Saved in:
Bibliographic Details
Main Authors: Bagy, A. C., Chbani, Z., Riahi, H.
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.00038
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910393710936064
author Bagy, A. C.
Chbani, Z.
Riahi, H.
author_facet Bagy, A. C.
Chbani, Z.
Riahi, H.
contents Given a proper convex lower semicontinuous function defined on a Hilbert space and whose solution set is supposed nonempty. For attaining a global minimizer when this convex function is continuously differentiable, we approach it by a first-order continuous dynamical system with a time rescaling parameter and a Tikhonov regularization term. We show, along the generated trajectories, fast convergence of values, fast convergence of gradients towards origin and strong convergence towards the minimum norm element of the solution set. These convergence rates now depend on the time rescaling parameter, and thus improve existing results by choosing this parameter appropriately. The obtained results illustrate, via particular cases on the choice of the time rescaling parameter, good performances of the proposed continuous method and the wide range of applications they can address. Numerical illustrations for continuous example are provided to confirm the theoretical results.
format Preprint
id arxiv_https___arxiv_org_abs_2404_00038
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Strong convergence towards the minimum norm solution via temporal scaling and Tikhonov approximation of a first-order dynamical system
Bagy, A. C.
Chbani, Z.
Riahi, H.
Optimization and Control
Dynamical Systems
65K15, 65K15, 90C30
Given a proper convex lower semicontinuous function defined on a Hilbert space and whose solution set is supposed nonempty. For attaining a global minimizer when this convex function is continuously differentiable, we approach it by a first-order continuous dynamical system with a time rescaling parameter and a Tikhonov regularization term. We show, along the generated trajectories, fast convergence of values, fast convergence of gradients towards origin and strong convergence towards the minimum norm element of the solution set. These convergence rates now depend on the time rescaling parameter, and thus improve existing results by choosing this parameter appropriately. The obtained results illustrate, via particular cases on the choice of the time rescaling parameter, good performances of the proposed continuous method and the wide range of applications they can address. Numerical illustrations for continuous example are provided to confirm the theoretical results.
title Strong convergence towards the minimum norm solution via temporal scaling and Tikhonov approximation of a first-order dynamical system
topic Optimization and Control
Dynamical Systems
65K15, 65K15, 90C30
url https://arxiv.org/abs/2404.00038