An Improved Multi-objective Genetic Algorithm for Revealing Community Structures of Complex Networks

Abstract

Community detection a crucial task in the study of complex networks aims at identifying structural patterns of the networks. Recently, evolutionary methods are successfully applied to reveal communities of complex networks. Most of them employ only one quality measure in their search processes. Since each objective covers a different aspect of network’s property, investigating this problem with more than one objectives results in identifying more accurate community structure. To handle this issue in this paper, a multi-objective genetic algorithm integrated with a local search strategy called Enhanced Multi-Objective Genetic Algorithm for Community Detection (EMOGACD) is proposed. The main goal of using the local search strategy is speeding up the convergence and improving the accuracy of the proposed method. the proposed method uses the vector-based method is used to represent the solutions. This type of representation reduces the search space and does not need to know the number of communities at the beginning of the search process. Performed experiments performed on both real-world and synthetic networks demonstrate the relatively high capacity of the proposed method in detecting high quality communities within lower generations.