A family of Barzilai-Borwein steplengths from the viewpoint of scaled total least squares

  • Shiru Li
  • , Tao Zhang
  • , Yong Xia*
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The Barzilai-Borwein (BB) steplengths play great roles in practical gradient methods for solving unconstrained optimization problems. Motivated by the observation that the two well-known BB steplengths correspond to the ordinary and the data least squares, respectively, we introduce a novel family of BB steplengths from the viewpoint of scaled total least squares. Numerical experiments demonstrate that high performance can be received by a carefully-selected BB steplength in the new family.

Original languageEnglish
Pages (from-to)1011-1031
Number of pages21
JournalComputational Optimization and Applications
Volume87
Issue number3
DOIs
StatePublished - Apr 2024

Keywords

  • BB steplength
  • Gradient descent
  • Total least squares
  • Unconstrained optimization

Fingerprint

Dive into the research topics of 'A family of Barzilai-Borwein steplengths from the viewpoint of scaled total least squares'. Together they form a unique fingerprint.

Cite this