Application of Genetic Algorithm for Optimization Photonic Crystals

In this paper, great attention has been paid to photonic crystals due to their potential applications in ultra compact photonic integrated circuits. The goal of this work is applying a genetic algorithm to search for photonic crystals with large band gaps. The algorithm adopted in this work is based on introducing single-pixel perturbations to an initial seed with small relative band gap. Each filling pattern of the grids is translated into a matrix of binary elements. By applying innovative crossover and mutation operators, the number of generations needed to achieve the fittest photonic crystal structure is considerably reduced. Besides, the required time to get optimal structure is significantly decreased. This is due to decreasing both the time needed to evaluate a band structure of each chromosome and generation number.  Because of this efficient method, the required computation time of the fitness function is significantly decreased. The paper presents two optimized photonic crystal structures with almost 21% relative band gap.