Polyhedral Gauß-Seidel Converges
Please always quote using this URN:urn:nbn:de:0296-matheon-6879
- We prove global convergence of an inexact polyhedral Gau\ss--Seidel method for the minimization of strictly convex functionals that are continuously differentiable on each polyhedron of a polyhedral decomposition of their domains of definition. While being known to be very slow by themselves, such methods are a cornerstone for fast, globally convergent multigrid methods. Our result generalizes the proof of Kornhuber and Krause [2006] for differentiable functionals on the Gibbs simplex. Example applications are given that require the generality of our approach.
Author: | Carsten Gräser, Oliver Sander |
---|---|
URN: | urn:nbn:de:0296-matheon-6879 |
Referee: | Christof Schütte |
Document Type: | Preprint, Research Center Matheon |
Language: | English |
Date of first Publication: | 2010/04/14 |
Release Date: | 2010/03/22 |
Tag: | |
Institute: | Freie Universität Berlin |
Weierstraß-Institut für Angewandte Analysis und Stochastik (WIAS) | |
Zuse Institute Berlin (ZIB) | |
Preprint Number: | 696 |