Quantitative system risk assessment from incomplete data with belief networks and pairwise comparison elicitation

TL;DR

This paper introduces a Bayesian belief network approach for quantitative risk assessment from incomplete data, using pairwise comparison for prior elicitation and Bayesian updating, demonstrated on spacecraft re-entry risk.

Contribution

It presents a novel method combining belief networks and pairwise comparison elicitation for risk assessment with incomplete data, applied to spacecraft explosion risk.

Findings

Effective Bayesian updating with incomplete data

Successful application to spacecraft re-entry risk

Improved elicitation of prior probabilities

Abstract

A method for conducting Bayesian elicitation and learning in risk assessment is presented. It assumes that the risk process can be described as a fault tree. This is viewed as a belief network, for which prior distributions on primary event probabilities are elicited by means of a pairwise comparison approach. A Bayesian updating procedure, following observation of some or all of the events in the fault tree, is described. The application is illustrated through the motivating example of risk assessment of spacecraft explosion during controlled re-entry.