Implicit steepest descent algorithm for optimization with orthogonality constraints

Research output: Contribution to journalArticlepeer-review

Abstract

Optimization problems with orthogonality constraints appear widely in applications from science and engineering. We address these types of problems from a numerical approach. Our new framework combines the steepest gradient descent, using implicit information, with a projection operator in order to construct a feasible sequence of points. In addition, we adopt an adaptive Barzilai–Borwein steplength mixed with a globalization technique in order to speed-up the convergence of our procedure. The global convergence, and some theoretical related to our algorithm are proved. The effectiveness of our proposed algorithm is demonstrated on a variety of problems including Rayleigh quotient maximization, heterogeneous quadratics minimization, weighted orthogonal procrustes problems and total energy minimization. Numerical results show that the new procedure can outperform some state of the art solvers on some practically problems.

Original languageEnglish
Pages (from-to)1773-1797
Number of pages25
JournalOptimization Letters
Volume16
Issue number6
DOIs
StatePublished - Jul 2022
Externally publishedYes

Keywords

  • Eigenvalue problem
  • Gradient type methods
  • Orthogonality constrained optimization
  • Riemannian optimization
  • Total energy minimization

Fingerprint

Dive into the research topics of 'Implicit steepest descent algorithm for optimization with orthogonality constraints'. Together they form a unique fingerprint.

Cite this