遗传算法用于衍射光学元件的优化设计

提出了一种基于遗传算法的衍射光学元件优化设计方法;在衍射光学元件设计中遗传算法运行参数对遗传算法性能有一定的影响:采用较大的群体规模,遗传算法越容易获得最优解;交叉算子越大,遗传算法全局搜索能力越强;选择算子对遗传算法的影响不是太大;如果要进一步提高解的精度,可选取较大的终止代数。数值计算结果表明,用遗传算法优化设计的衍射光学元件,其误差小于 5.2%,衍射效率达到 91.2%。遗传算法很适合衍射光学元件的优化设计。

Genetic algorithms used for optimization design of diffractive optical elements

A diffractive optical elements optimization design method based on genetic algorithm is proposed. For designing diffractive optical elements, the function parameters of genetic algorithm have some influence on the performance of genetic algorithm: when a bigger population scale is selected, genetic algorithm can obtain optimal solution easily; when cross operator is larger the searching capability of genetic algorithm will be more powerful. The influence of selecting operator on genetic algorithm is not important.If solution accuracy is to be further improved, a bigger termination generation can be selected. The numerical calculation results shows that the error of the diffractive optical elements optimally designed by genetic algorithm is less than 5.2% and its diffraction efficiency has attained 91.2%. Genetic algorithm is very suitable for the optimization design of diffractive optical elements.