基于复杂性理论的军事通信网络中心化方法比较

 复杂网络的中心化有助于发现复杂网络中的重要节点,具有重要的应用价值。本文介绍了中心化指标的验证条件,讨论了常用的度中心、偏心率中心、邻近中心、子图中心和信息中心5种中心化方法,并以某一具体的军事通信网络为例进行了实例仿真应用。仿真结果显示,不同方法下得到的网络中心节点不同,反映出各种中心化方法侧重点的差异。分析表明,在实际网络中单一方法不能准确反映网络的重要节点；对于复杂网络的中心化问题,应结合具体应用背景,综合考虑几种中心化方法。

Comparision of Centrality Measures for Military Communication Network Based on Complexity Theory

Centralization of complex networks, which can help us find important nodes in complex networks, is of great practical value in many applications. Certification requirements of centrality index are introduced. Degree centrality, eccentricity centrality, closeness centrality, subgraph centrality and information centrality are discussed and their applicability is demonstrated on a certain military communication network. The simulation results show that the center node of network is different in different measures; it reflects differences in the emphasis of various centrality measures. Analysis shows that single measure can't reflect important nodes of networks exactly. As a result, several measures should be considered for the centralization analysis of complex networks based on specific application background.