A multi-objective systems reliability approach for infrastructure design under aleatory and epistemic uncertainty

Abstract

Robust structural design requires joint consideration of both aleatory (random) and epistemic (lack of knowledge) uncertainty. In a multi-objective decision-making context, aleatory uncertainties are associated with uncertainty inherent to the built environment, whereas epistemic uncertainties accrue due to imperfect representation of decision maker values. This paper presents a systematic decision framework for combining treatment of these types of uncertainties using the First Order Reliability Method (FORM). Probabilistic descriptions of the performance of various alternative structural design configurations are combined with decision maker defined utility functions (indifference curves) that are subjected to epistemic uncertainty, supporting assessment of the sensitivity of the decision analysis results to various sources of error. The proposed decision analysis model allows full and consistent representation of all relevant uncertainties, multiple decision maker utilities, alternative design configurations, and alternative decision criteria. A demonstration identifies the optimal of 12 alternative configurations of a nine-story office building with the decision variables of building damage and business interruption cost, casualty cost, and CO2 emission cost under aleatory and epistemic uncertainties.