【作者】 邵桢；

【导师】 任涛；

【作者基本信息】 长春理工大学 ， 计算机应用技术， 2004， 硕士

【摘要】 智能优化算法通过模拟某些自然现象或过程的机理而得到了广泛地发展，本文针对光学CAD系统中目前正研究的热门课题提出了新的观点，重点研究了智能优化算法在光学系统设计中的应用。 首先，由于现有的光学系统结构设计的优化算法都是在选定了初始光学结构的基础上，只对光学结构参数进行优化选择，在优化过程中不能够自动改变光学面数，导致光学系统设计的智能化程度还不是很高，采用遗传算法，实现了光学面的自动增减； 其次，区别于传统的二元光学设计方法，为避免其量化台阶数过大导致的成本及加工复杂度的增加，提出了运用混合优化策略遗传-模拟退火算法直接设计二元光学元件的新的计算机方法。 最后，通过构造数学模型和编程实现了算法，由优化的效果和效率分析来证实智能优化算法是提高光学系统设计智能化程度的有效工具。更多还原

【Abstract】 Intelligent optimization algorithms get to broad development by imitating or announcing to some natural phenomenas or processes.This paper brings up the new standpoint aiming at the optics CAD system’s popular lesson,and makes a particular study of Intelligent optimization algorithms’ application in the optical system design.First,the current optical system construction’s optimization algorithms are based on the just optical system’s preliminary construction which designers have selected. Optical system program can not automatically change the optical elements’nember.So that optical system design’s intelligentized degree is not high.Now this paper applies the genetic algorithms to lens design and shows that the genetic algorithms can effectively work in automatically changing the optical elements’nember.Next, this pape present a new computer design method that applies the genetic-simulated annealing algorithms to directly design the binary optical element. This method avoids the high cost and facture’ complexity and is different of the conventional design method.At last, we implement these algorithms by structuring the mathematic model and programming. We show that the effectiveness of Intelligent optimization algorithms in increasing the optical system design’s intelligentized degree by analyzing the algorithms’ effect and efficiency.更多还原