Discovering communities from disjoint complex networks using Multi-Layer Ant Colony Optimization

Discovering communities is one of the important features of complex networks, as it reveals the structural features within such networks. Community detection is an optimization problem, and there have been significant efforts devoted to detecting communities with dense intra-links. However, single objective optimization approaches are inadequate for complex networks. In this work, we propose the Multi-Layer Ant Colony Optimization (MLACO) to detect communities in complex networks. This algorithm takes Ratio Cut (RC) and Kernel K-means (KKM) as an objective function and attempts to find the optimal solution. The findings from our evaluation of MLACO using both synthetic and real world complex networks demonstrate that it outperforms other competing approaches, in terms of normalized mutual information (NMI) and modularity (Q). Moreover, we also evaluate our algorithm for small-scale and large-scale networks showing the utility of our proposed approach. (C) 2020 Elsevier B.V. All rights reserved.

Automated summary

Discovering communities is important in complex networks, and community detection is an optimization problem. The proposed MLACO algorithm, using RC and KKM as objective functions, outperforms other methods, demonstrating its utility.