Optimal Restart Strategies for Parameter-dependent Optimization Algorithms

Lisa Schönenberger and Hans-Georg Beyer · 2025

Read Optimal Restart Strategies for Parameter-dependent Optimization Algorithms by Lisa Schönenberger and Hans-Georg Beyer (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-10173.txt.

Description

Lisa Schönenberger and Hans-Georg Beyer study restart plans for algorithms that need an unknown threshold, in work from 2025. Success comes only once a parameter passes its hidden best value. Each restart raises the parameter, but higher values also cost more run time.

Strategy types are grouped by how the parameter grows across restarts. A loss score measures wasted compute cost against the best possible plan. Bounds show that not every growth type keeps that loss bounded. Steady multiplicative growth does keep loss bounded for all cases. Inside that family, one growth factor cuts the worst loss most. In the long run that best factor does not depend on the hidden threshold.

Catalog record

CREATORLisa Schönenberger and Hans-Georg Beyer
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-10173

Browse the full library · About the Tower of Babel