Assessing a search direction within a truncated-newton method

  • Authors:

  • Stephen G. Nash;Ariela Sofer

  • Affiliations:

  • Department of Operations Research and Applied Statistics, George Mason University, Fairfax, VA 22030, USA;Department of Operations Research and Applied Statistics, George Mason University, Fairfax, VA 22030, USA

  • Venue:
  • Operations Research Letters
  • Year:
  • 1990

Quantified Score

Hi-index 0.00

Visualization

Abstract

Truncated-Newton methods for nonlinear optimization compute a search direction by approximately solving the Newton equations, typically via the conjugate-gradient algorithm. The search direction is usually assessed using the norm of the residual. This note shows that the norm of the residual can be an arbitrarily poor predictor of a good search direction.