聚类系数指标对复杂网络鲁棒性的影响分析
|
摘要
分析了采用度分布相同且聚类系数不同的三种类型网络(中性网络、同配网络和异配网络)在遇到随机故障或者蓄意攻击时,网络的初始聚类系数变化对网络鲁棒性的影响.实验分析表明,网络的初始聚类系数越大,网络在受到随机故障或蓄意攻击时网络中最大连通子图的直径和网络中最大连通子图的平均路径长度的起伏也就越大.初始聚类系数的变化在异配网中对网络鲁棒性的作用最明显,中性网次之,对同配网的鲁棒性不明显.
The complex networks can be divided into three types according to chaining rules of their nodes, namely neutral network, assortative network, and hetero-assortative network. In this paper, three types of network with identical degree distribution and different clustering coefficient are used to analyze the influence of their initial clustering coefficient on their robustness when they are subjected to random failure and deliberate attack. Experimental analysis shows that the larger the initial clustering coefficient of the network is, the larger the fluctuation of the diameter and average path length of the maximum connected subgraph in network will be when the network is subjected to random failures or deliberate attacks. And the effect of initial clustering coefficient in hetero-assortative network on the network robustness will be most obvious, the effect will be less for neutral network, and there will be no obvious effect for assortative network.
The complex networks can be divided into three types according to chaining rules of their nodes, namely neutral network, assortative network, and hetero-assortative network. In this paper, three types of network with identical degree distribution and different clustering coefficient are used to analyze the influence of their initial clustering coefficient on their robustness when they are subjected to random failure and deliberate attack. Experimental analysis shows that the larger the initial clustering coefficient of the network is, the larger the fluctuation of the diameter and average path length of the maximum connected subgraph in network will be when the network is subjected to random failures or deliberate attacks. And the effect of initial clustering coefficient in hetero-assortative network on the network robustness will be most obvious, the effect will be less for neutral network, and there will be no obvious effect for assortative network.
引文
[1] PATEL S J,PATTEWAR T M.Software birthmark based theft detection of JavaScript programs using agglomerative clustering and frequent subgraph minming [C]//Embedded System (ICES),2014 International Conference on.[S.l.]:IEEE,2014.
[2] ALBERT R,JEONG H,SI A L.Error and attack tolerance of complex networks [J].Nature,2000,406(4):378-382.
[3] CRUCITTI P,LATORA V,MARCHIORI M,et al.Error and attack tolerance of complex networks [J].Physica A,2004,340:388-397.
[4] LIU Suyu,RONG Gang.Analysis on refinery system as a complex task-resource network [J].Chinese Journal of Chemical Engineering,2013,21(3):253-262.
[5] 毛凯.复杂网络结构的稳定性与鲁棒性研究 [J].计算机科学,2015,42:85-88.
[6] MAO J,GAO J,LIU Y.Power allocation over finding cognitive MIMO channels:an ergodic capacity perspective [J].Trans Veh Technol,2016,61:1162-1173.
[7] 陆靖桥,傅秀芬,蒙在桥.复杂网络的鲁棒性与中心性指标的研究 [J].计算机应用与软件,2016,33:302-310.
[8] MYLES G,COLLBERG C.Softwate watermarking via opaque predicates:Implementation analysis and attacks [J].Electronic Commerce Research,2006,6(2):155-171.
[9] MYLES G,COLLBERG C,HEIDEPEIEM Z,et al.The evaluation of two software watermaking algorithms [J].Software:Practice and Experience,2015,35(10):923-938.
[10] MELIáN C J,BASCOMPTE J.Complex netwokrs:two ways to be robust [J].Ecology Letters,2010,5(6):705-708.
[11] 刘飞飞,蔺靖娜,刘潇潇.基于贝叶斯复杂网络的复杂网络攻击方法研究 [J].计算机工程与应用,2017(53):18-25.
[12] CHRISTIAN S.Shortest-path queries in static networks [J].ACM Computing Surveys,2014,46(4):1-31.
[13] 刘宏鲲,周涛.中国城市航空网络的实证研究与分析 [J].物理学报,2007,56(5):106-112.
[14] 刘一奎,刘天琪,李茜,等.一种高聚类系数的无标度网络演化模型 [J].网络安全技术与应用,2015(1):55-56.
[2] ALBERT R,JEONG H,SI A L.Error and attack tolerance of complex networks [J].Nature,2000,406(4):378-382.
[3] CRUCITTI P,LATORA V,MARCHIORI M,et al.Error and attack tolerance of complex networks [J].Physica A,2004,340:388-397.
[4] LIU Suyu,RONG Gang.Analysis on refinery system as a complex task-resource network [J].Chinese Journal of Chemical Engineering,2013,21(3):253-262.
[5] 毛凯.复杂网络结构的稳定性与鲁棒性研究 [J].计算机科学,2015,42:85-88.
[6] MAO J,GAO J,LIU Y.Power allocation over finding cognitive MIMO channels:an ergodic capacity perspective [J].Trans Veh Technol,2016,61:1162-1173.
[7] 陆靖桥,傅秀芬,蒙在桥.复杂网络的鲁棒性与中心性指标的研究 [J].计算机应用与软件,2016,33:302-310.
[8] MYLES G,COLLBERG C.Softwate watermarking via opaque predicates:Implementation analysis and attacks [J].Electronic Commerce Research,2006,6(2):155-171.
[9] MYLES G,COLLBERG C,HEIDEPEIEM Z,et al.The evaluation of two software watermaking algorithms [J].Software:Practice and Experience,2015,35(10):923-938.
[10] MELIáN C J,BASCOMPTE J.Complex netwokrs:two ways to be robust [J].Ecology Letters,2010,5(6):705-708.
[11] 刘飞飞,蔺靖娜,刘潇潇.基于贝叶斯复杂网络的复杂网络攻击方法研究 [J].计算机工程与应用,2017(53):18-25.
[12] CHRISTIAN S.Shortest-path queries in static networks [J].ACM Computing Surveys,2014,46(4):1-31.
[13] 刘宏鲲,周涛.中国城市航空网络的实证研究与分析 [J].物理学报,2007,56(5):106-112.
[14] 刘一奎,刘天琪,李茜,等.一种高聚类系数的无标度网络演化模型 [J].网络安全技术与应用,2015(1):55-56.