1.College of Electronic and Optical Engineering, College of Microelectronics, Nanjing University of Posts and Telecommunications, Nanjing 210023, China 2.National and Local Joint Engineering Laboratory of RF Integration and Micro-Assembly Technology, Nanjing 210003, China
Fund Project:Project supported by the National Natural Science Foundation of China (Grant Nos. 71671093, 61271334).
Received Date:16 January 2019
Accepted Date:05 April 2019
Available Online:01 June 2019
Published Online:20 June 2019
Abstract:The measurement of node importance is significant for analyzing a network structure. To comprehensively reflect the global and local network features, in this paper we abstract the propagating process of epidemic diseases based on the network topology structure, and then respectively sets each node as an infection source. After a certain dissemination time K, the number of infected nodes in the network is defined as the K-order propagation number, and the weighted sum of K-order propagation numbers under different values of K is taken as the important index of nodes. The simulation experiments of Watts-Strogatz small-world networks and a dolphin network show that the weighted K-order propagation number algorithm is more effective than the traditional method in evaluating the importance of nodes. For the Watts-Strogatz small-world networks, it can reflect the influence of long-distance connections on information transmission in detail. For the dolphin network, the weighted K-order propagation number algorithm significantly raises the profile of those nodes which play a key role in the information communication between different dolphin communities. In addition, in this paper we use a deliberate attacking method to analyze the western power grid of the United States, the road transportation network of the Chicago region, the co-authorship network in network science and the axonal tracts’ network between neurons of mouse. According to the specific order of the node importance from high to low, network nodes are attacked in turn, that is, all edges of the attacked nodes are removed. The analysis results of network parameters such as the network efficiency and the node number of the maximum sub-graph changing with the attacking times show that comparing with other evaluation indices of node importance such as degree, Ren method, Chen method, eigenvector method, Katz index, PageRank, CI method and K-shell, the weighted K-order propagation number algorithm focuses much on destroying the major structure, and all of the above four networks will break down if only a small number of important nodes are attacked. For example, after attacking only 100 nodes, the network efficiency of the western power grid of the United States is down by 90%, and after attacking 200 nodes, the network scale of the maximum sub-graph is nearly 3% of the original network. Keywords:complex network/ node importance/ propagation model
小世界网络是指同时具有较短特征路径长度以及较大平均聚类系数的一种网络类型. Watts和Strogatz[25]最早提出了一种构造小世界网络的方法, 即将最近邻耦合网络中的边依概率进行随机重连, 通常将这种网络称为WS小世界网络. 图1(a)基于100节点最近邻耦合网络随机生成了一个WS小世界网络, 该最近邻耦合网络中每个节点与其邻近的左右两侧各两个节点相连. 在随机重连后, 边31-33和95-96分别被改连为31-72和95-53, 见图1(b)和图1(c), 其余边的位置不变. 图 1 随机生成的100节点WS小世界网络 (a) 网络结构; (b) 边31-33被改连至边31-72; (c) 边95-96被改连至边95-53 Figure1. A random WS small-world network with 100 nodes: (a) Network structure; (b) the edge 31-33 is rescheduled to the edge 31-72; (c) the edge 95-96 is rescheduled to the edge 95-53.
表1WS小世界网络以及同规模随机网络的网络参数 Table1.Network features of the WS small-world network and random networks.
现利用加权K-阶传播数法对该网络节点的重要性进行分析. 图2(a)为网络的K-阶结构熵${H^K}$, 图2(b)为权重系数${c^K}$, 图2(c)为归一化的节点重要性. 此外, 图2(c)中的小图按色度对重要性进行了标注. 图 2 基于加权K-阶传播数法的WS小世界网络节点重要性结果 (a) K-阶结构熵; (b) 权重系数; (c) 节点重要性 Figure2. Node importance of the WS small-world network based on the weighted K-order propagation number algorithm: (a) The K-order structure entropy; (b) the weight coefficient; (c) the importance of nodes.
为了进一步验证加权K-阶传播数法的有效性, 本文以海豚网络为例进行研究. Lusseau等[26,27]对新西兰道特富尔峡湾(Doubtful Sound) 62只宽吻海豚进行研究并构建了海豚社会网络. 基于加权K-阶传播数法的海豚网络节点重要性结果如图4所示, 图4(a)为网络的K-阶结构熵${H^K}$, 图4(b)为权重系数${c^K}$, 图4(c)为归一化后的节点重要性. 网络中节点代表海豚, 边表示海豚间的相关接触. 图 4 基于加权K-阶传播数法的海豚网络节点重要性结果 (a) K-阶结构熵; (b) 权重系数; (c) 节点重要性 Figure4. Node importance of the dolphin network based on the weighted K-order propagation number algorithm: (a) The K-order structure entropy; (b) the weight coefficient; (c) the importance of nodes.
为了进一步验证加权K-阶传播数法的优越性, 本文对美国西部电网[25]、芝加哥公路网络[28]、网络科学家合著网络[29]以及小鼠神经纤维束网络(Kasthuri_graph_v4)[30]进行了节点重要性研究. 其中美国西部电网共有4941个节点和6594条边, 描述了美国西部的高压输电电网结构; 芝加哥公路网络共有1467个节点和1298条边, 描述了美国芝加哥地区的公路运输情况; 网络科学家合著网络共有1589个节点和2742条边, 描述了从事网络理论和实验的科学家合著关系; 小鼠神经纤维束网络共有1029个节点和1700条边, 描述了小鼠部分脑皮层中神经元间的轴突束连接. 美国西部电网、芝加哥公路网络和网络科学家合著网络均是无权无向的; 小鼠神经纤维束网络是有权无向的, 为了适应加权K-阶传播数法, 本文忽略了该网络的边权信息. 以上网络的K-阶结构熵HK如图5所示. 图 5K-阶结构熵 (a) 美国西部电网; (b) 芝加哥公路网络; (c) 科学家合作网络; (d) 小鼠神经纤维束网络 Figure5. The K-order structure entropy: (a) The western power grid of the United States; (b) the road transportation network of the Chicago region; (c) the co-authorship network in network science; (d) the axonal tracts network between neurons of mouse.
其中${e_0}$为未被攻击的原始网络的网络效率. 由(7)式可见, ε的取值为$[0,\;1]$, 当网络未被攻击时$\varepsilon = 0$, 当网络被攻击时ε会随之升高; 最终, 当所有的边被删除时网络效率降至最低, 此时$\varepsilon = 1$. 此外, 为了进一步分析攻击前后网络拓扑结构的变化情况, 依照文献[33]将网络最大子图节点数设为γ. 图6和图7分别绘制了美国西部电网、芝加哥公路网络、网络科学家合著网络以及小鼠神经纤维束网络基于不同节点重要性的排序方法, 其网络效率下降率ε和最大子图节点数γ随攻击次数的变化曲线. 其中, Katz指标的权重衰减因子为1.5倍邻接矩阵最大特征值的倒数; PageRank算法的阻尼系数取0.85; CI法的权重系数取0.5. 图 6 网络效率下降率ε随攻击次数的变化情况 (a) 美国西部电网; (b) 芝加哥公路网络; (c) 网络科学家合著网络; (d) 小鼠神经纤维束网络 Figure6. Change of the network efficiency decrease rate ε with attacking times: (a) The western power grid of the United States; (b) the road transportation network of the Chicago region; (c) the co-authorship network in network science; (d) the axonal tracts network between neurons of mouse.
图 7 最大子图节点数γ随攻击次数的变化情况 (a) 美国西部电网; (b) 芝加哥公路网络; (c) 网络科学家合著网络; (d) 小鼠神经纤维束网络 Figure7. Change of the node number of the maximum sub-graph ε with attacking times: (a) The western power grid of the United States; (b) the road transportation network of the Chicago region; (c) the co-authorship network in network science; (d) the axonal tracts network between neurons of mouse.