Metamodeling of high-contrast-index gratings for color reproduction.

We propose a metamodel-based optimization technique to tailor the chromatic response of high-contrast-index gratings. The algorithm, which couples a population-based metaheuristic with a neural network, is used to retrieve the optimal geometrical parameters of a grating to reproduce a prescribed color. By means of some examples, we assess the possibilities and limitations of our optimization scheme. The numerical evidence found shows that the metamodel approach offers an alternative to traditional metaheuristic techniques that not only provides the best solution for a given geometry and a material but also significantly improves the computing time required for the optimization process.