Selecting model fidelity for antenna design using surrogate-based optimization

Efficient electromagnetic (EM)-driven antenna design can be performed using surrogate-based optimization (SBO). Surrogate models of antennas are typically constructed using coarse-discretization EM simulations. A proper selection of the model fidelity is a key factor that influences both the computational cost of the design process and the final design. Here, we carry out a numerical study that aims at finding a trade-off between the design cost and reliability of the SBO algorithms. Numerical results obtained for two antenna design cases, are followed by recommendations regarding the proper selection of the surrogate model coarseness.