Fuzinsumo ‐ Fuzzy investigation with surrogate models

Oberleiter T, Willner K (2018)


Publication Language: English

Publication Type: Conference contribution, Conference Contribution

Publication year: 2018

Publisher: WILEY‐VCH Verlag GmbH & Co. KGaA

City/Town: Weinheim

Pages Range: 725-726

Conference Proceedings Title: PAMM, Volume 17

Event location: Ilmenau/Weimar

DOI: 10.1002/pamm.201710331

Abstract

In this contribution, we investigate the accuracy of surrogate models for fuzzy systems. Within a virtual tolerance analysis, unknown uncertainties are considered by fuzzy numbers. These fuzzy numbers consist of higher‐dimensional intervals, where each interval is assigned a membership value. To reproduce fuzzy numbers correctly, a large number of system evaluations is needed. Instead of evaluating numerically expensive full systems, we employ surrogate models with radial basis functions. In order to find sufficiently accurate surrogate models with as few as possible sampling points, we investigate the influence of the sampling points on the respective fuzzy number intervals. As an example, we investigate the tilting of an X‐ray aperture, which shows different accuracy for different sampling strategies. The reference results for the nonlinear multi‐body system are calculated by a Matlab optimizer and are compared to the results of the surrogate models.

Authors with CRIS profile

Related research project(s)

How to cite

APA:

Oberleiter, T., & Willner, K. (2018). Fuzinsumo ‐ Fuzzy investigation with surrogate models. In PAMM, Volume 17 (pp. 725-726). Ilmenau/Weimar: Weinheim: WILEY‐VCH Verlag GmbH & Co. KGaA.

MLA:

Oberleiter, Thomas, and Kai Willner. "Fuzinsumo ‐ Fuzzy investigation with surrogate models." Proceedings of the 88th GAMM Annual Meeting, Ilmenau/Weimar Weinheim: WILEY‐VCH Verlag GmbH & Co. KGaA, 2018. 725-726.

BibTeX: Download