ABSTRACT
This chapter develops techniques for Belief Risk Assessment, and the Loss of Coolant Accidents fault tree from the Interim Reliability Evaluation Program study provided an example in which to apply those techniques. It then looks at the extension of the belief function models to both decision analysis problems and risk analysis in a public policy setting. The chapter introduces second-order belief function models for some of the types of events typically found in system reliability models. It extends the models of Dempster produces belief function models for component failures and component failure rates for both data-available and data-free components. In belief function models, information about the type parameters comes in the form of random intervals rather than random point estimates. The 95% intervals produced by the belief function model are slightly smaller, but this could be due to sampling variability from the Monte Carlo experiment or the unpooling used by the Maximus method.
