Shamraev A. Neural networks control of nonlinear dynamic objects on the multilayer perceptron basis

The thesis is devoted to the solution of neural network control problems for nonlinear dynamic objects using neural network strategies on the multilayer perceptron (MLP) basis in prior uncertainty conditions concerning properties of objects and influencing noises properties. In work the problems of construction adaptive and neural network control systems are conducted, and some new control and learning algorithms are realized on the MLP neural network basis. Neural network pruning algorithm is developed. Direct adaptive neural network control procedures are improved. A simulation of different MLP learning algorithms is carried out. The solution process of identification and control tasks for nonlinear dynamic objects that are subjected to disturbances is research. The authenticity of the dissertation outcomes proves out by experimental researches and results of their introduction.