基于饱和泊松分布的网络边的Birnbaum重要度计算方法  

Network Birnbaum Importance Measure under Saturated Poisson Distribution

作　　者：杜永军 侯沛勇[1] 郭雅琪[1] 

机构地区：[1]西北工业大学机电学院,陕西西安710072 [2]兰州理工大学理学院,甘肃兰州730050

出　　处：《西北工业大学学报》2017年第5期870-875,共6页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(7147147);高等学校学科创新引智计划(B13044)资助

摘　　要：重要度是系统可靠性领域用以识别系统薄弱环节的主要方法之一,也是系统可靠性优化的前提和基础。在对传统Birnbaum重要度计算方法研究的基础上,假设网络边故障个数概率分布已知的前提下,基于网络结构谱,给出了一种基于概率分布的Birnbaum重要度计算方法,重点探讨了网络故障边个数服从饱和泊松分布时Birnbaum重要度的性质。最后,结合算例给出了基于饱和泊松分布的Birnbaum重要度的应用方法。Network reliability analysis aim to quantify the impact of component failures on the network failure and to identify the weakness in a network. Importance measures provides numerical indicator to determine which compo- nents are more important for network reliability improvement or more critical for network failure. In this paper, the number of failed components is introduced, which is a random variable having a distribution. Under the condition of the distribution have been known, based on the structural spectrum, a new method to compute Birnbaum importance measure is derived. In particularly, when the number of failed components follows a saturated Poisson distribution, several results concerning ranking of components according to Birnbaum importance measure is presen- ted. Finally, an example is presented to explain the application of the results.

关 键 词：网络 Birnbaum重要度 饱和泊松分布 网络结构谱 

分 类 号：N945.17[自然科学总论—系统科学]