An effective interpolation scheme for multi-resonant antenna responses: generalized Stoer-Bulirsch algorithm with adaptive sampling

Multiple analysis calls and the high computational demand for each analysis present themselves as important bottlenecks in conducting practical electromagnetic design optimization studies particularly for complex material designs in metamaterial studies. To enable efficient reanalysis in such studies, in this paper we develop an algorithm for rational interpolation by generalizing the Stoer-Bulirsch algorithm to allow for non-diagonal Neville paths rather than the known standard diagonal path to enhance its interpolation capability. The algorithm is then integrated to an adaptive sampling strategy that exploits the non-diagonality and finds an optimum Neville path that provides more efficient and reliable fittings for multi-resonance antenna response functions. The resulting technique is applied to return loss responses of antenna models with textured material substrates and complex conductor topologies. Interpolation results are compared to the performance of a standard Stoer-Bulirsch algorithm. Results show that the proposed generalized scheme outperforms the existing Stoer-Bulirsch technique in terms of computational accuracy by detecting resonances while still maintaining minimum number of support points. To demonstrate the capability of the proposed generalized algorithm, it is adapted to a large-scale antenna design optimization example achieving significant bandwidth performance enhancements for a novel antenna structure within practical timespans.