T-S Fuzzy Model Identification With a Gravitational Search-Based Hyperplane Clustering Algorithm

• Published in: IEEE transactions on fuzzy systems Vol. 20; no. 2; pp. 305 - 317
• Main Authors: Li, Chaoshun, Zhou, Jianzhong, Fu, Bo, Kou, Pangao, Xiao, Jian
• Format: Journal Article
• Language: English
• Published: IEEE 01.04.2012
• Subjects: Clustering algorithms; Data models; Fuzzy model identification; gravitational search algorithm (GSA); hyperplane clustering; Optimization; Parameter estimation; Partitioning algorithms; Prototypes; Takagi-Sugeno (T-S) fuzzy model; Vectors
• ISSN: 1063-6706; 1941-0034
• DOI: 10.1109/TFUZZ.2011.2173693

In order to improve the performance of the fuzzy clustering algorithm in fuzzy space partition in the identification of the Takagi-Sugeno (T-S) fuzzy model, a hyperplane prototype fuzzy clustering model is proposed. To solve the clustering objective function, which could not be handled by the gradient method as the traditional clustering method fuzzy c-means does, a newly developed excellent global search method, which is the gravitational search algorithm (GSA), is employed. Then, the GSA-based hyperplane clustering algorithm (GSHPC) is proposed and illuminated. GSHPC is used to partition the fuzzy space and identify premise parameters of the T-S fuzzy model, and orthogonal least squares is exploited to identify the consequent parameters. Comparative experiments are designed to verify the validity of the proposed clustering algorithm and the T-S fuzzy model identification method, and the results show that the new method is effective in describing a complicated nonlinear system with significantly high accuracies compared with approaches in the literature.