“…Various metamodeling techniques have been developed over the years which are applied to expedite computationally expensive tasks such as multi-disciplinary design and optimization of aircraft wing geometries, robustness-and/or reliability-based optimization of antenna structures, sensitivity analysis, etc. For example, implementation of low-cost antenna models is possible using various approximation techniques such as polynomial regression [2], radial basis function interpolation [3], Kriging [4,5] support vector regression [6][7][8][9], fuzzy systems [10,11], multidimensional Cauchy approximation [12], or artificial neural networks [13][14][15][16], etc. A common problem associated with most of the metamodeling approaches is the fact that they demand for a high model setup cost: in order to ensure usable accuracy a large number of training sample points is necessary, which quickly grows with the dimensionality of the design space (a problem often referred to as the curse of dimensionality) [2,17,18].…”