“巨”型负磁导率材料的优化设计

Optimal design of negative permeability material with "巨"

针对单回路"巨"型负磁导率材料结构和设计要求,基于神经网络和遗传算法理论进行了结构优化设计.通过正交实验法选出了负磁导率材料结构参数样本;利用HFSS得到的仿真实验样本对BP神经网络进行训练,建立了结构参数和结构性能之间的映射关系;根据设计要求定出相应的适应度函数表达式,经过遗传进化得到了优化的负磁导率材料结构参数.

In this paper,based on a research on the integrating of genetic algorithms（GA） and neural network（NN）,and using them in the optimization of single-circuit negative prAmeability material with a shape of Chinese character ＂ju＂,we picked some samples by orthogonal experiment method and calculated the performance parameter with high frequency structure simulator（HFSS） and output the results,and then established the BP neural network model using the former results by HFSS to map the relation between the structure and the performance.After that,we used the GA to search the one satisfying our requirement,by adopting different fitness functions.