一种新型二分网络类局域世界演化模型

A novel local-world-like evolving bipartite network model

现实世界中复杂网络的演化存在很明显的局域选择现象,然而目前关于二分网络中的局域世界演化模型研究较少.因此,本文建立了一个基于二分网络的类局域世界演化模型.首先定义了网络节点度值的饱和度.在此基础上提出了一种新型二分网络局域世界演化模型.新节点加入系统不需要全局知识,而是通过节点在网络演化的不同时刻度值饱和度为选择条件构造新节点的局域世界,然后利用择优连接从局域世界中选择节点增加连边完成网络演化.此类模型中新节点的局域世界是通过节点饱和度的限制被动生成,因此又称为类局域世界模型.通过模拟分析发现在节点度值饱和度的限制下择优连接并没有产生具有幂率特性的度分布,而是生成了度分布相对均匀的二分网络,即节点度值分布区间较小.此外,本文还给出了该网络的混合系数计算结果,该结果显示网络同配性与网络参数的选择有关,这一结果与网络邻点平均度的模拟结果一致.

In complex networks, node degree values are limited by some practical factors. The saturation of node degree, which is a function of network evolution time, is defined first. We propose a novel evolving bipartite network model based on preferential attachment in local-world, which is generated by node saturation restrictions, not new node selection. So we also call it local-world-like model. However, the numerical simulation results display that the degree distribution does not obey the power-law distribution. We find that the degree value interval of this local-world-like bipartite network is small. There is no hub node. In addition to these, we analyze mixing coefficient of the network and find that the assortativities of the network are different when the network is generated by different initial parameters Such a result accords with our simulated result.