Value of information under random decision, model, and measurement errors

Value of Information (VoI) analysis provides a quantitative measure to justify the adoption of a structural health monitoring (SHM) strategy for civil infrastructure maintenance. The VoI analysis adopts predictive and measurement-based models to characterise the infrastructure state and monitoring quality. The uncertainties in such models have been explicitly considered in the current VoI framework and the decision maker is expected to choose the action maximizing the expected utility defined by a deterministic utility function. However, under the influence of some tangible or intangible factors like budget availability, bias, etc., the decision maker may end up choosing a sub-optimal action. The present study characterizes such decisions by introducing a concept of cost of decision and models such decisions by a random error, in addition to the predictive model and measurement errors, to assess their influence on the VoI. The results indicate the negative influence of such decisions on VoI.