A Novel Particle Swarm Optimization Algorithm for Network Clustering

The use of complex network analysis has gathered momenta in both theoretical and empirical studies. Network clustering plays an important role in network analysis. This paper models the networkclustering task as an optimization problem. A novel discrete particle swarm optimization algorithm is introduced to solve the modeled optimization problem. Particle swarm optimization is a stochastic searching algorithm, and it cannot avoid prematurity. To improve the performance of the algorithm, a new particle status update principle is defined, a novel turbulence operation is proposed to improve exploration, and a novel local search strategy is developed to enhance exploitation. Extensive experiments on both synthetic and real-world networks are carried out. Several state-of-the-art network-clustering approaches are compared with the proposed method. The experiments demonstrate that the proposed method is effective and promising.