一种基于最大流的网络结构熵

熵是可用来反映网络结构异质性的指标.针对传统熵指标不能很好反映网络全局异构性的不足,本文引入网络流的概念,综合考虑径向测度和中间测度,提出一种新的网络结构熵.特殊网络(如公用数据集Dolphins网络)的分析结果表明,本文提出的熵指标在一定程度上克服了其他网络熵指标的不足,更能够反映网络的真实拓扑结构;对随机网络、最近邻耦合网络、星型网络、无标度网络、Benchmark网络和小世界网络等典型网络的理论分析和仿真实验,进一步证明本文提出的熵指标在刻画一般复杂网络结构特征上的有效性和适用性.

A new network structure entropy based on maximum flow

Entropy is an index to reflect the heterogeneity of network structure. By introducing the concept of network flow which comprehensively considers radial measurement and betweenness measurement, we define a new network structure entropy index to solve the problem that classical entropy indices cannot effectively reflect heterogeneity of the global network. Analysis results concerning specific network (e.g. public data set Dolphins network) indicate that this new entropy index can reflect the real topological structure of network, and effectively overcome the shortcomings of other network entropy indices to some extent. The theoretical analyses and simulation experiments on Erdös-Renyi random network, nearest-neighbor coupled network, star network, Barabási-Albert scale-free network, Benchmark network, and the Watts-Strogatz small-world network further prove the effectiveness and applicability of this new network structure entropy index to describe the characteristics of ordinary complex network structures.
Keywords：complex networkheterogeneityentropy