Support Vector Regression Model for Millimeter Wave Transitions

Meng, Jicheng; Lei, Xia

doi:10.1007/s10762-007-9212-1

Cited by 36 publications

(21 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Allocate N cd samples in X * cd and obtain the coarse-dicretization EM model data (design of experiments can be performed using, e.g., Latin Hypercube Sampling [33]); 4. Create the coarse model by approximating coarse-discretization EM model data (using, e.g., RBF interpolation [34], kriging [35], support vector regression [36,37], etc. ).…”

Section: Manifold Mapping Optimization With Function Approximation Comentioning

confidence: 99%

Reliable Simulation-Driven Design Optimization of Microwave Structures Using Manifold Mapping

Koziel¹,

Ciaurri²

2010

PIER B

View full text Add to dashboard Cite

Abstract-A computationally efficient surrogate-based framework for reliable simulation-driven design optimization of microwave structures is described. The key component of our algorithm is manifold mapping, a response correction technique that aligns the coarse model (computationally cheap representation of the structure under consideration) with the accurate but CPU-intensive (fine) model of the optimized device. The parameters of the manifold mapping surrogate are explicitly calculated based on the fine model data accumulated during the optimization process. Also, manifold mapping does not use any extractable parameters, which makes it easy to implement. Robustness and excellent convergence properties of the proposed algorithm are demonstrated through the design of several microwave devices including microstrip filters and a planar antenna.

show abstract

Section: Manifold Mapping Optimization With Function Approximation Comentioning

confidence: 99%

Reliable Simulation-Driven Design Optimization of Microwave Structures Using Manifold Mapping

Koziel¹,

Ciaurri²

2010

PIER B

View full text Add to dashboard Cite

show abstract

“…Therefore, the use of full-wave simulations for executing design tasks such as parametric optimization or tolerance-aware design may be impractical. Fast replacement antenna models (or surrogates) [1], [2] can be obtained by approximating EM simulation data, using, e.g., neural networks [3]- [5], radial basis functions [6], support-vector regression [7]- [9], fuzzy systems [10], or Gaussian process regression (GPR) [11]- [13]. A bottleneck of approximation-based models is high cost of acquiring the training data, with typically a few thousands samples required to ensure reasonable predictive power.…”

Section: Introductionmentioning

confidence: 99%

Cost-efficient dual-stage Gaussian process modeling of antennas

Jacobs

Koziel

2014

The 8th European Conference on Antennas and Propagation (EuCAP 2014)

View full text Add to dashboard Cite

Cost-efficient antenna modeling exploiting variable-fidelity EM simulations is proposed. Our approach is based on a new dual-stage Gaussian process regression method and allows for significant (by 80% or more) reduction of the computational effort necessary to set up the training data sets for the high-fidelity models with insignificant loss in predictive power. The proposed method is using a broadband slot antenna example. Application for design optimization is also discussed.

show abstract

“…The surrogate can be created by approximating high-fidelity EM data using loworder polynomials [10], Kriging [11], neural networks [12], [13], support vector regression [14], [15], or rational approximation [16], etc. However, obtaining an accurate model requires dense sampling of the design space (hundreds or thousands of sampled may be necessary), which makes sense for multiple-use library models but not so much for ad-hoc antenna optimization.…”

Section: Introductionmentioning

confidence: 99%

Cost-efficient electromagnetic-simulation-driven antenna design using co-Kriging

Koziel¹,

Ogurtsov²,

Couckuyt³

et al. 2012

IET Microwaves, Antennas &Amp; Propagation

View full text Add to dashboard Cite

An efficient and reliable technique for design optimization of antennas is presented.Our approach exploits coarse-discretization electromagnetic (EM) simulations of the antenna of interest that are used to create its fast initial model (a surrogate) through Kriging. During the design process, predictions obtained by optimizing the surrogate are verified using highfidelity EM simulations, and these high-fidelity data are used to enhance the surrogate using co-Kriging interpolation that accommodates all EM simulation data into one surrogate model. The co-Kriging-based optimization algorithm is simple, elegant and is capable of yielding a satisfactory design at a low cost equivalent to a few high-fidelity EM simulations of the antenna under design. To our knowledge, this is a first application of co-Kriging to antenna design. Efficiency of our approach is demonstrated with several antenna design cases and compared to other techniques including pattern search and space mapping.

show abstract

Support Vector Regression Model for Millimeter Wave Transitions

Cited by 36 publications

References 8 publications

Reliable Simulation-Driven Design Optimization of Microwave Structures Using Manifold Mapping

Reliable Simulation-Driven Design Optimization of Microwave Structures Using Manifold Mapping

Cost-efficient dual-stage Gaussian process modeling of antennas

Cost-efficient electromagnetic-simulation-driven antenna design using co-Kriging

Contact Info

Product

Resources

About