Benefits of Bayesian Network Models

The application of Bayesian Networks (BR) in dependability is a relatively recent development. Their popularity grew in the area of reliability analysis of systems, since the 1990s. A large number of scientific publications in this area show the interest in the applications of BN in the field of dependability and risk analysis. The most important publications demonstrate equivalence with probabilistic methods conventionally used in dependability. We have now a number of survey papers that gives a good view of the ability of BN application to dependability. Unfortunately, this modeling formalism is not fully accepted in the industry. The questions posed by today's engineers focus on the validity of BN models and the resulting estimates. Indeed, the modeling formalism by BN is not based on a specific semantic in dependability but offers a general formalism for modeling problems under uncertainty. This book explains the principles of knowledge structuration to ensure a valid BN model and illustrate the flexibility and efficiency of representation by PGMs.