Generating a Multi-fidelity Simulation Model Estimating the Models’ Applicability with Machine Learning Algorithms

Christian Hürten, Philipp Sieberg, Dieter Schramm

2022

Abstract

Having access to large data sets recently gained increasing importance, especially in the context of automation systems. Whether for the development of new systems or for testing purposes, a large amount of data is required to satisfy the development goals and admission standards. This data is not only measured from real-world tests, but with growing tendency generated from simulations, facing a trade-off between computational effort and simulation model fidelity. This contribution proposes a method to assign individual simulation runs the simulation model that has the lowest computation costs while still being capable of producing the desired simulation output accuracy. The method is described and validated using support vector machines and artificial neural networks as underlying vehicle simulation model classifiers in the development of a lane change decision system.

Paper Citation

in Harvard Style

Hürten C., Sieberg P. and Schramm D. (2022). Generating a Multi-fidelity Simulation Model Estimating the Models’ Applicability with Machine Learning Algorithms. In Proceedings of the 12th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH, ISBN 978-989-758-578-4, pages 131-141. DOI: 10.5220/0011318100003274

in Bibtex Style

@conference{simultech22,
author={Christian Hürten and Philipp Sieberg and Dieter Schramm},
title={Generating a Multi-fidelity Simulation Model Estimating the Models’ Applicability with Machine Learning Algorithms},
booktitle={Proceedings of the 12th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH,},
year={2022},
pages={131-141},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011318100003274},
isbn={978-989-758-578-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH,
TI - Generating a Multi-fidelity Simulation Model Estimating the Models’ Applicability with Machine Learning Algorithms
SN - 978-989-758-578-4
AU - Hürten C.
AU - Sieberg P.
AU - Schramm D.
PY - 2022
SP - 131
EP - 141
DO - 10.5220/0011318100003274