Fuzzy Reasoning Based On Causality Diagram And Its Importance Analysis

At present,artificial intelligence(AI)technology is more and more used in many practical problems.Fault diagnosis of complex systems is one of the core problems of artificial intelligence.Its expression is varied,such as reliability network,directed graph,cognitive map and Causality diagram.Causality diagram can make all kinds of causality in the system present in a graphical way,and make all kinds of causality more intuitive for subsequent analysis and processing.In recent years,causal graph algorithm has been continuously optimized,which has important value in fault analysis,fault diagnosis,disaster prediction and risk prediction.This paper mainly deduces the fuzzy reasoning and importance analysis of causality diagram.Its main contents are as follows:Firstly,in the fault diagnosis of complex systems,the uncertainty of event occurrence probability is caused by the complexity of fault mechanism.In this paper,the hyperellipsoid model is proposed to deal with the range of uncertain intervals in causal graphs,and T-S fuzzy gate is used instead of AND gate,OR gate,so as to effectively diagnose the fault of complex systems and make them more adaptable and flexible.Secondly,based on the advantage that binary connection number can connect certainty and uncertainty in a certain range,it is applied to interval causality diagram fault diagnosis method.According to the error distribution form of interval number,the interval number and connection number can be converted to each other,and the uncertainty of interval causality diagram can be objectively expressed to alleviate the uncertainties such as complexity,diversity and lack of data.Then,in order to improve the accuracy and scientificalness of fault diagnosis results,this paper applies the entropy-weighting TOPSIS method to the analysis of the importance of basic events in causality diagram.Taking "importance" as an index,the entropy-weighting coefficient is introduced to quantitatively determine the weight of parameters,so as to avoid the subjective determination of the weight of parameters leading to the reasonableness of the results.Then the TOPSIS method is used tocalculate the relative proximity between positive and negative ideal solutions to obtain accurate diagnostic sequences.Finally,in order to expand the application field of causality diagram,it is applied to the ship navigation risk diagnosis,quantitative and qualitative analysis of the impact of basic events on the ship navigation risk,and gives the applicable suggestions.