uncertainty associated to the poorly known parameters of such probability distributions.A hybrid method is introduced to hierarchically propagate the two types of uncertainty, and the results are compared with those of a Monte Carlo-based Dempster–Shafer approach employing independent random sets and a purely probabilistic, two-level Monte Carlo approach: the risk estimates produced are similar to those of the Dempster–Shafer method and more conservative than those of the two-level Monte Carlo approach.

► We perform the hierarchical propagation of hybrid uncertainties in risk assessment. ► We study the effects of dependence between epistemically uncertain parameters. ► We study the effects of different representations of epistemic uncertainty. ► Different methods generate different results and decisions in risk problems. ► Non-probabilistic representations of uncertainty produce conservative results.