An approach for the design of reflectarrays using CG-FFT and PSO

A method applied to the design of reflectarray antennas that combines a real-valued particle swarm optimization (PSO) algorithm, used to carry out the synthesis of the phases at each reflectarray element, with the conjugate gradient fast Fourier transform (CG-FFT) method considered to obtain the design curves for the elements, is presented in this work. The analysis method has been validated against results available in the literature for rectangular patches. Furthermore, the method is able to analyze any geometry that can be properly modeled by a uniform mesh on both directions in the unit cell. At this moment, only one metallic layer, apart from the ground plane, can be analyzed. Radiation patterns for both medium and large size reflectarrays are included, achieving accurate results and demonstrating the usefulness of the approach and the ability of the PSO algorithm to manage with such high-dimensional optimization problems. It also includes a comparison between PSO and other methods such as micro-genetic algorithms (μGΑ) or hybrid schemes combining PSO and μGΑ.