A Generalized Conjugate Gradient Method for Non-Symmetric Systems of Linear Equations.

Abstract

A new iterative method is presented for solving non-symmetric linear systems of equations. The method requires that the symmetric of the matrix of the linear system be positive definite, and the method is efficient only if the symmetric part is easily invertible. The method is modeled on the conjugate gradient method for symmetric positive definite systems and has the finite termination property. The results from several numerical experiments are presented and compared with a similar method proposed by Concus, Golub, and Widlund.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 1981
Accession Number
ADA110311

Entities

People

  • John C. Strikwerda

Organizations

  • University of Wisconsin–Madison

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Classification
  • Computer Science
  • Computers
  • Contracts
  • Convergence
  • Differential Equations
  • Equations
  • Linear Systems
  • Mathematics
  • North Carolina
  • Numerical Analysis
  • Partial Differential Equations
  • Steepest Descent Method
  • Theorems
  • United States
  • Wisconsin

Fields of Study

  • Mathematics

Readers

  • Finite Element Method (FEM) for solving Partial Differential Equations (PDEs)
  • Forest Ecology
  • Operations Research