A multi-objective evolutionary algorithm based on mixed encoding for community detection

Community structure is one of the most significant features in complex networks and community detection a crucial method to analyze structure. Existing representations have characteristics inflexibility easily generate invalid solutions. To address drawbacks, this paper proposed multi-objective evolutionary algorithm based on mixed encoding (MOGAME). The combines locus-based representation labels-based representation, which can avoid generating solution improve performance. Extensive experiments both synthetic real-word show that performs better than existing algorithms with respect accuracy stability.