Inverse modelling-based multi-objective evolutionary algorithm with decomposition for community detection in complex networks

Community structure is an important topological property of complex networks representing real-world systems, and it is believed to be a highly important tool for understanding how complex networks are organized and function. Generally, community detection can be considered to be a single-objective or multi-objective optimization problem, and a great number of population-based optimization algorithms have been explored to address this problem in the past several decades. In this study, we present a novel discrete inverse modelling-based multi-objective evolutionary algorithm with decomposition (DIM-MOEA/D) for community detection in complex networks. First, the population is initialized by a problem-specific method based on label propagation. Next, inverse models based on the network topology are constructed to generate offspring by sampling the objective space, and the problem-specific mutation is introduced to maintain the diversity of the population and avoid being trapped in the local optima. Next, the decomposition-based selection is introduced as the updating rule of individuals. Finally, several real-world networks are considered to evaluate the performance of the proposed algorithm. The experimental results demonstrate that compared with the state-of-the-art approaches, DIM-MOEA/D is an effective and promising method for solving community detection in complex networks.