利用改良式差分進化及文化演算法於遞迴式函數類神經模糊網路之設計與應用

Design of a Recurrent Functional Neural Fuzzy Network Using Modified Differential Evolution and Cultural Algorithm

翁嘉俊(Chia-Chun Weng)

指導教授 ： 王德譽 ； 林正堅

朝陽科技大學/資訊學院/資訊工程系/碩士(2009年)

摘要

在本文中，我們提出了遞迴式函數鏈結類神經模糊網路結合改良式差分進化演算法和文化改良式差分進化演算法來解決預測和控制的問題。提出的遞迴式函數鏈結類神經模糊網路在歸屬函數層中加入了迴饋訊號的連結，以用來解決時間性的問題。除此之外，兩個有效的學習演算法，稱為改良式差分進化演算法和文化改良式差分進化演算法作為遞迴式函數鏈結類神經模糊網路的參數調整。在差分進化演算法中為了能夠增加突變的差異性，我們從母體中選擇了四條個體進行突變，有助於搜尋解能力的提升。而在文化改良式差分進化演算法中，它是結合了文化演算法和改良式差分進化演算法，在進化的過程中，利用文化演算法中的信仰空間來粹取並使用這些資訊，能夠有助於效能上的提升。模擬結果顯示，我們所提出的方法在收斂速度以及均方根值誤差都要比其它的方法擁有更好的效能。

關鍵字

類神經模糊網路 ； 函數鏈結神經網路 ； 遞迴式網路 ； 差異進化演算法 ； 文化演算法 ； 預測 ； 控制

並列摘要

In this thesis, a recurrent functional neural fuzzy network (RFNFN) with modified differential evolution (MDE) and cultural-based modified differential evolution (CMDE) is proposed for solving prediction and control problems. The proposed RFNFN model has feedback connections added in the membership function layer that can solve temporal problems. Moreover, two efficient learning algorithms, called modified differential evolution (MDE) and cultural-based modified differential evolution (CMDE) for tuning parameters of the RFNFN. In order to increase the diversity of mutations in differential evolution, we randomly choose four individual from the population to mutation. The solution can search capacity more efficiently. In cultural based modified differential evolution (CMDE) combined the cultural algorithm and modified differential evolution. It during the evolutionary process, the belief spaces extraction and use of the information is very effective in increase the performance. Simulation results show that the converging speed and root mean square error (RMSE) of the proposed method has a better performance than those of other methods.

並列關鍵字

cultural algorithm ； prediction ； control ； neural fuzzy networks ； functional link neural network ； recurrent networks ； differential evolutionary algorithm

參考文獻

[4] C. J. Lin, C. H. Chen, and C. Y. Lee, “Efficient immune-based particle swarm optimization learning for neuro-fuzzy networks design,” Journal of Information Science and Engineering, Sep. 2008, Vol. 24, No. 5, pp. 1505-1520.

[1] C. H. Chen, C. J. Lin, and C. T. Lin, “A functional-link-based neuro-fuzzy network for nonlinear system control,” IEEE Trans. on Fuzzy Systems, Oct. 2008, Vol. 16, No. 5, pp. 1362-1378.

[2] J. S. R. Jang, “ANFIS: Adaptive-network-based fuzzy inference system,” IEEE Trans. on Syst., Man, and Cybern., 1993, Vol. 23, pp. 665-685.

[3] T. Takagi and M. Sugeno, “Fuzzy identification of systems and its applications to modeling and control,” IEEE Trans. on Syst., Man, and Cybern., 1985, Vol. 15, pp. 116-132.

[5] H. Takagi, N. Suzuki, T. Koda, and Y. Kojima, “Neural networks designed on approximate reasoning architecture and their application,” IEEE Trans. On Neural Networks, Sep. 1992, Vol. 3, pp. 752-759.