Randomization of Low-discrepancy Sampling Designs by Cranley-Patterson Rotation

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27240%2F23%3A10253619" target="_blank" >RIV/61989100:27240/23:10253619 - isvavai.cz</a>

Výsledek na webu

<a href="https://dl.acm.org/doi/10.1145/3628454.3631564" target="_blank" >https://dl.acm.org/doi/10.1145/3628454.3631564</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1145/3628454.3631564" target="_blank" >10.1145/3628454.3631564</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Randomization of Low-discrepancy Sampling Designs by Cranley-Patterson Rotation

Popis výsledku v původním jazyce

Complex problems are often addressed by methods from the domain of computational intelligence, including metaheuristic algorithms. Different metaheuristics have different abilities to solve specific types of problems and the selection of suitable methods has a large impact on the ability to find good problem solutions. Problem characterization became an important step in the application of intelligent methods to practical problems. A popular approach to problem characterization is the exploratory landscape analysis. It consists of a sequence of operations that approximate and describe the hypersurfaces formed by characteristic problem properties from a limited sample of solutions. Exploratory landscape analysis uses a particular strategy to select just a small subset of problem solutions for which are the characteristic properties evaluated and high-level landscape features computed. Low-discrepancy sequences have been recently used to design a family of sampling strategies. They have useful space-filling properties but their effective and efficient randomization might represent an issue. In this work, we study the Cranley-Patterson rotation, a lightweight randomization strategy for low-discrepancy sequences, compare it with other randomization methods, and observe the effect its use has on the randomization of sets of sampling points in the context of exploratory landscape analysis.

Název v anglickém jazyce

Randomization of Low-discrepancy Sampling Designs by Cranley-Patterson Rotation

Popis výsledku anglicky

Complex problems are often addressed by methods from the domain of computational intelligence, including metaheuristic algorithms. Different metaheuristics have different abilities to solve specific types of problems and the selection of suitable methods has a large impact on the ability to find good problem solutions. Problem characterization became an important step in the application of intelligent methods to practical problems. A popular approach to problem characterization is the exploratory landscape analysis. It consists of a sequence of operations that approximate and describe the hypersurfaces formed by characteristic problem properties from a limited sample of solutions. Exploratory landscape analysis uses a particular strategy to select just a small subset of problem solutions for which are the characteristic properties evaluated and high-level landscape features computed. Low-discrepancy sequences have been recently used to design a family of sampling strategies. They have useful space-filling properties but their effective and efficient randomization might represent an issue. In this work, we study the Cranley-Patterson rotation, a lightweight randomization strategy for low-discrepancy sequences, compare it with other randomization methods, and observe the effect its use has on the randomization of sets of sampling points in the context of exploratory landscape analysis.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GF22-34873K" target="_blank" >GF22-34873K: Vícekriteriální optimalizace s omezeními pomocí analýzy potenciálních ploch</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2023

Kód důvěrnosti údajů

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

Název statě ve sborníku

ACM International Conference Proceeding Series 2023

ISBN

979-8-4007-0849-7

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

1-8

Název nakladatele

Association for Computing Machinery

Místo vydání

New York

Místo konání akce

Bangkok

Datum konání akce

6. 12. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—