Detecting community structure in complex networks using simulated annealing with kk-means algorithms

Identifying the community structure in a complex network has been addressed in many different ways. In this paper, the simulated annealing strategy is used to maximize the modularity of a network, associating with a dissimilarity-index-based and with a diffusion-distance-based kk-means iterative procedure. The proposed algorithms outperform most existing methods in the literature as regards the optimal modularity found. They can not only identify the community structure, but also give the central node of each community during the cooling process. An appropriate number of communities can be efficiently determined without any prior knowledge about the community structure. The computational results for several artificial and real-world networks confirm the capability of the algorithms.