Online Black-Box Algorithm Portfolios for Continuous Optimization

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F14%3A00220064" target="_blank" >RIV/68407700:21230/14:00220064 - isvavai.cz</a>

Výsledek na webu

<a href="http://pasky.or.cz/sci/cocopf-opt13/optpf-extended.pdf" target="_blank" >http://pasky.or.cz/sci/cocopf-opt13/optpf-extended.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Online Black-Box Algorithm Portfolios for Continuous Optimization

Popis výsledku v původním jazyce

In black-box function optimization, we can choose from a wide variety of heuristic algorithms that are suited to different functions and computation budgets. Given a particular function to be optimized, the problem we consider in this paper is how to select the appropriate algorithm. In general, this problem is studied in the field of algorithm portfolios; we treat the algorithms as black boxes themselves and consider online selection (without learning mapping from problem features to best algorithms apriori and dynamically switching between algorithms during the optimization run). We study some approaches to algorithm selection and present two original selection strategies based on the UCB1 multi-armed bandit policy applied to unbounded rewards. We benchmark our strategies on the BBOB workshop reference functions and demonstrate that algorithm portfolios are beneficial in practice even with some fairly simple strategies, though choosing a good strategy is important.

Název v anglickém jazyce

Online Black-Box Algorithm Portfolios for Continuous Optimization

Popis výsledku anglicky

In black-box function optimization, we can choose from a wide variety of heuristic algorithms that are suited to different functions and computation budgets. Given a particular function to be optimized, the problem we consider in this paper is how to select the appropriate algorithm. In general, this problem is studied in the field of algorithm portfolios; we treat the algorithms as black boxes themselves and consider online selection (without learning mapping from problem features to best algorithms apriori and dynamically switching between algorithms during the optimization run). We study some approaches to algorithm selection and present two original selection strategies based on the UCB1 multi-armed bandit policy applied to unbounded rewards. We benchmark our strategies on the BBOB workshop reference functions and demonstrate that algorithm portfolios are beneficial in practice even with some fairly simple strategies, though choosing a good strategy is important.

Druh

D - Stať ve sborníku

CEP obor

BB - Aplikovaná statistika, operační výzkum

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2014

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Parallel Problem Solving from Nature - PPSN XIII

ISBN

978-3-319-10762-2

ISSN

0302-9743

e-ISSN

—

Počet stran výsledku

10

Strana od-do

40-49

Název nakladatele

Springer

Místo vydání

Heidelberg

Místo konání akce

Ljubljana

Datum konání akce

13. 9. 2014

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—