Multipass cell design with the random walk and gradient descent optimization algorithms

An automated approach is presented for optimizing the multipass cell (MPC) design with dense patterns in this paper. First, a strategy based on the random walk (RW) algorithm is implemented for global exploration to determine the parameters of the target MPC configuration and accelerate the design p...

Full description

Saved in:
Bibliographic Details
Published in:Applied physics. B, Lasers and optics Vol. 127; no. 9
Main Authors: Kong, Rong, Liu, Peng, Zhou, Xin
Format: Journal Article
Language:English
Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.09.2021
Springer Nature B.V
Subjects:
Algorithms
Applied physics
Clustering
Configuration management
Design optimization
Engineering
Identification methods
Lasers
Optical Devices
Optics
Photonics
Physical Chemistry
Physics
Physics and Astronomy
Quantum Optics
Random walk
Regular Paper
ISSN:0946-2171, 1432-0649
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:An automated approach is presented for optimizing the multipass cell (MPC) design with dense patterns in this paper. First, a strategy based on the random walk (RW) algorithm is implemented for global exploration to determine the parameters of the target MPC configuration and accelerate the design process. Second, the gradient descent (GD) algorithm is performed for local exploitation to optimize the re-entrant condition in a fast and automatic way. In addition, we apply the clustering method to identify the desired spot patterns with specific properties automatically. Finally, the proposed algorithms are tested in the optimization of two types of densely patterned MPC under the re-entrant condition. The results presented in this paper clearly show that the proposed approach is effective and efficient in optimizing the MPC design automatically and can be further utilized in more complex optical configurations.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0946-2171
1432-0649
DOI:10.1007/s00340-021-07679-6