基于免疫遗传算法的二元光学元件的位相设计

Design of phase distribution of binary optics elements based on immune genetic algorithm

在激光诱导扩散中,需要利用二元光学元件对激光器输出的高斯光束进行整形,以实现曝光区的温度分布均匀化。为了得到二元光学元件的位相分布,采用免疫遗传对相位分布进行设计。免疫遗传算法中采取变频率的交叉操作、变异操作,克服了遗传算法在局部搜索解空间上效率差的缺点,并使算法跳出局部极大值的能力得到了增强。采取由正向记忆细胞库提取的免疫疫苗对抗体群进行接种,使群体的进化方向得到引导,提高了算法的进化效率;采取由反向记忆细胞库提取的劣化疫苗对抗体群进行反向接种,减少算法的重复运算,极大地抑制了群体退化;采用B、T细胞的作用机制,保持群体在进化过程中的多样性,很大程度上抑制了算法未成熟收敛。运算结果表明,免疫遗传算法较遗传算法具有更高的算法效率和更强的寻优能力。最后考虑到实际加工,对最优解做适当调整得到了更适合于实际加工的二元光学元件的位相分布。

In laser assisted diffusion, to realize the homogenization of the temperature distribution in the small exposed region, the laser beam is transformed through Binary Optics Elements （BOE）. In order to determine the phase distribution of BOE, an immune genetic algorithm is adopted to design BOE. Various frequency crossover operation and mutation operation have been applied into this algorithm to improve searching efficiency in local solution space and enhance the ability to escape local maximum. The immune vaccine extracted from orthodox memory cells bank has been injected into antibody population, which guides the evolution direction of antibody population and accelerate evolution. Meanwhile, to decrease reiteration calculation and restrain degeneration, the inverse vaccine extracted from inverse memory cells bank has been inversely injected into antibody population. The operation mechanism between B/T cells has been taken to maintain the diversity of population during the evolution process, and to restrain the premature convergence greatly. The experiment results show that this algorithm is more effective and has stronger ability to get the optimum solution compared with genetic algorithm. The optimum solution has been adjusted to get the expected phase distribution of BOE, which can be easily made.