Two novel gradient methods with optimal step sizes

Harry Oviedo, Oscar Dalmau, Rafael Herrera

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

In this work we introduce two new Barzilai and Borwein-like steps sizes for the clas- sical gradient method for strictly convex quadratic optimization problems. The proposed step sizes employ second-order information in order to obtain faster gradient-type meth- ods. Both step sizes are derived from two unconstrained optimization models that involve approximate information of the Hessian of the objective function. A convergence anal- ysis of the proposed algorithm is provided. Some numerical experiments are performed in order to compare the efficiency and effectiveness of the proposed methods with similar methods in the literature. Experimentally, it is observed that our proposals accelerate the gradient method at nearly no extra computational cost, which makes our proposal a good alternative to solve large-scale problems.

Original languageEnglish
Pages (from-to)375-391
Number of pages17
JournalJournal of Computational Mathematics
Volume39
Issue number3
DOIs
StatePublished - Apr 2021
Externally publishedYes

Keywords

  • Convex quadratic optimization
  • Gradient methods
  • Hessian spectral properties
  • Steplength selection

Fingerprint

Dive into the research topics of 'Two novel gradient methods with optimal step sizes'. Together they form a unique fingerprint.

Cite this