一种带节点失效的BDD网络可靠性分析新方法  

A new BDD-based reliability evaluation method for networks with imperfect vertices

作　　者：潘竹生[1] 莫毓昌[1] 

机构地区：[1]浙江师范大学数理与信息工程学院,浙江金华321004

出　　处：《浙江师范大学学报（自然科学版）》2017年第4期406-414,共9页Journal of Zhejiang Normal University:Natural Sciences

基　　金：国家自然科学基金资助项目(61572442);浙江省重中之重学科"计算机软件与理论"开放基金资助项目(ZSDZZZZXK24)

摘　　要：针对边扩展图EED方法和边界集BS方法存在的存储空间浪费和可靠度结果精度不高等缺陷,提出了一种基于依赖集的分析方法,解决了节点和边随机失效的网络可靠性问题.新方法不再受排序起点约束,可以灵活选择高质量排序以获得紧凑的BDD模型;同时,基于依赖集的分区能正确表征网络特征以获得精确解.综合实例和大量实验表明,所提出的方法是正确和高效的.Edge expansion diagram（ EED） and boundary set（ BS） methods had various constraints leading to problems in accuracy of reliability and space efficiency to build BDD model for the network with imperfect nodes. To overcome these problems,a new BDD-based algorithm called OBDD-DS was proposed for K-terminal network reliability analysis considering both edge and node failures. The proposed method had no restrictions on the starting node for the compact BDD model construction. In addition,based on a newly-defined concepts ＂ dependency set＂,the proposed methods could accurately compute the reliability of networks with imperfect nodes relying on identifying sub network correctly. Comprehensive examples and experiments in wide variations of networks were provided to show correctness and effectiveness of the proposed approach.

关 键 词：网络可靠度 二叉决策图 依赖集 边扩展 

分 类 号：TB114[理学—概率论与数理统计]