Share Email Print

Proceedings Paper

Design of a gradient-index beam shaping system via a genetic algorithm optimization method
Author(s): Neal C. Evans; David L. Shealy
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Geometrical optics - the laws of reflection and refraction, ray tracing, conservation of energy within a bundle of rays, and the condition of constant optical path length - provides a foundation for design of laser beam shaping systems. This paper explores the use of machine learning techniques, concentrating on genetic algorithms, to design laser beam shaping systems using geometrical optics. Specifically, a three-element GRIN laser beam shaping system has been designed to expand and transform a Gaussian input beam profile into one with a uniform irradiance profile. Solution to this problem involves the constrained optimization of a merit function involving a mix of discrete and continuous parameters. The merit function involves terms that measure the deviation of the output beam diameter, divergence, and irradiance from target values. The continuous parameters include the distances between the lens elements, the thickness, and radii of the lens elements. The discrete parameters include the GRIN glass types from a manufacturer's database, the gradient direction of the GRIN elements (positive or negative), and the actual number of lens elements in the system (one to four).

Paper Details

Date Published: 25 October 2000
PDF: 14 pages
Proc. SPIE 4095, Laser Beam Shaping, (25 October 2000); doi: 10.1117/12.405265
Show Author Affiliations
Neal C. Evans, Univ. of Alabama/Birmingham (United States)
David L. Shealy, Univ. of Alabama/Birmingham (United States)

Published in SPIE Proceedings Vol. 4095:
Laser Beam Shaping
Fred M. Dickey; Scott C. Holswade, Editor(s)

© SPIE. Terms of Use
Back to Top