遗传算法在光栅面形优化设计中的应用

衍射效率高、偏振相关损耗低的衍射光栅是集成波分复用／解复用器件的一种核心元件，在光通信中具有很重要的应用价值，而一般面形的光栅很难达到此要求，为此需要在光栅衍射效率计算的基础上，利用遗传算法对光栅面形进行优化。采用二次函数来描述光栅面形，选取一定的原始样本，利用耦合波方法计算光栅的衍射效率；然后采用遗传算法中的各种算子对样本进行变换，计算变换后光栅面形的衍射效率，选择其中接近所确定条件的样本进行保留；再继续下一次变换和选择，直到满足设定的要求。通过优化最后得到衍射效率高、偏辗相关损耗小的光栅面形，这些面形的光栅衍射效率达到90％以上，而相应的偏振相关损耗小于0．1dB，已经达到光通信中的相应要求。在光栅面形设计中采用了遗传算法进行光栅面形优化为有特定要求光栅的制作提供了理论指导，可以节省一定的人力和渤力、

Application of genetic algorithm in the design optimization of grating profile

Diffraction grating of high diffraction efficiency and low polarization dependent loss is a key component of the integrated wavelength division-multiplexing device and is of importance in optical communication. But it is difficult for traditional gratings to meet the special requirement of optical communication. A genetic algorithm was used to optimize the profile of grating based on the theoretical calculation method of grating efficiency. A quadratic function was used to describe the grating profile, some original samples were selected, and diffraction efficiency of the sample gratings wascalculated by using the couple-wave approach. Furthermore a few transformations of coefficient of quadratic tuncnons were made by using different operators of genetic algorithm. The diffraction efficiencies of the gratings described by those transformed quadratic function were calculated, the grating profiles that were close to the special requirement were chosen, and others were rejected. Next the transformations of the new sample gratings were made again ,the diffraction efficiency of the gratings described by those transformed quadratic function were calculated again, and then above process was repeated, it was continued until the request was satisfied. By using the method mentioned before, finally we have achieved some optimizational grating profiles that have high diffraction efficiency and low polarization dependence loss. Diffraction efficiency of the gratings achieves more than 90 percent and polarization dependence loss is less than 0.1 dB. It meets the special requirement of optical communication. It is showed that genetic algorithm was used to optimize the profile of grating based on the theoretical calculation method of grating efficiency, it provides theoretical guidance for grating fabrication, and much manpower and material resource were saved.