用于光束整形的衍射光学元件设计的混合算法

提出了一种用于衍射光学元件优化设计的混合遗传迭代爬山算法,该算法将迭代量化傅里叶变换算法融入到遗传算法中,然后在整体遗传算法结束后,对找到的当前最优解再用爬山法进行局部寻优,从而得到最优的衍射光学元件表面相位分布.用该混合方法设计了衍射光学元件,可以将入射的高斯光束整形成方形的均匀光斑.模拟结果表明:该混合算法具有收敛速度快、设计准确度高等优点.相比于其它设计方法,本文提出的方法能较好地改善整形效果,特别适用于光束整形的衍射元件设计.

Design Diffractive Optical Elements for Beam Shaping with Hybrid Algorithm

A novel algorithm,which combines a genetic algorithm with an iterative quantization fourier transform algorithm and hill-climbing algorithm, is presented to design diffractive optical elements for beam shaping.This hybrid algorithm firstly takes the genetic algorithm which combines the iterative soft quantization fourier transform algorithm to obtain a solution.Secondly,the hill-climbing algorithm is applied to search the local best solution around the solution derived firstly.It can acquire good efficiency and uniformity with only a small quantity of iterative number. Simulation result shows that the hybrid algorithm is effective and robust.Compared with other algorithms,this hybrid algorithm can greatly improve the effect of the beam shaping,especially the place where the uniformity is important.