The Application And Reasoning Improvement Of Causality Diagram

The essential function of Artificial Intelligence is the capability solving uncertain information in practical researches. Thus, uncertain reasoning models are the core subject. Dynamic Causality Diagram is an uncertain model based on probability. It takes great advantage in uncertain reasoning by absorbing some advantages of both the Belief Network and the Fault Tree, and it solves some faults. With the rapid development of research, it grows quickly in the Fault Diagnosing field. Therefore, the more researches on Causality Diagram have academic and practical value.This dissertation discusses and studies to surround reasoning and application, and the main contents include:First, the reasoning improvement of Causality Diagram. In this paper, it uses the basic event matrix to get the Minimal Cut Sets by extracting matrix elements. And it proposes a method transforming Causality Diagram into the Binary Decision Diagram. Avoiding the process of disjoint cut sets. It can effectively reduce reasoning complexity.Second, extending the practical field. In this paper, it transforms Causality Diagram into the Bayesian Network and use the Depth First Left Most to find modules. It can improve practical fields of Causality Diagram. Using Causality Diagram to risky model of bank, the loss of importance of basic events is introduced to develop Causality Diagram. It calculates the influence of basic events of bank and analyzes these acting forms, then quantitative analysis the risk of bank.