904. Acceleration strategies for gradient methods
Invited abstract in session MC-37: Optimization algorithms and applications 1, stream Nonlinear Optimization.
Monday, 10:30-12:00Room: JUR – Seminar-Raum 43
Authors (first author is the speaker)
| 1. | Anna De Magistris
|
| Department of Mathematics and Physics, University of Campania “Luigi Vanvitelli” |
Abstract
Gradient methods are fundamental to large-scale optimization, yet acceleration remains a key challenge. This work moves beyond standard negative gradients by employing composite search directions. The first strategy executes a minimization process within low-dimensional affine subspaces, specifically targeting convex quadratic problems. The second adapts the “twin” concept from Kaczmarz iterations as a means of acceleration. Numerical experiments demonstrate that both strategies significantly enhance convergence and robustness.
Keywords
- Optimization Models and Methods
- Programming, Quadratic
- Programming, Mathematical
Status: accepted
Back to the list of papers