Alternating-Direction Incomplete Factorizations.

Abstract

To solve the system of linear equations Aw = r that arises from the discretization of a two-dimensional self-adjoint elliptic differential equation, iterative methods employing easily computed incomplete factorization, LU = A+B, are frequently used. Dupont, Kendall, and Rachford showed that, for the DKR factorization, the number of iterations (arithmetic operations) required to reduce the A-norm of the error by a factor of epsilon is O(h to the minus 1/2 power log 1 epsilon) (O(h to the minus 2 and 1/2 power log 1 epsilon)), where h is the stepsize used in the discretization. We present some error estimates which suggest that, if a pair of Alternating-Direction DKR Factorizations are used, then the number of iterations (arithmetic operations) may be decreased to O(h to the minus 1/3 power log 1 epsilon) (O(h to the minus 2 and 1/3 power log 1 epsilon)). Numerical results supporting this estimate are included. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 19, 1981
Accession Number
ADA115430

Entities

People

  • Benren Zhu
  • Kenneth R. Jackson
  • Tony F. Chan

Organizations

  • Yale University

Tags

Communities of Interest

  • C4I

DTIC Thesaurus Topics

  • Arithmetic
  • Chebyshev Polynomials
  • Coefficients
  • Computer Science
  • Convergence
  • Differential Equations
  • Eigenvalues
  • Equations
  • Errors
  • Inequalities
  • Intervals
  • Iterations
  • Numbers
  • Polynomials
  • Square Roots
  • Theorems
  • Two Dimensional

Fields of Study

  • Mathematics

Readers

  • Analytical Mechanics
  • Approximation Theory.
  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)