A Machine-Learning Method for Time-Dependent Wave Equations over Unbounded Domains

Abstract

Time-dependent wave equations represent an important class of partial differential equations (PDE) for describing wave propagation phenomena, which are often formulated over unbounded domains. Given a compactly supported initial condition, classical numerical methods reduce such problems to bounded domains using artificial boundary condition (ABC). In this work, we present a machine-learning method to solve this type of equations as an alternative to ABCs. Specifically, the mapping from the initial conditions to the PDE solution is represented by a neural network, trained using wave packets that are parameterized by their band width and wave numbers. The accuracy is tested for both the second-order wave equation and the Schrodinger equation, including the nonlinear Schrodinger equation. We examine the accuracy from both interpolations and extrapolations. For initial conditions lying in the training set, the learned map has good interpolation accuracy, due to the approximation property of deep neural networks. The learned map also exhibits some good extrapolation accuracy. We also demonstrate the effectiveness of the method for problems in irregular domains. Overall, the proposed method provides an interesting alternative for finite-time simulation of wave propagation.

References in corpus (14)

• Deep Residual Learning for Image Recognition
• DGM: A deep learning algorithm for solving partial differential equations
• NSFnets (Navier-Stokes Flow nets): Physics-informed neural networks for the incompressible Navier-Stokes equations
• B-PINNs: Bayesian Physics-Informed Neural Networks for Forward and Inverse PDE Problems with Noisy Data
• Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
• Weak Adversarial Networks for High-dimensional Partial Differential Equations
• Computational methods for the dynamics of the nonlinear Schrodinger/Gross-Pitaevskii equations
• Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations
• Matching Conditions in Atomistic-Continuum Modeling of Materials
• Overcoming the curse of dimensionality in the approximative pricing of financial derivatives with default risks
• Numerical simulations for full history recursive multilevel Picard approximations for systems of high-dimensional partial differential equations
• MIM: A deep mixed residual method for solving high-order partial differential equations
• Absorbing boundary conditions for the time-dependent Schrödinger-type equations in ${\mathbb{R}}^3$
• Learning Unsplit-field-based PML for the FDTD Method by Deep Differentiable Forest