论文信息 - A new global optimizing algorithm for fuzzy neural networks

A new global optimizing algorithm for fuzzy neural networks

In this paper, a new global optimizing algorithm that combines the modified quasiNewton method and the improved genetic algorithm is proposed to find the global minimum of the total error function of a fuzzy neural network. A global linear search algorithm based on fuzzy logic and combinative interpolation techniques is developed in the modified quasi-Newton model. It is shown that this algorithm ensures convergence to a global minimum with probability 1 in a compact region of a weight vector space. The results of computer simulations also reveal that this algorithm has a better convergence property and the times of global search are obviously decreased.

[1] Marius Iosifescu,et al. Finite Markov Processes and Their Applications , 1981 .

[2] Günter Rudolph,et al. Convergence analysis of canonical genetic algorithms , 1994, IEEE Trans. Neural Networks.

[3] José Carlos Príncipe,et al. A Simulated Annealing Like Convergence Theory for the Simple Genetic Algorithm , 1991, ICGA.

[4] Yoshiki Uchikawa,et al. On fuzzy modeling using fuzzy neural networks with the back-propagation algorithm , 1992, IEEE Trans. Neural Networks.

[5] Derek A. Linkens,et al. Fuzzified RBF network-based learning control: structure and self-construction , 1993, IEEE International Conference on Neural Networks.

[6] R. Katayama,et al. Self generating radial basis function as neuro-fuzzy model and its application to nonlinear prediction of chaotic time series , 1993, [Proceedings 1993] Second IEEE International Conference on Fuzzy Systems.