Mathematics > Numerical Analysis
[Submitted on 20 Jan 2008 (this version), latest version 16 Mar 2009 (v3)]
Title:Gradient flow approach to geometric convergence analysis of preconditioned eigensolvers
View PDFAbstract: Preconditioned eigenvalue solvers (eigensolvers) are gaining popularity, but their convergence theory remains sparse and complex. We consider the simplest preconditioned eigensolver--the gradient iterative method with a fixed step size--for symmetric generalized eigenvalue problems, where we use the gradient of the Rayleigh quotient as an optimization direction. We prove a known sharp and simple convergence rate bound using novel geometric ideas. The proof is short and self-contained. We first demonstrate that, for a given initial eigenvector approximation, the next iterative approximation belongs to a cone if we apply all admissible preconditioners. Second, using the classical Temple inequality we easily determine that a subspace spanned by two specific eigenvectors gives the smallest norm of the gradient of the Rayleigh quotient. Then we analyse a corresponding continuation method of a gradient flow of the Rayleigh quotient and show that the slowest convergence of the continuation method is reached when the initial vector belongs to this two-dimensional subspace. Next, we extend this result by integration to our fixed step gradient method to conclude that the point on the cone, which corresponds to the poorest convergence and thus gives the guaranteed convergence rate bound, belongs to the same two-dimensional invariant subspace. This reduces the convergence analysis to a two-dimensional case, where a sharp convergence rate bound is derived.
Submission history
From: Andrew Knyazev [view email][v1] Sun, 20 Jan 2008 18:25:57 UTC (87 KB)
[v2] Mon, 16 Jun 2008 18:47:37 UTC (88 KB)
[v3] Mon, 16 Mar 2009 16:56:29 UTC (84 KB)
Current browse context:
math.NA
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.