Invulnerability and node importance evaluation method for complex networks复杂网络抗毁度和节点重要性评价方法.pdfVIP

摘要 摘 要 几乎所有的复杂系统都可以抽象为复杂网络模型，这些网络大多数具有大量 的节点，并且节点之间有着复杂的连接关系．深入研究这些网络就可以揭示隐藏 在自然界中的复杂系统的一般规律．复杂网络作为大量真实复杂系统的高度抽象， 由于受计算机仿真和大规模的实际网络数据库的支持近年来在学术界倍受关 注．规模越庞大，结构越复杂，网络的故障就越多，从而使网络面临的威胁和攻 击也越来越多，因此复杂网络的抗毁性逐渐成为学术界的一个新兴研究热点． 由于对复杂网络的拓扑结构知之甚少，其抗毁性测度的研究一直是一个比较 困难的问题．本文在对复杂网络抗毁性问题相关技术分析的基础上，从网络拓扑 结构的一般属性出发，阐述复杂网络拓扑特性及形成机制，研究了复杂网络各种 特性对抗毁性的影响；探讨从网络连通性的角度出发，在随机失效率的前提下对 网络的抗毁性测度进行了定义．并分别在只有选择性攻击和随机性攻击的情况下 验证了所提的全网连通度的有效性．进一步，对无标度网络和随机网络的抗毁性 能进行了比较和分析．结果表明考虑了随机失效率的网络抗毁度定义更加确切有 效．其次，在本文所提的全网连通度的基础上，提出了一种对连通网络和非连通 网络都适用的节点重要性评价方法，该方法简单、直观，且能有效地评估网络中 节点的重要度． 关键词：复杂网络 抗毁度 节点重要性 随机失效率 连通测度 Abstract Abstract Almost all of the complex systems can be abstracted as complex network models, and the networks mostly have a large number of nodes with complicated connection relationships. By delving into the networks deeply, the hidden general rule of the complex systems in nature can be revealed. Complex networks, as the high abstraction of the real complex systems in most cases, have drawn great attentions of researchers in recent years due to the supports of computer simulation and large-scale real network database. However, with the increase of the scale, the structure becomes more complicated and results in more network errors which will definitely produce more threats and attacks to the network. Consequently, complex network invulnerability gradually becomes a new focus of the research interests. For complex networks, the study of their invulnerability measure is still a hard problem because of their unavailable topology structures. Based on the analyses of the related methods aim