RBF网络与BP网络在非线性电子器件建模中的性能比较

Performance Comparison between the BP and RBF in Modeling Nonlinear Electron Devices

针对BP网络在非线性电子器件建模应用中的不足,尝试了利用RBF网络建模的新方法,即首先用RBF网络逼近器件的非线性特性,然后再将得出的网络结构用Pspice电路描述语言描述来建立非线性电子器件的模型,并对这两种网络在建模时的优缺点进行了比较.

Considering the shortage of the BP neural network in modeling the nonlinear electron devices, a new method based on the RBF network is tried. The model can be obtained by describing the architecture of the neural network in Pspice, which is achieved by training the characteristic curve of the devices. Comparisons of the merit and shortcoming between using these two modeling methods are presented.