BINet: Learn to Solve Partial Differential Equations with Boundary Integral Networks

BINet: Learn to Solve Partial Differential Equations with Boundary Integral Networks

Year:    2023

Author:    Guochang Lin, Pipi Hu, Fukai Chen, Xiang Chen, Junqing Chen, Jun Wang, Zuoqiang Shi

CSIAM Transactions on Applied Mathematics, Vol. 4 (2023), Iss. 2 : pp. 275–305

Abstract

We propose a method combining boundary integral equations and neural networks (BINet) to solve (parametric) partial differential equations (PDEs) and operator problems in both bounded and unbounded domains. For PDEs with explicit fundamental solutions, BINet learns to solve, as a proxy, associated boundary integral equations using neural networks. The benefits are three-fold. Firstly, only the boundary conditions need to be fitted since the PDE can be automatically satisfied with single or double layer potential according to the potential theory. Secondly, the dimension of the boundary integral equations is less by one, and as such, the sample complexity can be reduced significantly. Lastly, in the proposed method, all differential operators have been removed, hence the numerical efficiency and stability are improved. Adopting neural tangent kernel (NTK) techniques, we provide proof of the convergence of BINets in the limit that the width of the neural network goes to infinity. Extensive numerical experiments show that, without calculating high-order derivatives, BINet is much easier to train and usually gives more accurate solutions, especially in the cases that the boundary conditions are not smooth enough. Further, BINet outperforms strong baselines for both one single PDE and parameterized PDEs in the bounded and unbounded domains.

You do not have full access to this article.

Already a Subscriber? Sign in as an individual or via your institution

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/csiam-am.SO-2022-0014

CSIAM Transactions on Applied Mathematics, Vol. 4 (2023), Iss. 2 : pp. 275–305

Published online:    2023-01

AMS Subject Headings:    Global Science Press

Copyright:    COPYRIGHT: © Global Science Press

Pages:    31

Keywords:    Partial differential equation boundary integral neural network learning operator.

Author Details

Guochang Lin

Pipi Hu

Fukai Chen

Xiang Chen

Junqing Chen

Jun Wang

Zuoqiang Shi

  1. NSNO: Neumann Series Neural Operator for Solving Helmholtz Equations in Inhomogeneous Medium

    Chen, Fukai | Liu, Ziyang | Lin, Guochang | Chen, Junqing | Shi, Zuoqiang

    Journal of Systems Science and Complexity, Vol. 37 (2024), Iss. 2 P.413

    https://doi.org/10.1007/s11424-024-3294-x [Citations: 1]
  2. Boundary integrated neural networks and code for acoustic radiation and scattering

    Qu, Wenzhen | Gu, Yan | Zhao, Shengdong | Wang, Fajie | Lin, Ji

    International Journal of Mechanical System Dynamics, Vol. 4 (2024), Iss. 2 P.131

    https://doi.org/10.1002/msd2.12109 [Citations: 2]
  3. Boundary integrated neural networks for 2D elastostatic and piezoelectric problems

    Zhang, Peijun | Xie, Longtao | Gu, Yan | Qu, Wenzhen | Zhao, Shengdong | Zhang, Chuanzeng

    International Journal of Mechanical Sciences, Vol. 280 (2024), Iss. P.109525

    https://doi.org/10.1016/j.ijmecsci.2024.109525 [Citations: 3]
  4. Learning domain-independent Green’s function for elliptic partial differential equations

    Negi, Pawan | Cheng, Maggie | Krishnamurthy, Mahesh | Ying, Wenjun | Li, Shuwang

    Computer Methods in Applied Mechanics and Engineering, Vol. 421 (2024), Iss. P.116779

    https://doi.org/10.1016/j.cma.2024.116779 [Citations: 0]