A Combination of Fuzzy Theory and Genetic-Neural Network Algorithm

Nowadays, the BP network algorithm has achieved a great success and many nonlinear problems can be solved well. However, standard BP network algorithm has some Shortcomings. Such as local minimum, low convergence and oscillation effects etc. GA has a strong macro-search capability. It has some advantages. Such as simple and universal, robust, parallel computing features, so use it to complete the pre-search can overcome the shortcomings of BP. Fuzzy system is good at express people´s experiential knowledge. It can deal with vague information. It can solve the intelligent questions better. Fuzzy clustering methods have been used widely in pattern recognition. Combine fuzzy systems with genetic-neural Network Algorithm not only make the algorithm more efficient, but also to address the intelligent questions better. It has become a hot research.