Abstract:Recent years have witnessed growing interests in solving partial differential equations by deep neural networks, especially in the high-dimensional case. Unlike classical numerical methods, such as finite difference method and finite element method, the enforcement of boundary conditions in deep neural networks is highly nontrivial. One general strategy is to use the penalty method. In the work, we conduct a comparison study for elliptic problems with four different boundary conditions, i.e., Dirichlet, Neumann, Robin, and periodic boundary conditions, using two representative methods: deep Galerkin method and deep Ritz method. In the former, the PDE residual is minimized in the least-squares sense while the corresponding variational problem is minimized in the latter. Therefore, it is reasonably expected that deep Galerkin method works better for smooth solutions while deep Ritz method works better for low-regularity solutions. However, by a number of examples, we observe that deep Ritz method can outperform deep Galerkin method with a clear dependence of dimensionality even for smooth solutions and deep Galerkin method can also outperform deep Ritz method for low-regularity solutions. Besides, in some cases, when the boundary condition can be implemented in an exact manner, we find that such a strategy not only provides a better approximate solution but also facilitates the training process.

Uniform convergence guarantees for the deep Ritz method for nonlinear problems

Deep Ritz Method with Adaptive Quadrature for Linear Elasticity

Convergence of the Dirichlet-Neumann method for semilinear elliptic equations

Deep Ritz Methods for Laplace Equations with Dirichlet Boundary Condition

The Deep Ritz Method: A Deep Learning-Based Numerical Algorithm for Solving Variational Problems

Convergence Analysis of a Method for Variational Inclusions

Convergence Rate Analysis for Deep Ritz Method

A Deep Uzawa-Lagrange Multiplier Approach for Boundary Conditions in PINNs and Deep Ritz Methods

The Deep Ritz Method for Parametric $p$-Dirichlet Problems

Rothe Method for Parabolic Variational–hemivariational Inequalities

The repulsion property in nonlinear elasticity and a numerical scheme to circumvent it

Uniform Convergence of Global Least Energy Solutions to Dirichlet Systems in Non-reflexive Orlicz–Sobolev Spaces

A comparison study of deep Galerkin method and deep Ritz method for elliptic problems with different boundary conditions

Finite Element Approximation of Hamilton-Jacobi-Bellman equations with nonlinear mixed boundary conditions

On inertial Levenberg-Marquardt type methods for solving nonlinear ill-posed operator equations

Weak Solutions and Energy Estimates for a Class of Nonlinear Elliptic Neumann Problems

Convergence Analysis for Bregman Iterations in Minimizing a Class of Landau Free Energy Functionals

Error Analysis of Deep Ritz Methods for Elliptic Equations.

Convergence analysis of a numerical scheme for the porous medium equation by an energetic variational approach

Monte Carlo method and the random isentropic Euler system