基于多种群遗传算法的复杂网络社区结构发现

Community structure discovery in complex networks based on multi-population genetic algorithm

提出了一种基于多种群遗传算法的复杂网络社区结构发现新算法,该算法无须预先知道社区内节点的数量以及任何门限值,同时引入并行遗传算法的思想,进一步提高了算法的运行效率。实验结果表明,与传统算法相比,在无先验信息的条件下,使用该算法对不同规模的网络图Zachary和Dophins网络结构进行验证时,能够以较低的时间复杂度、高效并准确地完成对网络社区的有效划分。

This paper proposed a new model to detect community structure in complex network based on multiple-population genetic algorithm.It didn＇t need any prior knowledge about the numbers of community and any threshold values,introduced simultaneously parallel genetic algorithm to enhance the efficiency.The numerical experiments show that this algorithm can greatly reduce the time complexity and get more accurate optimum partiton of network structure without any prior information compared with traditional algorithm,by using this new algorithm to test the two networks with different scale named Zachary and Dophins.