梯形模糊因果图及归一化研究  

Trapezoidal Fuzzy Causality Diagram and its Normalization

作　　者：梁新元[1] 吴淑皇[2] 石庆喜[1] 

机构地区：[1]重庆工商大学计算机科学与信息工程学院,重庆400067 [2]重庆大学自动化学院,重庆400030

出　　处：《计算机科学》2006年第12期183-186,共4页Computer Science

基　　金：高等学校博士点专项基金(99061116);重庆市自然科学基金项目(8084)联合资助。

摘　　要：因果图理论是一种基于概率论的推理方法。文章在分析因果图理论发展和存在的问题基础上,将模糊数学引入因果图理论,即模糊因果图,从而可以克服因果图分析中概率难以精确赋值的缺点,将因果图理论应用扩大到了模糊领域。文章主要对事件概率为梯形模糊数进行讨论,提出了模糊因果图的算子,得到了模糊条件概率的计算公式,讨论了模糊概率的归一化方法。最后,以核电站的一个子系统为例进行仿真实验,实验结果与实际一致,归一化方法可行。研究表明模糊因果图能有效地用于故障分析,比原来的因果图方法具有更大的灵活性和适应性。Causality Diagram is a probabilistic reasoning method. In this paper, fuzzy set theory is introduced to develop Causality Diagram Methodology after discussing the development and the restriction of conventional Causality Diagram, Fuzzy Causality Diagram can overcome the shortcomings that it is difficult to gain the accurate probability of the event in conventional Causality Diagram. The application of Causality Diagram is extended to fuzzy field by introducing fuzzy set theory. Based on the trapezoidal fuzzy number, operator, fuzzy conditional probability and the normalization method are discussed in this paper. The result of numerical simulating of a subsystem in nuclear plant is coincident with the fact, and it shows the normalizing method is effective. The research shows that Fuzzy Causality Diagram is so effective in fault analysis, and it is more flexible and adaptive than conventional method.

关 键 词：因果图 模糊因果图 事件概率 模糊数 梯形模糊数 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]