Complex Network Clustering by Multiobjective Discrete Particle Swarm Optimization Based on Decomposition

The field of complex network clustering has been very active in the past several years. In this paper, a discrete framework of the particle swarm optimization algorithm is proposed. Based on the proposed discrete framework, a multiobjective discrete particle swarm optimization algorithm is proposed...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on evolutionary computation Ročník 18; číslo 1; s. 82 - 97
Hlavní autoři: Gong, Maoguo, Cai, Qing, Chen, Xiaowei, Ma, Lijia
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 01.02.2014
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
Algorithms
Clustering
Clustering algorithms
Complex networks
Computational fluid dynamics
Decomposition
evolutionary algorithm
Fluid flow
multiobjective optimization
Networks
Optimization
Optimization algorithms
Particle swarm optimization
Sociology
Statistics
Swarm intelligence
Turbulence
ISSN:1089-778X, 1941-0026
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The field of complex network clustering has been very active in the past several years. In this paper, a discrete framework of the particle swarm optimization algorithm is proposed. Based on the proposed discrete framework, a multiobjective discrete particle swarm optimization algorithm is proposed to solve the network clustering problem. The decomposition mechanism is adopted. A problem-specific population initialization method based on label propagation and a turbulence operator are introduced. In the proposed method, two evaluation objectives termed as kernel k-means and ratio cut are to be minimized. However, the two objectives can only be used to handle unsigned networks. In order to deal with signed networks, they have been extended to the signed version. The clustering performances of the proposed algorithm have been validated on signed networks and unsigned networks. Extensive experimental studies compared with ten state-of-the-art approaches prove that the proposed algorithm is effective and promising.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1089-778X
1941-0026
DOI:10.1109/TEVC.2013.2260862