Design of high‐frequency structures, including microwave and antenna components, heavily relies on full‐wave electromagnetic (EM) simulation models. Their reliability comes at a price of a considerable computational cost. This may lead to practical issues whenever numerous EM analyses are to be executed, e.g., in the case of parametric optimization. The difficulties entailed by massive simulations may be mitigated by the use of fast surrogates, among which data‐driven models are the most popular ones due to their versatility and accessibility. Unfortunately, conventional modeling techniques are significantly affected by the curse of dimensionality. It is particularly restrictive in the case of high‐frequency components, typically exhibiting highly nonlinear characteristics. Recently, the concept of performance‐driven modeling has been proposed where the surrogate model setup is focused on a small subset of the parameter space, containing the designs that are optimal or nearly optimal with respect to the considered performance figures. Domain confinement allows for a dramatic reduction of the number of training data samples needed for rendering reliable surrogates valid over wide ranges of the system parameters. In this paper, we review some of the recent techniques employing these concepts, discuss their properties, and illustrate them using real‐world examples of antenna and microwave components.
Authors
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1002/jnm.2706
- Category
- Publikacja w czasopiśmie
- Type
- artykuły w czasopismach
- Language
- angielski
- Publication year
- 2020
