Detail publikace

FORMULATION OF POTENTIAL FOR DYNAMICAL PARTICLE SYSTEM APPLIED TO MONTE CARLO SAMPLING

MAŠEK, J. VOŘECHOVSKÝ, M.

Originální název

FORMULATION OF POTENTIAL FOR DYNAMICAL PARTICLE SYSTEM APPLIED TO MONTE CARLO SAMPLING

Typ

článek ve sborníku ve WoS nebo Scopus

Jazyk

angličtina

Originální abstrakt

The presented paper investigates the effect of formulation of energy potential of a dynamical particle system as used for optimization of statistical point sampling. The dynamical particle system, originally developed as a physical analogy of the Audze-Eglajs (AE) optimization criterion and its periodical modification (PAE), effectively demonstrated that the originally proposed potential performs well only in poorly applicable scenarios of design spaces of low dimension filled with rather high number of design points. A remedy lying in a refined formulation of energy potential as well as its derivation and reasoning are presented.

Klíčová slova

Particle dynamical system, statistical sampling optimization, periodic space, particle interaction laws, short- and long-range interactions, self-similar designs.

Autoři

MAŠEK, J.; VOŘECHOVSKÝ, M.

Vydáno

15. 6. 2017

Nakladatel

National Technical University of Athens

Místo

Rhodes Island

ISBN

978-6-1882-8444-9

Kniha

UNCECOMP 2017 2nd ECCOMAS Thematic Conference on Uncertainty Quantification in Computational Sciences and Engineering

Strany od

467

Strany do

477

Strany počet

11

URL

https://doi.org/10.7712/120217.5384.17000

BibTex

@inproceedings{BUT146057,
  author="Jan {Mašek} and Miroslav {Vořechovský}",
  title="FORMULATION OF POTENTIAL FOR DYNAMICAL PARTICLE SYSTEM APPLIED TO MONTE CARLO SAMPLING",
  booktitle="UNCECOMP 2017 2nd ECCOMAS Thematic Conference on
Uncertainty Quantification in Computational Sciences and Engineering",
  year="2017",
  pages="467--477",
  publisher="National Technical University of Athens",
  address="Rhodes Island",
  doi="10.7712/120217.5384.17000",
  isbn="978-6-1882-8444-9",
  url="https://doi.org/10.7712/120217.5384.17000"
}