一個基於蟻群演算法之新式模糊建模方法

A Novel Fuzzy Modeling Method Based on Ant Colony Algorithm

葉建廷(Cheng-Ting Yeh)

指導教授 ： 蔡舜宏

國立臺北科技大學/機電學院/自動化科技研究所/碩士(2014年)

https://doi.org/10.6841/NTUT.2014.00716

摘要

本文針對Takagi-Sugeno(T-S)模糊模型，提出一個新式的建模方法，首先將樣本資料點用相對式模糊C平均值(Alternative Fuzzy C-Means, AFCM)做分群並利用Xie-Beni指數標準找出合適的分群數，且以此數目作為規則數，將目標系統分成數個線性子系統。此外，結合蟻群最佳化演算法(Ant Colony Optimization, ACO)和模糊C回歸模型(Fuzzy C-Regression Model, FCRM)演算法以找出資料點和線性系統的模糊關係，建立出模糊規則參數初始值。最後再用權重遞推最小平方法(weight recursive least square)得到系統參數的最佳值，以建構出T-S模糊模型。
最後，我們使用本文的方法針對一些範例進行建模，並與其它方法做比較，經由實驗結果證明了本文提出的方法所建構的模型，能夠更加的近似複雜的目標系統。

關鍵字

T-S模糊模型 ； 模糊分群法 ； 蟻群演算法 ； 參數估計

並列摘要

In this thesis, a novel modeling method for Takagi-Sugeno (T-S) fuzzy model is proposed. At first, the sample data points are classified by alternative fuzzy c-means (AFCM) algorithm. Based on Xie-Beni index criterion, the optimal numbers of cluster can be obtained and then the numbers of cluster numbers are set as the rule numbers of fuzzy. In addition, by utilizing fuzzy c-regression model (FCRM) algorithm several linear subsystems can be divided from the unknown system. By examining the fuzzy relationship, ant colony optimization (ACO) algorithm and fuzzy c-regression model (FCRM) algorithm are adopted to find the fuzzy relationship between data points and linear subsystems, and construct the initial value of the fuzzy rule parameters. Moreover, the weight recursive least squares (WLRS) method is utilized to obtain the initial premise variables of each fuzzy rule for each linear subsystem and establish the T-S fuzzy model. Lastly, some examples are given to illustrate that our modeling method can provide the better approximation results than some studies.

並列關鍵字

T-S Fuzzy model ； Fuzzy Clustering ； Ant Swarm Optimization ； Parameter Estimation

參考文獻

[1] L. A. Zadeh, “Fuzzy sets,” Information and Control, vol. 8, 1965, pp. 335-353.

[2] H. J. Zimmermann, Fuzzy set theory and its applications, Springer Science, 1991

[3] T. Takagi and M. Sugeno, “Fuzzy identification of systems and its applications to modeling and control,” IEEE Transactions on Fuzzy Systems, vol. 15, no. 1, 1985, pp. 116-132.

[4] J. M. Dias and A. Dourado, “A self-organizing fuzzy controller with a fixed maximum number of rules and an adaptive similarity factor,” Fuzzy Sets and Systems, vol. 103, no. 1, 1999, pp. 27-48

[5] J. S. R. Jang, “Self-learning fuzzy controllers based on temporal back propagation,” IEEE Trans. Neural Networks, vol. 3, no. 5, 1992, pp. 714-723.