Microwave design optimization using local response surface approximations and variable-fidelity electromagnetic models

A computationally efficient algorithm for simulation driven design optimization of microwave structures is proposed. Our approach exploits variable-fidelity electromagnetic models of the structure under consideration. The low-fidelity model is optimized using a response surface approximation technique. The high-fidelity model is refined by space mapping with polynomial interpolation of the low-fidelity model data used as an underlying coarse model. Our algorithm is carefully developed to minimize the number off evaluations of both the low- and high-fidelity model in the optimization process. The operation and efficiency of the approach is demonstrated through design of a microstrip filter and an ultrawideband monopole antenna. A comparison with other design approaches, including the direct high-fidelity model optimization, is also presented.