人工神经网络和遗传算法在微带交指电容器设计中的应用

The Application of Artificial Neural Network and Genetic Algorithms to the Design of Microstrip Interdigital Capacitor

将神经网络技术 (ANN)与遗传算法 (GA)相结合对交指电容器 (IDC)进行了分析与设计。采用多层感知器神经网络 (MLPNN)建立了交指电容器的模型 ,并利用遗传算法的全局搜索能力根据实际需要优化设计交指电容器的结构。模型训练样本的S参数由时域有限差分 (FDTD)方法得到。结果证明该方法具有较高的准确性 。

In this paper, a novel approach combining artificial neural network (ANN) and ge netic algorithms (GA) to analyze and design interdigital capacitor (IDC) is desc ribed. A multilayer perceptron neural network (MLPNN) is applied to describe the model of IDC. According to the engineering requirement, the dimensions of IDC c an be designed by the trained models and GA which has global searching ability. The scattering parameters of the training samples are computed by the finite di fference time domain (FDTD). This design procedure is proved to be time saving and of high accuracy.