Aplicação do algoritmo Artificial Bee Colony na otimização inversa de cristais fotônicos bidimensionais
Data
2020-07-23
Autores
Delalibera, Fabianno Gianelli
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Neste trabalho, apresenta-se uma nova proposta de aplicação do algoritmo Artificial Bee Colony (ABC) em um problema eletromagnético para otimização de estruturas cristalinas fotônicas bidimensionais em arranjo triangular, buscando maximizar a banda proibida fotônica em estruturas compostas por silício e ar, tanto para o modo transversal elétrico quanto para o transversal magnético. Estruturas otimizadas via Genetic Algorithm (GA), apresentadas na literatura, foram utilizadas como parâmetro comparativo. O estudo foi realizado, inicialmente, submetendo os algoritmos clássicos do ABC e GA à busca pelo valor máximo de duas funções bidimensionais previamente conhecidas, com o objetivo de compreender o comportamento de busca do ABC para, posteriormente, ser configurado à otimização de estruturas fotônicas. Foram necessárias algumas modificações no ABC clássico, para otimização de problemas em um espaço de buscas discreto. Os resultados apresentados pelo ABC foram bastante satisfatórios, tendo encontrado estruturas com percentual de banda proibida fotônica superiores ao encontrado pelo GA. Sendo assim, o ABC apresentou uma melhor eficiência para o problema proposto em comparação ao GA.
In this work, it is show a novel proposal for the application of the Artificial Bee Colony (ABC) algorithm in an electromagnetic problem, for optimization of two-dimensional photonic crystal structure in triangular lattice, seeking to maximize the photonic band gap (PBG) in structures composed of silicon and air, both for the modes transverse electric and transverse magnetic. Structures optimized by Genetic Algorithm (GA), presented in the literature, were used as a comparative parameter. The study was initially carried out by submitting the classical algorithms of ABC and GA to search for the maximum value of two previously known two-dimensional functions, with the objective of understanding the search behavior of ABC to later configure the optimization of photonic structures. Some modifications were needed in the classic ABC to optimize problems in a discrete search space. The results presented by the ABC were quite satisfactory, finding structures with a percentage of PBG higher than that found by GA. Therefore, the ABC presented a better efficiency for the proposed problem compared to GA.
In this work, it is show a novel proposal for the application of the Artificial Bee Colony (ABC) algorithm in an electromagnetic problem, for optimization of two-dimensional photonic crystal structure in triangular lattice, seeking to maximize the photonic band gap (PBG) in structures composed of silicon and air, both for the modes transverse electric and transverse magnetic. Structures optimized by Genetic Algorithm (GA), presented in the literature, were used as a comparative parameter. The study was initially carried out by submitting the classical algorithms of ABC and GA to search for the maximum value of two previously known two-dimensional functions, with the objective of understanding the search behavior of ABC to later configure the optimization of photonic structures. Some modifications were needed in the classic ABC to optimize problems in a discrete search space. The results presented by the ABC were quite satisfactory, finding structures with a percentage of PBG higher than that found by GA. Therefore, the ABC presented a better efficiency for the proposed problem compared to GA.
Descrição
Palavras-chave
Fotônica, Ondas eletromagneticas, Otimização matemática