基于神经网络的微波电路建模与优化

本文讨论用神经网络对微波电路进行建模、优化。借助电磁声理论计算或基于实际测量，可得到微波电路的输入、输出样本数据，从而可训练神经网络，在兼顾它的推广性能的基础上，对微波电路建模。进一步，通过优化神经网络对应参数，可优化微波电路。文章用ＲＢＦ（Ｒａｄｉａｌ Ｂａｓｉｓ Ｆｕｎｃｔｉｏｎ）神经网络对微带变阻器建模、优化，以此为例，进行了较为详细的阐述。

Modeling and Optimization of Microwave Circuits Based on Neural Networks

The modeling and optimization of microwave circ uits using RBF neural networks are discussed in this paper.According to the elec tromagnetic computation methods or some measuring methods we get a series of inp ut and output circuit data ,which could be used to train the network.While mode ling the circuit the generation ability as well as the accuracy of the network m ust be considered.Further more,by optimizing the network parameters the circuit structure may be optimized also.In this article we elaborate on the modeling an d optimization of the microstrip impedance transformer as an example.