A new justification of the Jacobi–Davidson method for large eigenproblems

First published in
Preprints des Institutes für Mathematik 
Preprints des Institutes für Mathematik;Bericht 99
Number in series
99
Citation
Preprint. Published in: Linear Algebra and its ApplicationsVolume 424, Issues 2–3, 15 July 2007, Pages 448-455
Publisher DOI
10.1016/j.laa.2007.02.013
Scopus ID
2-s2.0-34248554446
The Jacobi–Davidson method is known to converge at least quadratically if the correction equation is solved exactly, and it is common experience that the fast convergence is maintained if the correction equation is solved only approximately. In this note we derive the Jacobi–Davidson method in a way that explains this robust behavior.
Schlagworte
large eigenvalue problem
iterative projection method
Jacobi–Davidson method
inexact Krylov subspace methods
DDC Class
510: Mathematik