南华大学计算机科学与技术学院毕业设计（论文）计算机科学与技术学院毕业设计(论文)论文题目基于复杂网络的因特网抗毁性分析 指导教师职 称博士讲师学生姓名学 号专 业网络工程班 级系 主 任院 长起止时间2013年10月11日至2014年5月23日2014年5月23日南华大学计算机科学与技术学院毕业设计（论文） 目 录 摘 要iAbstractii第一章 绪论11.1研究背景及意义11.2本文的主要工作31.3本文的组织结构3第二章 复杂网络的背景知识42.1从图论到复杂网络42.2复杂网络的结构参数52.3复杂网络的经典模型7第三章 复杂网络抗毁性113.1网络抗毁性的定义113.2网络抗毁性分析12 3.2.1抗毁性度量指标12 3.2.2实证分析与仿真分析13第四章 因特网抗毁性分析164.1因特网的拓扑结构及加权网络介绍164.2考虑信息条件下因特网的加权网络基于节点的抗毁性分析19 4.2.1因特网抗毁性模型描述19 4.2.2局部信息下的抗毁性仿真结果20 4.2.3全局信息下的抗毁性仿真结果234.3考虑信息条件下因特网的加权网络基于边的抗毁性分析26 4.3.1抗毁性模型26 4.3.2抗毁性仿真结果26第五章 总结32参考文献33谢 辞35基于复杂网络的因特网抗毁性分析摘要：自从上世纪末科学家们发现复杂网络的小世界及无标度结构特性以来，复杂网络就一直是最重要的多学科交叉研究领域之一。因特网作为现实世界中重要的复杂网络之一，其与安全性和可靠性相关的结构抗毁性一直是受关注的焦点。本篇文章首先阐明了复杂网络的基础理论，主要包括图论、复杂网络的结构参数、经典复杂网络拓扑模型。其次概述了复杂网络抗毁性研究现状。然后利用仿真软件Matlab，基于经典的加权无标度网络模型对因特网抗毁性开展仿真研究。本实验选定网络中较为常用的测度鲁棒性及网络效率，来评测网络的抗毁性。研究关注于网络在不同的信息条件和攻击方式下的抗毁性表现。从不同角度对因特网抗毁性进行了仿真，有基于节点的抗毁性仿真和基于边的抗毁性仿真。其中，在基于节点的抗毁性仿真中，研究了局部信息条件下遭遇ID攻击（对初始网络按节点的度大小顺序来移除节点）时网络的抗毁性，以及全局信息条件下遭遇IB攻击（对初始网络按照节点的介数大小顺序来移除节点）和RB攻击（每次移除的节点是当前网络中节点或边介数最大的节点）时网络的抗毁性；在基于边的抗毁性仿真中，研究了不同信息获取情况下网络的抗毁性。不同信息条件下基于节点的因特网抗毁性仿真结果表明：在基于局部拓扑信息的各种攻击策略下，权重系数越大，节点移除对于网络效率和连通性的影响越小，网络的抗毁性越强。相比于基于局部拓扑信息的攻击策略，基于全局信息的攻击策略下，网络应对节点移除的鲁棒性要更弱。而不同信息条件下基于边的因特网抗毁性仿真结果表明：当网络信息获取为随机获取时，网络的抗毁性取决于网络信息量的获取程度。其中，当网络的信息获取为零时，网络具有较强的抗毁性。在网络信息为优先获取情况下，只需要获取少量重要的边信息就会导致网络变得异常脆弱。这些研究结果为从信息角度给出因特网更合理有效的保护措施提供了理论依据。关键词：复杂网络；因特网；加权网络；小世界；抗毁性 1Invulnerability Based On The Complexity Of The Internet NetworkAbstract：Since the last century that scientists have found that the characteristics of the small-world and scale-free structure of complex networks, complex networks has been one of the most important multidisciplinary research field. Internet as one of the important real-world complex networks, and its safety and reliability associated with the structure of invulnerability has been the focus of attention.This article first clarifies the basic theory of complex networks, including graph theory, the structural parameters of complex networks, classical complex network topology model. Followed by an overview of the complex network of anti-status quo survivability research. Then use Matlab, classic weighted scale-free network model based on Internet invulnerability to carry out simulation studies. In this study, selected the more commonly used measure of the network - network efficiency and robustness to evaluating network survivability. Research focuses on network survivability performance under different conditions and attacks of the information. From a different perspective on the Internet invulnerability of the simulation, there is a node invulnerability simulation and survivability simulation based on the edge of the anti-Based. Among them, the survivability simulation based on nodes, studied under attack when confronted ID information under local conditions (initial node of the network according to the order to remove the node) network survivability, as well as the conditions encountered IB global information attack (the initial referral network in accordance with the number of nodes in the order to remove the node) and RB attacks (each node in the network to remove the maximum number of nodes or edges mediated current node) network of invulnerability; based side survivability simulation to study the different access to information under the network survivability. Under different conditions of node-based Internet Information invulnerability simulation results show that: in a variety of attack strategies based on local topology information, the greater the weighting factor, remove the node connectivity for network efficiency and the impact of the smaller network survivability the stronger sex. Compared to the attack strategy based on local topology information, under attack strategy based on global information network robustness to cope with weaker nodes removed. But under different conditions based on the information side of the Internet invulnerability simulation results show that: when the network is random access to information access, the network survivability depends on the extent of the network to get the amount of information. Wherein when the information network to get to zero, the network has a strong invulnerability. Priority access to network information case, you only need to get a small amount of side information will lead to important network tenuous. The results of these studies give a more reasonable and effective measures to protect the Internet from an information point of view provides a theoretical basis.Key words: complex networks; internet; weighted network; small world; invulnerability2第一章 绪论1.1研究背景及意义 自从上世纪九十年代以来，由于总体经济的高速发展，人们逐渐进入了一个以因特网为代表的网络时代，并且在生活中的各个领域因特网都广为普及，与我们的日常生活紧密相关。纵观整个