Publication detail

Integer simulation based optimization by local search

SKLENÁŘ, J. POPELA, P.

Czech title

Celočíselná optimalizace založená na simulaci pomocí lokálního vyhledávání

English title

Integer simulation based optimization by local search

Type

journal article - other

Language

en

Original abstract

Simulation-based optimization combines simulation experiments used to evaluate the objective and/or constraint functions with an optimization algorithm. Compared with classical optimization, simulation based optimization brings its specific problems and restrictions. These are discussed in the paper. Evaluation of the objective function is based on time consuming, typically repeated simulation experiments. So we believe that the main objective in selecting the optimization algorithm is minimization of the number of objective function evaluations. In this paper we concentrate on integer optimization that is typical in simulation context. Local search algorithms that try to minimize the number of objective function evaluations are described. Examples with both analytical and simulationbased objective functions are used to demonstrate the performance of the algorithms.

Czech abstract

Ve srovnání s klasickou optimalizací, optimalizace založená na simulaci přináší své specifické problémy a omezení. Vyhodnocení účelové funkce je často založeno na časově náročných, obvykle opakovaných simulačních experimentech. Proto se zaměřujeme na hlavní cíl, kterým je při výběru optimalizačního algoritmu minimalizace výpočtů hodnot účelové funkce. Zabýváme se celočíselnou optimalizací, které je typická v kontextu simulace. Jsou popsány algoritmy lokálního prohledávání zohledňující výše popsaný problém. Uvedené příklady slouží k ilustraci navržených algoritmů.

English abstract

Simulation-based optimization combines simulation experiments used to evaluate the objective and/or constraint functions with an optimization algorithm. Compared with classical optimization, simulation based optimization brings its specific problems and restrictions. These are discussed in the paper. Evaluation of the objective function is based on time consuming, typically repeated simulation experiments. So we believe that the main objective in selecting the optimization algorithm is minimization of the number of objective function evaluations. In this paper we concentrate on integer optimization that is typical in simulation context. Local search algorithms that try to minimize the number of objective function evaluations are described. Examples with both analytical and simulationbased objective functions are used to demonstrate the performance of the algorithms.

Keywords in Czech

Celočíselná optimalizace, lokální prohledávání, simulace

Keywords in English

Integer optimization; Local search; Simulation

Released

30.05.2010

ISSN

1877-0509

Volume

1

Number

1

Pages from–to

1341–1348

Pages count

8

BIBTEX


@article{BUT124192,
  author="Jaroslav {Sklenář} and Pavel {Popela},
  title="Integer simulation based optimization by local search",
  year="2010",
  volume="1",
  number="1",
  month="May",
  pages="1341--1348",
  issn="1877-0509"
}