Expected Utility Theory For Monitoring-Based Decision Support System

Abstract

In theory, we all recognize that Structural Health Monitoring (SHM) should help infrastructure owners to make decisions on the management of their assets. In practice, we often observe owners who are very sceptical of the benefit of SHM and act based on their experience, disregarding the actions suggested by SHM. In this contribution, we wish to fill this gap by outlining a formal approach to rational decisions based on monitoring information. SHM and decision are two separate processes, occurring one downstream of the other: whereas SHM allows assessing the state of a structure based on monitoring data, decision provides the optimal action, based on the structural state assessed and the possible consequences of any individual choice. The first process is rationally tackled by using Bayesian inference, the second by using the axioms of expected utility theory (EUT). We illustrate this approach on a real-life case study, the Colle Isarco viaduct, among the longest spans in the European Alpine region. The decision support system (DSS) developed interprets the data coming from a robotized topographic monitoring system, infers information on the state of the structure, and suggests the optimal management policy after a possible damaging event.