Publication detail
Monte carlo based detection of parameter correlation in simulation models
NAJMAN, J. BRABLC, M. RAJCHL, M. BASTL, M. SPÁČIL, T. APPEL, M.
English title
Monte carlo based detection of parameter correlation in simulation models
Type
conference paper
Language
en
Original abstract
Simulation models which are of high order or are automatically generated via modelling software are usually depended on high number of unknown parameters. In this paper we present a method for detecting correlation between these parameters and identifying the subspace shape for their uncorrelated complements. This can be further used to lower the order of the optimization problem. For our low-order examples the methods’ operating principle is visualized and the subspace is shown.
English abstract
Simulation models which are of high order or are automatically generated via modelling software are usually depended on high number of unknown parameters. In this paper we present a method for detecting correlation between these parameters and identifying the subspace shape for their uncorrelated complements. This can be further used to lower the order of the optimization problem. For our low-order examples the methods’ operating principle is visualized and the subspace is shown.
Keywords in English
MATLAB;Model-Based Design;Monte Carlo;Parameter correlation;Parameter estimation;PCA;Simulation;Simulink;
Released
16.08.2019
Publisher
Springer Verlag
Location
Warsaw
ISBN
9783030299927
ISSN
2194-5357
Book
Advances in Intelligent Systems and Computing - Mechatronics 2019: Recent Advances Towards Industry 4.0
Volume
1044
Edition number
1044
Pages from–to
54–61
Pages count
8
BIBTEX
@inproceedings{BUT160010,
author="Jan {Najman} and Martin {Brablc} and Matej {Rajchl} and Michal {Bastl} and Tomáš {Spáčil} and Martin {Appel},
title="Monte carlo based detection of parameter correlation in simulation models",
booktitle="Advances in Intelligent Systems and Computing - Mechatronics 2019: Recent Advances Towards Industry 4.0",
year="2019",
volume="1044",
month="August",
pages="54--61",
publisher="Springer Verlag",
address="Warsaw",
isbn="9783030299927",
issn="2194-5357"
}