论文部分内容阅读
针对遗传算法(GA)优化超多参量光学系统时鲁棒性较差的问题,在混入逃逸函数实数编码GA(MERCGA)的基础上,进一步结合参量归一化和自适应变异概率的措施,提出了自适应归一化实数编码GA(ANRCGA)。用ANRCGA对鱼眼镜头光学系统案例进行优化设计,并应用评价函数和Zemax光线追迹方法对MERCGA和ANRCGA的优化结果作比较。结果表明,应用本文的ANRCGA比引自专利的参考设计及MERCGA优化得到光学系统的成像质量有明显提高,算法的鲁棒性和计算效率也到了改善。
Aiming at the poor robustness of genetic algorithm (GA) in optimizing superparametric optical system, based on MERCGA (Escalation Code GA) and further measures of parameter normalization and adaptive mutation probability The adaptive normalized real coding GA (ANRCGA). The case of fisheye lens optical system was optimized by ANRCGA, and the result of MERCGA and ANRCGA was compared with the evaluation function and Zemax ray tracing method. The results show that the proposed method improves the imaging quality of the optical system using ANRCGA than the reference design and the MERCGA optimization. The robustness and computational efficiency of the proposed method are also improved.