Optimal design of frequency selective surfaces with fractal motifs

An alternative electromagnetic (EM) optimisation technique for the optimal design of frequency selective surfaces (FSSs) with fractal motifs is described. Based on computational intelligence tools, the proposed technique overcomes the high computational cost associated with FSS parametric full-wave analysis. In an application example, a fast and accurate multilayer perceptrons model of a FSS band-stop spatial filter with a Vicsek fractal motif is developed. This neural network model is used for repetitive cost function computations in population-based search algorithm simulations. A bees algorithm, continuous genetic algorithm and particle swarm optimisation are used for FSS optimisation with specific resonant frequency and bandwidth. The performance of these algorithms is compared in terms of numerical convergence. Consistent results are presented for a second-pass of designed FSS prototype with Vicsek fractal elements.