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...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Applied physics. B, Lasers and optics Jg. 127; H. 9
Hauptverfasser: Kong, Rong, Liu, Peng, Zhou, Xin
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Berlin/Heidelberg Springer Berlin Heidelberg 01.09.2021
Springer Nature B.V
Schlagworte:
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-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung: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.
Bibliographie: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