Гибридный метод обучения нейронных сетей для классификации каталитической стадии процесса Клауса

Presented a hybrid technique of neural networks as applied to neural network control systems. Technique includes the classic gradient methods, as well as a modified genetic algorithm for adjusting weights. Considers the application of this technique to train the model neuroregulator, for classification of catalytic stage Claus process. The conclusions about the rate of convergence of the proposed methodology.