中文摘要：

社区发现是复杂网络研究中的一个重要领域,且应用广泛,但目前已有的大多数算法都需采用社区评判函数来确定社区结构的划分,且仅能得到一种划分结果。引入宇宙星系模型和万有引力定律,基于引力思想提出一种新的复杂网络社区发现算法,为网络中节点赋予质量并构建出社区框架,继而利用引力作用完成社区结构划分,并可对发现社区的粒度大小进行选择以得到多种划分结果,无需先验知识及相关参数。通过真实网络实验验证,并与现有的社区发现算法比较,本文提出的算法能有效且较为准确地挖掘出复杂网络中的社区结构。

英文摘要：

Community detection is an important field in the study of complex networks,and it is widely applied. But for most of the existing algorithms at present,community structure is determined by some community evaluation function,and only one division result can be obtained. Referenced from the galaxy model and the law of universal gravitation,a new community detection algorithm of complex network based on gravitational search is proposed,nodes in a network are given quality,and community framework is built. Then community structure is divided via gravitation. The granularity of the detected communities can be selected,and thereby a variety of division results can be obtained,without prior knowledge and the related parameters. Experiments in real networks,and comparison with other pre-existing community detection algorithms prove that,the community structure of complex networks can be effectively and accurately excavated via the proposed algorithm.