Journal of Bionic Engineering ›› 2023, Vol. 20 ›› Issue (5): 2331-2358.doi: 10.1007/s42235-023-00387-1
Discrete Improved Grey Wolf Optimizer for Community Detection
Mohammad H. Nadimi‑Shahraki1,2; Ebrahim Moeini1,2; Shokooh Taghian1,2; Seyedali Mirjalili3,4
- 1 Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad 8514143131, Iran 2 Big Data Research Center, Najafabad Branch, Islamic Azad University, Najafabad 8514143131, Iran 3 Centre for Artifcial Intelligence Research and Optimisation, Torrens University, Brisbane 4006, Australia 4 Yonsei Frontier Lab, Yonsei University, Seoul, South Korea
Discrete Improved Grey Wolf Optimizer for Community Detection
Mohammad H. Nadimi‑Shahraki1,2; Ebrahim Moeini1,2; Shokooh Taghian1,2; Seyedali Mirjalili3,4
- 1 Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad 8514143131, Iran 2 Big Data Research Center, Najafabad Branch, Islamic Azad University, Najafabad 8514143131, Iran 3 Centre for Artifcial Intelligence Research and Optimisation, Torrens University, Brisbane 4006, Australia 4 Yonsei Frontier Lab, Yonsei University, Seoul, South Korea
摘要： Detecting communities in real and complex networks is a highly contested topic in network analysis. Although many metaheuristic-based algorithms for community detection have been proposed, they still cannot effectively fulfill large-scale and real-world networks. Thus, this paper presents a new discrete version of the Improved Grey Wolf Optimizer (I-GWO) algorithm named DI-GWOCD for effectively detecting communities of different networks. In the proposed DI-GWOCD algorithm, I-GWO is first armed using a local search strategy to discover and improve nodes placed in improper communities and increase its ability to search for a better solution. Then a novel Binary Distance Vector (BDV) is introduced to calculate the wolves’ distances and adapt I-GWO for solving the discrete community detection problem. The performance of the proposed DI-GWOCD was evaluated in terms of modularity, NMI, and the number of detected communities conducted by some well-known real-world network datasets. The experimental results were compared with the state-of-the-art algorithms and statistically analyzed using the Friedman and Wilcoxon tests. The comparison and the statistical analysis show that the proposed DI-GWOCD can detect the communities with higher quality than other comparative algorithms.