Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 8.53 MB
Add to list
  1. Search results
  2. Bayesian optimization for microwave d...

Bayesian optimization for microwave devices using deep GP spectral surrogate models

Federico Garbuglia, Domenico Spina (UGent) , Dirk Deschrijver (UGent) , Ivo Couckuyt (UGent) and Tom Dhaene (UGent)
(2023) IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES. 71(6). p.2311-2318
Author
Organization
Project
Abstract
In microwave design, Bayesian optimization (BO) techniques have been widely applied to the optimization of the frequency response of components and devices. The common approach in BO is to model and maximize an objective function over the design parameters, in order to find the optimal spectral response. Such an approach avoids the direct modeling of spectral responses, which is a challenging task for the typical data-efficient surrogate models used in BO. Simple objective functions may lead to a suboptimal solutions, while complicated objectives require more powerful and less data-efficient surrogate models. To resolve this issue, this article proposes to adopt a deep Gaussian process (DGP) to directly model all relevant $S$ coefficients over the frequency and the design parameter ranges of interest. Subsequently, an objective probability distribution is retrieved from the DGP model and maximized using a BO scheme. The proposed approach is tested on two suitable microwave examples and compared to the standard BO approach. Results show increased accuracy in identifying the optimal frequency response for the given design parameters and the desired objective, while maintaining high data efficiency.
Keywords
DESIGN, Bayesian optimization (BO), deep Gaussian processes (DGPs), electronic, design automation (EDA), S coeffi-cients

Downloads

  • Download
    8232 acc.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 4.36 MB
  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 4.18 MB

Citation

Please use this url to cite or link to this publication:

MLA
Garbuglia, Federico, et al. “Bayesian Optimization for Microwave Devices Using Deep GP Spectral Surrogate Models.” IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, vol. 71, no. 6, 2023, pp. 2311–18, doi:10.1109/TMTT.2022.3228951.
APA
Garbuglia, F., Spina, D., Deschrijver, D., Couckuyt, I., & Dhaene, T. (2023). Bayesian optimization for microwave devices using deep GP spectral surrogate models. IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, 71(6), 2311–2318. https://doi.org/10.1109/TMTT.2022.3228951
Chicago author-date
Garbuglia, Federico, Domenico Spina, Dirk Deschrijver, Ivo Couckuyt, and Tom Dhaene. 2023. “Bayesian Optimization for Microwave Devices Using Deep GP Spectral Surrogate Models.” IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES 71 (6): 2311–18. https://doi.org/10.1109/TMTT.2022.3228951.
Chicago author-date (all authors)
Garbuglia, Federico, Domenico Spina, Dirk Deschrijver, Ivo Couckuyt, and Tom Dhaene. 2023. “Bayesian Optimization for Microwave Devices Using Deep GP Spectral Surrogate Models.” IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES 71 (6): 2311–2318. doi:10.1109/TMTT.2022.3228951.
Vancouver
1.
Garbuglia F, Spina D, Deschrijver D, Couckuyt I, Dhaene T. Bayesian optimization for microwave devices using deep GP spectral surrogate models. IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES. 2023;71(6):2311–8.
IEEE
[1]
F. Garbuglia, D. Spina, D. Deschrijver, I. Couckuyt, and T. Dhaene, “Bayesian optimization for microwave devices using deep GP spectral surrogate models,” IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, vol. 71, no. 6, pp. 2311–2318, 2023.
@article{01GPJQG9DGFRJ49CXYYS1RZJAJ,
  abstract     = {{In microwave design, Bayesian optimization (BO) techniques have been widely applied to the optimization of the frequency response of components and devices. The common approach in BO is to model and maximize an objective function over the design parameters, in order to find the optimal spectral response. Such an approach avoids the direct modeling of spectral responses, which is a challenging task for the typical data-efficient surrogate models used in BO. Simple objective functions may lead to a suboptimal solutions, while complicated objectives require more powerful and less data-efficient surrogate models. To resolve this issue, this article proposes to adopt a deep Gaussian process (DGP) to directly model all relevant $S$ coefficients over the frequency and the design parameter ranges of interest. Subsequently, an objective probability distribution is retrieved from the DGP model and maximized using a BO scheme. The proposed approach is tested on two suitable microwave examples and compared to the standard BO approach. Results show increased accuracy in identifying the optimal frequency response for the given design parameters and the desired objective, while maintaining high data efficiency.}},
  author       = {{Garbuglia, Federico and Spina, Domenico and Deschrijver, Dirk and Couckuyt, Ivo and Dhaene, Tom}},
  issn         = {{0018-9480}},
  journal      = {{IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES}},
  keywords     = {{DESIGN,Bayesian optimization (BO),deep Gaussian processes (DGPs),electronic,design automation (EDA),S coeffi-cients}},
  language     = {{eng}},
  number       = {{6}},
  pages        = {{2311--2318}},
  title        = {{Bayesian optimization for microwave devices using deep GP spectral surrogate models}},
  url          = {{http://doi.org/10.1109/TMTT.2022.3228951}},
  volume       = {{71}},
  year         = {{2023}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: