Prior Robustness for Bayesian Implementation of the Fault Tree Analysis

摘要: We propose a prior robustness approach for the Bayesian implementation of fault tree analysis (FTA). FTA is often used to evaluate risk in large, safety critical systems but has limitations due its static structure. approaches have been proposed as superior alternative it, however, this involves elicitation, which not straightforward. show that minor misspecification priors elementary events can result significant top event. A large amount data required correctly update misspecified and such may be available many complex, systems. In cases, equals posterior misspecification. Therefore, there need develop FTA, quantify effects on analysis. Here, we first specifically developed FTA. only prove few important mathematical properties approach, also easy use Monte Carlo sampling algorithms implement any given with and/or or gates. then two real-life examples: spacecraft re-entry example feeding control system example. provide step-by-step illustration how applied problem.