Design Goals and Optimization Methods
As mentioned previously, the typical primary requirement for an ideal horn liner is that it be a soft or hard surface. Certain periodic surfaces, such as corrugated metals, mushroom-like textured surfaces, and uniplanar compact band-gap structures, have been designed to exhibit soft and hard characteristics [19, 26]. Achieving a soft or hard boundary over a broad bandwidth is a significant challenge, particularly while maintaining the hybrid-mode condition. Global stochastic optimizers such as the genetic algorithm (GA) and more recent improvements [27, 28] provide a means for designing metamaterial surfaces that exhibit broadband and multiband operation , or that achieve a specific frequency dispersion characteristic. The examples presented here were based on a GA optimization of the thickness t, the unit cell width (periodicity) w, and the metallic pixel pattern. The following cost functions were applied to the various optimizations:
The simulations and optimizations of unit cells were performed using a periodic finite-element boundary integral solver, which is now commercially available as PFSS from E x H, Inc. . The following section presents several examples of metasurfaces that were designed using this optimization approach.