Multilayer perceptron models for band diagram prediction in bi-dimensional photonic crystals

Abstract

We modeled Multilayer Perceptron (MLP) Artificial Neural Network for predicting band diagrams (BD) of bi-dimensional photonic crystals. Datasets for MLP training were created by relating geometric and material properties to BDs of triangular-and square-lattice photonic crystals. We demonstrate that fast-Training MLP models are able to estimate accurate BDs and existing photonic band gaps through rapid computations.