伯努利节点网络模型的拓扑鲁棒性分析方法

摘要

基于网络连通和恢复能力提出连接鲁棒性和恢复鲁棒性两种测度指标,根据随机故障和恶意攻击两种网络失败类型将连接鲁棒性分为随机故障鲁棒性和恶意攻击鲁棒性,将恢复鲁棒性分为随机故障节点恢复鲁棒性、随机故障边恢复鲁棒性、恶意攻击节点恢复鲁棒性、恶意攻击边恢复鲁棒性,并给出了这六个测度指标的确切定义.利用这六个测度指标分析了伯努利节点网络模型的拓扑鲁棒性,得出不同情形下拓扑结构与这六个测度指标的关系,结果表明:无线网络平面拓扑结构的恶意攻击鲁棒性要优于层次拓扑结构,而其随机故障鲁棒性要劣于层次拓扑结构；平面拓扑结构的边恢复鲁棒性要优于层次拓扑结构,而其节点恢复鲁棒性要劣于层次拓扑结构.%Robustness is divided into connectivity robustness and recovery robustness according to connectivity and recovery. Then we divide connectivity robustness into random-fault robustness and hostile-attacks robustness, and divide recovery robustness into node recovery robustness and edge recovery robustness according to type of network failure,besides we define the concepts of connectivity robustness and recovery robustness. Then these concepts are used to evaluate the degree of robustness of Bernoulli node model, and the conclusion is achieved that the hostile-attacks robustness of hierarchy structure is less robust than planar structure while the random-fault robustness is more robust than planar structure, and that the node recovery robustness of hierarchy structure is more robust than planar structure while edge recovery robustness is less robust than planar structure.