基于网络局域相同步的社团层次结构划分

Community detection based on local phase synchronization in complex networks

网络结构和网络上的动力学行为密不可分,网络结构影响网络上的动力学行为,反之网络上的动力学行为也能体现网络的结构,包括利用动力学过程进行群落结构的检测。文章假设网络同步过程分为局域同步行为和全局同步行为,群落结构导致局域同步趋同性。然后设计了一个同步模型,该模型通过抑制全局同步仅保留局域同步,并利用网络的稳态局域同步进行群落检测。将该局域同步模型应用于模型网络和实际网络,并与经典模块度极值优化算法对比,验证了本方法的有效性。

The dynamics in complex networks is relevant to the network topology and the phenomenon could be utilized to detect communities. In the paper, we suppose that the synchronization on complex networks could be divided into local and collective process and community structure leads to local synchronization. Then, we proposed a scheme via suppressing the collective synchronization and locking the local synchronization（phase locking） at a stable state at the same time. Through this scheme, community structure can naturally emerge in paths to synchronization. Finally, the community detection based on the scheme is performed on artificial networks and five real networks and the observed community structures are much better than the classical methods.