Optimal redundancy allocation in structural systems: a link between Bayesian utility theory and Info-Gaps
Lucas Rodrigues da Silva, Andre Teofilo Beck
SSRN Electronic Journal · 2026
What is the optimal redundancy of a structural system? Redundancy has a cost, due to the strengthening required to produce alternative load paths, but it may reduce the expected consequences of failure. Bayesian utility theory alone cannot answer the question, because: a) one can always find a cheaper non-redundant system which meets a prescribed target system reliability, and b) structural reliability theory does not address all sources of uncertainty.
In fact, structural reliability theory neglects many of the uncertainties which have historically contributed to structural collapses. The latent failure probability concept is an effort to close this gap. It recognizes the open boundaries of structural systems and the effects of uncertainties not dealt with in reliability theory.
Herein, it is shown that the latent failure probability concept provides a link between Bayesian utility theory and Information-Gaps, with implications to robust design under deep epistemic uncertainty. This leads to two complementary optimization problems: 1. for fixed redundancy, Bayesian utility theory provides the element cross-sections which maximize expected benefits or minimize total expected costs; 2. Info-gap theory provides the redundancy which maximizes robustness with respect to ignorance.
This leads to Pareto fronts of compromise solutions, where the trade-offs between maximizing utility and maximizing robustness can be decided by the analyst.