The Random Feature Method for Solving Interface Problems

Xurong Chi,Jingrun Chen,Zhouwang Yang

2023-08-08

Abstract:Interface problems have long been a major focus of scientific computing, leading to the development of various numerical methods. Traditional mesh-based methods often employ time-consuming body-fitted meshes with standard discretization schemes or unfitted meshes with tailored schemes to achieve controllable accuracy and convergence rate. Along another line, mesh-free methods bypass mesh generation but lack robustness in terms of convergence and accuracy due to the low regularity of solutions. In this study, we propose a novel method for solving interface problems within the framework of the random feature method. This approach utilizes random feature functions in conjunction with a partition of unity as approximation functions. It evaluates partial differential equations, boundary conditions, and interface conditions on collocation points in equal footing, and solves a linear least-squares system to obtain the approximate solution. To address the issue of low regularity, two sets of random feature functions are used to approximate the solution on each side of the interface, which are then coupled together via interface conditions. We validate our method through a series of increasingly complex numerical examples. Our findings show that despite the solution often being only continuous or even discontinuous, our method not only eliminates the need for mesh generation but also maintains high accuracy, akin to the spectral collocation method for smooth solutions. Remarkably, for the same accuracy requirement, our method requires two to three orders of magnitude fewer degrees of freedom than traditional methods, demonstrating its significant potential for solving interface problems with complex geometries.

Numerical Analysis,Computational Physics

What problem does this paper attempt to address?

The paper primarily proposes a new solution for interface problems by utilizing the Random Feature Method (RFM) to solve such issues. Interface problems are a key research area in scientific computing, involving the interfaces between different materials or fluids. Mathematically, they are usually represented by partial differential equations that have different forms in different regions and are coupled together through interface conditions.

Traditional methods are divided into two main categories: grid-based methods and mesh-free methods. Grid-based methods require the construction of grids that adapt to complex geometries, which is often time-consuming. On the other hand, mesh-free methods avoid the problem of grid generation but have issues with convergence and accuracy.

The RFM method proposed in the paper is a mesh-free method. Its core idea is to use random feature functions combined with Partition of Unity (PoU) functions as approximation functions. This method evaluates the partial differential equations, boundary conditions, and interface conditions at each sampling point simultaneously and obtains an approximate solution by solving a linear least squares problem. To handle low regularity problems, the method uses two sets of random feature functions on either side of the interface, coupling them through the interface conditions.

The paper validates the effectiveness of this method through a series of complex numerical experiments, including 2D elliptic interface problems, 3D Stokes interface problems, 3D elasticity interface problems, moving interface problems, and fluid-structure interaction problems. The results show that although the solutions may be merely continuous or even discontinuous, RFM not only eliminates the need for grid generation but also maintains high accuracy, requiring two to three orders of magnitude fewer degrees of freedom than traditional methods for the same accuracy requirements.

In summary, the paper aims to address interface problems with complex geometries and proposes an efficient, accurate, and grid-free new method—the Random Feature Method (RFM). Through extensive numerical experiments, the advantages of this method in handling such problems are demonstrated.