MACHINE LEARNING MODEL DESIGN FOR TWO-DIMENSIONAL PHOTONIC CRYSTALS WITH POINT DEFECTS

The discovery of photonic crystals has opened up a brand-new path for controlling the propagation of light and electromagnetic waves. Its unique photonic bandgap characteristics enable photonic crystals to effectively restrict and regulate the behavior of photons, similar to the bandgap effect of electrons in semiconductors. In photonic crystals, by adjusting the parameters of the dielectric column, the band gap structure, band gap position, defect mode and the intrinsic field distribution of the defect mode can be adjusted. The plane wave propagation method (PWE) and the finite-difference time-domain method (FDTD), which have developed in recent years, as effective numerical calculation methods, have been widely applied in the design and analysis of optical waveguide devices. However, both of these calculation methods require a very large amount of computation, usually necessitating traversal calculations of design parameters, and the calculation time depends on the size, structure, and precision of the device. This study aims to compare the performance of various machine learning models in the prediction and optimization design of two-dimensional photonic crystal structure parameters with point defects, in order to achieve accurate prediction and optimization design of relevant parameters of two-dimensional photonic crystals with point defects.