Community detection algorithm with locally social spider optimized

Community detection in complex networks is very important for decision-making in the real world. Swarm intelligence optimization is an effective method for community detection. However, such algorithms are easy to fall into local optima and tend to ignore smaller communities. This paper proposed a community detection algorithm with locally social spider optimized (LSSO/CD). The network nodes and their relationships are initialized as spider populations, and the populations evolve by females and males respectively. The fitness function is defined by the degree of close connection among nodes, and the modular increment of community is used as the criterion of evolution. The whole process starts from a variety of local groups and the network is divided step by step. The results show that LSSO/CD can effectively balance global and local convergence, and solve complex network partition problem very well.