Optimization of optical parameters for the design of multilayer bandpass filter using genetic algorithm

Abstract

This work investigates the use of genetic algorithm to design a multilayer band pass filter in visible range. The effect of choice of optical parameters such as thickness and refractive index as optimization variables on the convergence of design-solution is studied to obtain the best design that matches the desired one. Firstly, the thickness is varied; high and low refractive index combination of dielectric materials are used to design the alternate multilayer stack of 28 layers. Secondly, refractive index is varied keeping the thickness of layers constant. In both cases, the algorithm either optimizes thickness or refractive index of each layer to get the best possible solution. Initially, best matching is obtained with thickness as optimisation variable. Further, genetic parameters, population size and crossover type were varied and performance of the filter was analysed. The result of this study indicates that effect of choice of variable is dependent on the design specifications and it is not distinct. Although choosing thickness as the optimization variable showed slightly better convergence in terms of merit function when genetic parameters were not varied but this cannot be generalized as it is purely dependent on the design specification.