Adaptive Weighted Expected Improvement With Rewards Approach in Kriging Assisted Electromagnetic Design
Adaptive Weighted Expected Improvement With Rewards Approach in Kriging Assisted Electromagnetic Design
The paper explores kriging surrogate modelling combined with expected improvement approach for the design of electromagnetic devices. A novel algorithm based on the concept of rewards is proposed, tested and demonstrated in the context of TEAM Workshop Problem 22. Balancing exploration and exploitation is emphasized and robustness of the design considered.
2057-2060
Xiao, Song
6ffa9657-513e-4b86-86a2-e560d3c09c72
Rotaru, M.
c53c5038-2fed-4ace-8fad-9f95d4c95b7e
Sykulski, J.K.
d6885caf-aaed-4d12-9ef3-46c4c3bbd7fb
May 2013
Xiao, Song
6ffa9657-513e-4b86-86a2-e560d3c09c72
Rotaru, M.
c53c5038-2fed-4ace-8fad-9f95d4c95b7e
Sykulski, J.K.
d6885caf-aaed-4d12-9ef3-46c4c3bbd7fb
Xiao, Song, Rotaru, M. and Sykulski, J.K.
(2013)
Adaptive Weighted Expected Improvement With Rewards Approach in Kriging Assisted Electromagnetic Design.
IEEE Transactions on Magnetics, 49 (5), Spring Issue, .
(doi:10.1109/TMAG.2013.2240662).
Abstract
The paper explores kriging surrogate modelling combined with expected improvement approach for the design of electromagnetic devices. A novel algorithm based on the concept of rewards is proposed, tested and demonstrated in the context of TEAM Workshop Problem 22. Balancing exploration and exploitation is emphasized and robustness of the design considered.
Text
IEEEvol49no5May2013page2057.pdf
- Version of Record
More information
Published date: May 2013
Organisations:
EEE
Identifiers
Local EPrints ID: 352697
URI: http://eprints.soton.ac.uk/id/eprint/352697
ISSN: 0018-9464
PURE UUID: 92f0f3e5-305b-4071-878c-161629eae5f0
Catalogue record
Date deposited: 17 May 2013 08:05
Last modified: 15 Mar 2024 02:34
Export record
Altmetrics
Contributors
Author:
Song Xiao
Author:
M. Rotaru
Author:
J.K. Sykulski
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics