Accession Number:

AD0742126

Title:

Linear Convergence of the Conjugate Gradient Method

Descriptive Note:

Corporate Author:

IBM THOMAS J WATSON RESEARCH CENTER YORKTOWN HEIGHTS NY

Personal Author(s):

• Crowder, Harlan
• Wolfe, Philip

Report Date:

1972-05-12

Pagination or Media Count:

8.0

Abstract:

It is shown that the method of conjugate gradients for the minimization of a quadratic function converges no better than linearly if the standard starting and restarting procedures are not used.

Descriptors:

• *MATHEMATICAL PROGRAMMING
• *ITERATIONS
• OPTIMIZATION
• TRANSFORMATIONS(MATHEMATICS)
• CONVERGENCE
• STEEPEST DESCENT METHOD

Subject Categories:

• Numerical Mathematics
• Operations Research

Distribution Statement:

APPROVED FOR PUBLIC RELEASE