一种具有指数截断和局部集聚特性的网络模型

针对真实网络局域演化的特点,提出了一种具有局部集聚特性的网络演化模型——局部集聚模型(LC模型). 理论分析和模拟实验表明,LC模型的节点度服从一种具有指数截断的幂律分布,同时它的平均聚类系数要远大于局域世界模型,接近真实网络. 模拟了LC模型对恶意攻击和随机错误的抵抗力,发现高聚类系数的LC模型对恶意攻击更加脆弱. 关键词： 局部集聚 指数截断 脆弱性 无标度网络

A network model with exponential cut-off degree distribution and local-cluster property

According to the character of local evolving mechanism in real network, we propose a new local-clustered network evolving model (LC model). Theoretical analysis and numerical simulations show that the LC model obeys a power-law degree distribution with an exponential cut-off, and its cluster coefficient is much bigger than the local-world model (LW model) proposed by Li and Chen and is more, close to real network. We simulate the performance of LC model against random errors and malice attacks, and discover that LC model is more fragile under attacks because of its large cluster coefficient.