基于RBFNN和GA优化设计DGS微带*线  

Design and Optimization of DGS Microstrip Based on RBFNN and GA

作　　者：杜道鹏[1] 高卫东[1] 李乐[1] 

机构地区：[1]解放军电子工程学院,安徽合肥230037

出　　处：《雷达科学与技术》2009年第4期316-320,324,共6页Radar Science and Technology

摘　　要：采用神经网络与遗传算法相结合的方法,对一种缺陷接地结构微带线进行优化设计。整个优化设计过程分为两步进行:首先采用基于混合递阶遗传算法优化训练的RBF神经网络对缺陷接地结构微带线进行建模,当神经网络模型训练成功后就实现了对这种缺陷地结构微带线传输系数快速和精确的输出仿真;然后利用该神经网络模型,并结合遗传算法共同优化设计这种缺陷地结构微带线的传输系数和尺寸参数。仿真结果验证了该方法的有效性和准确性。The RBF neural network and genetic algorithms are successfully applied to the design and optimization of a DGS microstrip. The design and optimization procedure can be divided into two steps. Firstly, this RBF neural network based on hybrid hierarchy genetic algorithms is applied to the modeling of the DGS microstrip, and the transmission coefficient of the DGS microstrip can be obtained from the neural network model which is trained successfully. Secondly, according to the design goal, the transmission coefficient and the dimensions of the DGS microstrip can be designed by the trained neural network model and GA. The re suits of simulation demonstrate that the approach presented in this paper is valid and accurate.

关 键 词：缺陷接地结构 RBF神经网络 混合递阶遗传算法 传输系数 

分 类 号：TN817[电子电信—信息与通信工程] TN957