Neural Network Control Under The Multi-step Predictive Index Function And Application

First, the study of neural networks has never been important today .Then,the scheme of using foreword NN to identify the plant is presented. The idea of predictive control is combined with that of NN inverse control. In the thesis, the neuro-control schemes for complex system are mainly studied, and the faults of predictive control are analyzed. So the idea of predictive control is combined with that of NN inverse control is presented. NN inverse control under the multi-step predictive index function is used in temperature of Air-conditioning control system. Using this method in the thesis, the system has better response performance. To the approximated error of the NN identifying apparatus, the method has stronger robustness. Under new method, the outputs of the plant have not complementary error.