The LM Algorithm Research in Brillouin Scattering Spectra Information Based on the Weighted Least Squares

Author

Yu, Wei ; Wan, ShengPeng ; Li, Baoji ; Zhong, Lin ; Hu, Wengang ; Wang, Wei

Author_Institution

Key Lab. of Nondestructive Test (Minist. of Educ.), Nanchang Hangkong Univ., Nanchang, China

fYear

2012

fDate

21-23 May 2012

Firstpage

1

Lastpage

5

Abstract

Compare LM algorithm in reference to the iteration because of slow convergence speed,low convergence precision,poor stability,easy to get bad data influence each other about the Weighted Least Squares. This paper introduces the LM algorithm of based on Weighted Least Squares. The algorithm is applied to Brillouin scattering spectra information processing. The LM algorithm is the improvement algorithm of the Weighted Least Squares.This paper expounds the principle of Weighted Least Squares and LM algorithm . That introduces the LM algorithm of parameters adjustment method and its convergence judgement basis . Finally,the paper gives corresponding simulation results and analysis according to the examples . The LM algorithm has better convergence and stability comparing with the traditional Weighted Least Squares. The LM algorithm has rapid convergence and higher precision and saves time for measurement influence Brillouin scattering parameters.

Keywords

convergence of numerical methods; least squares approximations; stimulated Brillouin scattering; Brillouin scattering spectra information processing; LM algorithm; convergence judgement; convergence precision; convergence speed; convergence stability; parameter adjustment method; rapid convergence; weighted least squares;

fLanguage

English

Publisher

ieee

Conference_Titel

Photonics and Optoelectronics (SOPO), 2012 Symposium on

Conference_Location

Shanghai

ISSN

2156-8464

Print_ISBN

978-1-4577-0909-8

Type

conf

DOI

10.1109/SOPO.2012.6270510

Filename

6270510