An Expedient DDF-Based Implementation of Perfectly Matched Monolayer

Alternative unsplit-field-based absorbing boundary condition (ABC) computation approach for the finite-difference time-domain (FDTD) is efficiently proposed based upon the deep differentiable forest (DDF) model which is introduced to replace the conventional perfectly matched layer (PML) ABC during the computation process of FDTD. The field component data on the interface of traditional PML are adopted to train the DDF-based model for implementing the perfectly matched monolayer (PMM). The DDF has the advantages of both trees and neural networks. Its tree structure is easy to use and explain for the numerical PML data. Meanwhile, it has full differentiability like neural networks. The DDF can be trained by powerful deep learning techniques. Numerical cases have been carried out to benchmark the performance of the proposed method. Results illustrate that the DDF-based PMM could not only replace the traditional PML but be integrated into FDTD computation process with satisfactory numerical accuracy.