Optimization of FSS with Sierpinski island fractal elements using population-based search algorithms and MLP neural network

Abstract

This work presents an electromagnetic optimization technique that blending full‐wave method, artificial neural network, and population‐based search algorithms for optimal design of frequency selective surfaces (FSSs) with fractal motifs. We consider a simple application of this technique in a single‐layer FSS with Sierpinski island fractal patch elements. The optimization technique replaces the computational intensive full‐wave method of moments simulations by a fast and accurate multilayer perceptrons neural network model of FSS spatial filter, which is used to compute the cost (or fitness) function in the search algorithms iterations. In the FSS optimization with specific resonant frequency and bandwidth, we use: bees algorithm, continuous genetic algorithm, and particle swarm optimization. A FSS prototype is built and measured for a first level of Sierpinski island fractal. The accuracy of the proposed optimization technique is verified through the excellent agreement (0.9%) obtained by means of comparisons between theoretical and experimental results. © 2014 Wiley Periodicals, Inc. Microwave Opt Technol Lett 56:827–831, 2014

Revisão Teórica Sobre as Técnicas de Otimização Metaheurística na Síntese de Superfícies Seletivas de Frequência

“…Embora a MLP já tenha sido implementada para trabalhos envolvendo FSS sem qualquer outro algoritmo auxiliador, como em [10], a utilização deste tipo de rede para otimização bioinspirada de FSS nota-se em [4,5,7,8]: um compilado de trabalhos científicos publicados entre 2009 e 2014. Em suma, a MLP é aplicada para sintetizar a estrutura, servindo também como a função de ajuste (também chamada de função de fitness) para o algoritmo genético ou PSO selecionados para compor a solução híbrida proposta.…”

“…Neste ínterim, autores têm realizado a junção de redes neurais artificiais e algoritmos de otimização natural, como os clássicos algoritmos genéticos e a otimização por enxame de partículas, ou ainda, os mais recentes algoritmos de busca cuco no projeto e otimização das superfícies seletivas de frequência [2]. Uma extensa pesquisa analítica tem sido realizada para descrever as propriedades de reflexão e transmissão das FSS com o emprego de técnicas de otimização para se obter as dimensões exatas, necessárias ao atendimento dos requisitos de projeto [3,4,5].…”

A new efficient and proper modeling of isotropic/uniaxial anisotropic substrate specifications in design procedures of metasurfaces

“…So, FSS design procedures consisting of these networks must be accompanied by the suitable optimization algorithm and desired convergence conditions. Three methods exist by which substrate specifications can be considered in the ANN-based design procedures: Both ε r and h are assumed to be constant. Both ε r and h are the direct input of ANNs, like other geometric dimensions [24–27]. Neither ε r nor h is the direct input/output of ANNs, but ANNs can transform geometric dimensions in proportion to the chosen substrate. The first method is suitable for the special cases, in which ε r and h are determinate, but in general-purpose algorithms this method may not support all of the applications.…”

A fast and accurate technique for FSS and antenna designs based on the social spider optimization algorithm

“…It is common to find studies using evolutionary algorithms with the aid of a neural network or even interpolation functions . The use of a neural network or interpolation is due to the need to evaluate candidate solutions.…”