Multi-objective Community Detection Method Using an Improved NSGA II Algorithm

Krista Rizman ┼Żalik


Complex networks, community detection, multiobjective optimization, evolutionary algorithms


Community detection methods provide help in understanding the structure and function of networks. Since single objective community detection methods are not able to detect multiple significant community structures, some methods formulate the community detection as multi-objective problems. This paper proposes a multi-objective approach using a combination of three objective functions in multi-objective framework that is able to discover wide specter of resulting partitions. The algorithm is based on the basic NSGA-II, and improvements are as follows. A hybrid model is established for population initialization and the mutation operator and the crossover operators compose only possible partitions of input network. Different three optimization functions are tested in order to detect a large spectrum of resulting partitions.

Important Links:

Go Back