Agile Bayesian belief networks and application on complex system reliability growth analysis

Abstract

Bayesian belief networks (BBN) provide an effective way of reasoning under uncertainty and diverse source information. BBN have a wide application of uncertainty modeling. With the application being more complex and dynamic, the modeling of BBN needs to be flexible and agile. In this paper, we have developed an improved BBN, called agile BBN, which emphasizes the structure and parameter learning of the model. An example is presented of using the agile BBN for a complex system reliability growth analysis.