Detecting community structure of networks using evolutionary coordination games

Community detection is a well-studied problem in network science. In this paper, a novel community detection algorithm based on evolutionary game dynamics is proposed, where individuals hold disperse opinions as their mixed strategies and play coordination games with their connected individuals in a network. Employing strategy updating processes among the population, the individuals finally fall into clusters of different opinionists which correspond to a community partition of the network. Using Zachary´s karate club network as a benchmark test, the validity of the proposed community detection method is verified.