A Fast Algorithm for Community Detection of Network Systems in Smart City

• Published in: IEEE access Vol. 7; pp. 51856 - 51865
• Main Authors: Liu, Fangyu, Xie, Gang
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; BBO algorithm; Clustering algorithms; community detection; complex network; Complex networks; Heuristic algorithms; Indexes; Mathematical model; NMW small world network; Nodes; Optimization; Partitioning algorithms; Robustness (mathematics); Smart cities; smart city
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2019.2910602

In this paper, a novel algorithm is designed to detect the community structure of network systems in the smart city based on the biogeography-based optimization (BBO) algorithm and the Newman, Moore, and Watts (NMW) small-world network. We have incorporated the NMW small-world network to the BBO algorithm to enhance the ability of migration of the habitat by using the connection mechanism of the NMW small-world network. With the help of small-world network information sharing, the convergence speed of the BBO algorithm has significantly improved. The first step of the algorithm design is to generate an NMW small-world network containing nodes equal to the number of habitats with good connectivity, which facilitates better information exchange between the nodes. In the second step, the habitat in the BBO algorithm is dynamically assigned to the small world network, and then, the BBO algorithm migrates and mutates according to the connection relationship of the NMW small-world network. Finally, the new designed NMW-BBO algorithm is evaluated for community detection via four real networks and computer-generated networks, and one of them is exhibited the characteristics of a large network. The numeric simulations are also employed to demonstrate that the new algorithm exhibits better accuracy and robustness.