Detecting Communities in Complex Networks Using an Adaptive Genetic Algorithm and Node Similarity-Based Encoding. (28th February 2023)

• Record Type: Journal Article
• Authors: Hesamipour, Sajjad; Balafar, Mohammad Ali; Mousazadeh, Saeed
• Other Names: Selişteanu Dan Academic Editor.
• Is Part Of: Complexity. Volume 2023(2023)
• Publication Date: 2023-02-28
• Subjects: Chaotic behavior in systems -- Periodicals; Complexity (Philosophy) -- Periodicals; 003
• DOI: 10.1155/2023/4796536
• Languages: English
• ISSNs: 1076-2787

Abstract

Abstract : Detecting communities in complex networks can shed light on the essential characteristics and functions of the modeled phenomena. This topic has attracted researchers from both academia and industry. Among different community detection methods, genetic algorithms (GAs) have become popular. Considering the drawbacks of the currently used locus-based and solution-vector-based encodings to represent the individuals, in this article, we propose (1) a new node similarity-based encoding method, MST-based encoding, to represent a network partition as an individual, which can avoid the shortcomings of the previous encoding schemes. Then, we propose (2) a new adaptive genetic algorithm for the purpose of detecting communities in networks, along with (3) a new initial population generation function to improve the convergence time of the algorithm, and (4) a new sine-based adaptive mutation function which adjusts the mutations according to the improvement in the fitness value of the best individual in the population pool. The proposed method combines the similarity-based and modularity-optimization-based approaches to find communities in complex networks in an evolutionary framework. Besides the fact that the proposed encoding can avoid meaningless mutations or disconnected communities, we show that the new initial population generation function and the new adaptive mutation function can improve the convergence time of the algorithm. Several experiments and comparisons were conducted to verify the effectiveness of the proposed method using modularity and NMI measures on both real-world and synthetic datasets. The results show that the proposed method can find the communities in a significantly shorter time than other GAs while reaching a better trade-off in the different measures.