A Discrete Random Drift Particle Swarm Optimization with Modularity in Community Detection

Abstract

In the field of complex networks, community detection is one of important research objects. To solve the problem of poor quality and the unstable result with community structure, we propose a community detection optimization algorithm based on random drift particle swarm optimization (RDPSO) algorithm (DRDPSO-net), in which we use discrete method to update the network information. Through the discrete particle evolution process and local greedy strategy with network topology character, DRDPSO-net can obtain a better quality of community division. In addition, several representative real networks are used to verify the performance of DRDPSO-net. By comparing them across several algorithms, DRDPSO-net has more desirable value among those algorithms. Furthermore, the experimental results demonstrated that DRDPSO-net obtain a valid and steady community structure.