Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Conformal mapping and convergence of Krylov iterations

Conference ·
OSTI ID:223848
;  [1]
  1. Cornell Univ., Ithaca, NY (United States)

Connections between conformal mapping and matrix iterations have been known for many years. The idea underlying these connections is as follows. Suppose the spectrum of a matrix or operator A is contained in a Jordan region E in the complex plane with 0 not an element of E. Let {phi}(z) denote a conformal map of the exterior of E onto the exterior of the unit disk, with {phi}{infinity} = {infinity}. Then 1/{vert_bar}{phi}(0){vert_bar} is an upper bound for the optimal asymptotic convergence factor of any Krylov subspace iteration. This idea can be made precise in various ways, depending on the matrix iterations, on whether A is finite or infinite dimensional, and on what bounds are assumed on the non-normality of A. This paper explores these connections for a variety of matrix examples, making use of a new MATLAB Schwarz-Christoffel Mapping Toolbox developed by the first author. Unlike the earlier Fortran Schwarz-Christoffel package SCPACK, the new toolbox computes exterior as well as interior Schwarz-Christoffel maps, making it easy to experiment with spectra that are not necessarily symmetric about an axis.

Research Organization:
Front Range Scientific Computations, Inc., Boulder, CO (United States); USDOE, Washington, DC (United States); National Science Foundation, Washington, DC (United States)
OSTI ID:
223848
Report Number(s):
CONF-9404305--Vol.1; ON: DE96005735
Country of Publication:
United States
Language:
English

Similar Records

Linear non-modal phenomena in numerical analysis and applied mathematics. First annual performance report, February 15, 1994--November 15, 1994
Technical Report · Sun Nov 13 23:00:00 EST 1994 · OSTI ID:10114048

Scharz Preconditioners for Krylov Methods: Theory and Practice
Technical Report · Fri May 10 00:00:00 EDT 2013 · OSTI ID:1079618

Krylov subspace recycling for evolving structures
Journal Article · Wed Jan 12 23:00:00 EST 2022 · Computer Methods in Applied Mechanics and Engineering · OSTI ID:1882856