A Comparison of the PLU and QR Methods for Determining Eigenvalues of Real Hessenberg Matrices.

Abstract

The LU algorithm with interchanges (herein after PLU) should be substantially faster than the QR, based on an order n squared count of floating point arithmetic operations, and is essentially stable. If this comparison held in actual executions of these two algorithms, and if the PLU provided convergence on most matrices, it would be a strong case for using the PLU instead of QR. In actual timings of the algorithms on test matrices, the PLU algorithm did not yield as substantial a savings as was predicted by this traditional operation count. This anomaly is considered. (Modified author abstract)

Document Details

Document Type
Technical Report
Publication Date
May 01, 1974
Accession Number
AD0785244

Entities

People

  • J. T. Panttaja

Organizations

  • University of California, Berkeley

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Arithmetic
  • Convergence
  • Eigenvalues
  • Floating Point Operations
  • Mathematics

Fields of Study

  • Engineering

Readers

  • Linear Algebra
  • Systems Analysis and Design