关键词:非线性系统;BP神经网络;多项式模型
Abstract
The nonlinear system itself mechanism is having the complexity, now to solves the non-linear problem method also to have the very big limitation. Therefore we must carry on the further comparison to several nonlinear system modelling method and their simulation effect, induces they respective good and bad points, in order to be better, more nimble being suitable in society each production system. Here analyzed introduced the neural network modelling and the multinomial modelling two nonlinear system modelling's method and the simulation, and have carried on the comparison and the analysis to them. the neural network model has the precision to be high, erroneous small characteristic. But when BP training, the request function's gradient, the operand is big, therefore the program execution is more time-consuming. But the multinomial system modeling has the precision to be low, erroneous major problem. Do not choose the multinomial system modeling as far as possible in the high request's situation. The multinomial system modeling operating speed is quick, the training time short is the multinomial system modeling biggest merit.
Utilizes two nonlinear system modelling method in the radio frequency power amplifier to use under a procedure to carry on the deep level to them the comparison, looked that the output chart analyzes they respective good and bad points.
Keywords: Nonlinear system; BP neural network; Multinomial model
目录