Optimal simplex optimization for optical design

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Simplex optimization is a powerful method of finding minima in noisy merit function spaces, such as those for illumination. The standard simplex routine and a modified one developed by the author are applied to three lens design problems found in the literature: singlet, cemented doublet, and triplet. The starting conditions of the size of the simplex and the location in merit function space are investigated. It is found that the modified simplex routine provides better results than the standard one as the solution converges to the optimal solution, which is called the "end game". The standard simplex tends to provide better results than the modified one when operating in the "start game". The simplex results are compared to those from a commercially available lens design code. In most circumstances the commercially available code provides better performance in both iterations to convergence and quality of the result. The results presented herein provide confirmation that the modified simplex algorithm is a viable means of optimization for noisy merit function determination when in the neighborhood of local optima.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
EditorsJ.M. Sasian, R.J. Koshel, P.K. Manhart, R.C. Juergens
Pages205-216
Number of pages12
Volume5524
DOIs
StatePublished - 2004
EventNovel Optical Systems Design and Optimization VII - Denver, CO, United States
Duration: Aug 2 2004Aug 3 2004

Other

OtherNovel Optical Systems Design and Optimization VII
CountryUnited States
CityDenver, CO
Period8/2/048/3/04

Fingerprint

Optical design
function space
lens design
games
optimization
Lenses
iteration
Lighting
illumination

Keywords

  • Figure of merit
  • Illumination design
  • Lens design
  • Optical design
  • Simplex optimization

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Koshel, R. J. (2004). Optimal simplex optimization for optical design. In J. M. Sasian, R. J. Koshel, P. K. Manhart, & R. C. Juergens (Eds.), Proceedings of SPIE - The International Society for Optical Engineering (Vol. 5524, pp. 205-216). [22] https://doi.org/10.1117/12.562916

Optimal simplex optimization for optical design. / Koshel, Richard John.

Proceedings of SPIE - The International Society for Optical Engineering. ed. / J.M. Sasian; R.J. Koshel; P.K. Manhart; R.C. Juergens. Vol. 5524 2004. p. 205-216 22.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Koshel, RJ 2004, Optimal simplex optimization for optical design. in JM Sasian, RJ Koshel, PK Manhart & RC Juergens (eds), Proceedings of SPIE - The International Society for Optical Engineering. vol. 5524, 22, pp. 205-216, Novel Optical Systems Design and Optimization VII, Denver, CO, United States, 8/2/04. https://doi.org/10.1117/12.562916
Koshel RJ. Optimal simplex optimization for optical design. In Sasian JM, Koshel RJ, Manhart PK, Juergens RC, editors, Proceedings of SPIE - The International Society for Optical Engineering. Vol. 5524. 2004. p. 205-216. 22 https://doi.org/10.1117/12.562916
Koshel, Richard John. / Optimal simplex optimization for optical design. Proceedings of SPIE - The International Society for Optical Engineering. editor / J.M. Sasian ; R.J. Koshel ; P.K. Manhart ; R.C. Juergens. Vol. 5524 2004. pp. 205-216
@inproceedings{61b6384d29474f0680fd95597227a6ce,
title = "Optimal simplex optimization for optical design",
abstract = "Simplex optimization is a powerful method of finding minima in noisy merit function spaces, such as those for illumination. The standard simplex routine and a modified one developed by the author are applied to three lens design problems found in the literature: singlet, cemented doublet, and triplet. The starting conditions of the size of the simplex and the location in merit function space are investigated. It is found that the modified simplex routine provides better results than the standard one as the solution converges to the optimal solution, which is called the {"}end game{"}. The standard simplex tends to provide better results than the modified one when operating in the {"}start game{"}. The simplex results are compared to those from a commercially available lens design code. In most circumstances the commercially available code provides better performance in both iterations to convergence and quality of the result. The results presented herein provide confirmation that the modified simplex algorithm is a viable means of optimization for noisy merit function determination when in the neighborhood of local optima.",
keywords = "Figure of merit, Illumination design, Lens design, Optical design, Simplex optimization",
author = "Koshel, {Richard John}",
year = "2004",
doi = "10.1117/12.562916",
language = "English (US)",
volume = "5524",
pages = "205--216",
editor = "J.M. Sasian and R.J. Koshel and P.K. Manhart and R.C. Juergens",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Optimal simplex optimization for optical design

AU - Koshel, Richard John

PY - 2004

Y1 - 2004

N2 - Simplex optimization is a powerful method of finding minima in noisy merit function spaces, such as those for illumination. The standard simplex routine and a modified one developed by the author are applied to three lens design problems found in the literature: singlet, cemented doublet, and triplet. The starting conditions of the size of the simplex and the location in merit function space are investigated. It is found that the modified simplex routine provides better results than the standard one as the solution converges to the optimal solution, which is called the "end game". The standard simplex tends to provide better results than the modified one when operating in the "start game". The simplex results are compared to those from a commercially available lens design code. In most circumstances the commercially available code provides better performance in both iterations to convergence and quality of the result. The results presented herein provide confirmation that the modified simplex algorithm is a viable means of optimization for noisy merit function determination when in the neighborhood of local optima.

AB - Simplex optimization is a powerful method of finding minima in noisy merit function spaces, such as those for illumination. The standard simplex routine and a modified one developed by the author are applied to three lens design problems found in the literature: singlet, cemented doublet, and triplet. The starting conditions of the size of the simplex and the location in merit function space are investigated. It is found that the modified simplex routine provides better results than the standard one as the solution converges to the optimal solution, which is called the "end game". The standard simplex tends to provide better results than the modified one when operating in the "start game". The simplex results are compared to those from a commercially available lens design code. In most circumstances the commercially available code provides better performance in both iterations to convergence and quality of the result. The results presented herein provide confirmation that the modified simplex algorithm is a viable means of optimization for noisy merit function determination when in the neighborhood of local optima.

KW - Figure of merit

KW - Illumination design

KW - Lens design

KW - Optical design

KW - Simplex optimization

UR - http://www.scopus.com/inward/record.url?scp=13444254168&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=13444254168&partnerID=8YFLogxK

U2 - 10.1117/12.562916

DO - 10.1117/12.562916

M3 - Conference contribution

VL - 5524

SP - 205

EP - 216

BT - Proceedings of SPIE - The International Society for Optical Engineering

A2 - Sasian, J.M.

A2 - Koshel, R.J.

A2 - Manhart, P.K.

A2 - Juergens, R.C.

ER -