Obtaining new local minima in lens design by constructing saddle points

Area of Science:

• Optics and Optical Engineering
• Computational Science
• Applied Mathematics

Background:

• Lens design optimization often encounters challenges with local minima.
• Existing saddle-point methods have limitations in lens design applications.

Purpose of the Study:

• To introduce a novel, systematic, and efficient saddle-point method for lens design.
• To leverage theoretical insights into saddle points for global optimization.

Main Methods:

• Developed a saddle-point method utilizing a-priori knowledge.
• Applied the method to identify new local minima in the lens design landscape.
• Demonstrated the method's ability to refine poor-quality local minima into good-quality ones.

Main Results:

• Established the existence of saddle points closely related to local minima in lens design.
• The developed method successfully obtains new local minima.
• The method is capable of creating both positive and negative lenses, unlike previous approaches.
• Successfully improved a poor-quality local minimum to a good-quality one through successive application.

Conclusions:

• The new saddle-point method offers a systematic and efficient approach to lens design optimization.
• This method expands the capability to generate diverse lens types (positive and negative).
• The approach shows potential applicability to other global optimization problems with similar requirements.