Essential Convergence Rates of Continuous-Time Models for Optimization Methods

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ushiyama, Kansei, Sato, Shun, Matsuo, Takayasu
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915698054266880
author Ushiyama, Kansei
Sato, Shun
Matsuo, Takayasu
author_facet Ushiyama, Kansei
Sato, Shun
Matsuo, Takayasu
contents Designing and analyzing optimization methods via continuous-time models expressed as ordinary differential equations (ODEs) is a promising approach for its intuitiveness and simplicity. A key concern, however, is that the convergence rates of such models can be arbitrarily modified by time rescaling, rendering the task of seeking ODEs with ``fast'' convergence meaningless. To eliminate this ambiguity of the rates, we introduce the notion of the essential convergence rate. We justify this notion by proving that, under appropriate assumptions on discretization, no method obtained by discretizing an ODE can achieve a faster rate than its essential convergence rate.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23317
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Essential Convergence Rates of Continuous-Time Models for Optimization Methods
Ushiyama, Kansei
Sato, Shun
Matsuo, Takayasu
Optimization and Control
65K05, 90C25, 65L20
Designing and analyzing optimization methods via continuous-time models expressed as ordinary differential equations (ODEs) is a promising approach for its intuitiveness and simplicity. A key concern, however, is that the convergence rates of such models can be arbitrarily modified by time rescaling, rendering the task of seeking ODEs with ``fast'' convergence meaningless. To eliminate this ambiguity of the rates, we introduce the notion of the essential convergence rate. We justify this notion by proving that, under appropriate assumptions on discretization, no method obtained by discretizing an ODE can achieve a faster rate than its essential convergence rate.
title Essential Convergence Rates of Continuous-Time Models for Optimization Methods
topic Optimization and Control
65K05, 90C25, 65L20
url https://arxiv.org/abs/2512.23317