Home          
 
Short Presentations Accepted
 
 
     LSQR Method Acceleration Using a Simulated LU Factorization
     Presenter: Leobardo Valera
     Co-Authors: Brigida Molina
Abstract

For a overdetermined linear system of equations Ax=b where A is a full rank matrix with m rows and n columns, b is a vector with m components, b is not in the range of A. The proposal of this work is to extract n+1 linear independent rows of the augmented matrix [A b] using a simulated LU factorization. Then apply the LSQR method to the resulting n+1 equation. When applying the LSQR method to the reduced system the same solution is obtained than the original system but in less time and iteration number.

 

SDSU: Computational Science and Engineering Gateway to Latin America

Last updated: June 2, 2010 10:52 AM