Table of Contents
Fetching ...

An efficient multigrid solver for 3D biharmonic equation with a discretization by 25-point difference scheme

Kejia Pan, Dongdong He, Runxin Ni

TL;DR

The paper tackles solving the 3D biharmonic equation $\Delta^2 u=f$ on $\Omega=[0,1]^3$ with Dirichlet boundary conditions, a challenging large-scale non-symmetric linear system arising from a 25-point FD discretization. It introduces the EXCMG_bi-cg framework, which uses Richardson extrapolation and quadratic interpolation to generate a third-order accurate initial guess on progressively finer grids and solves the resulting systems with Bi-CG under a relative residual tolerance. The main contributions are the EXCMG algorithm, its 3D extension with extrapolation on embedded meshes, and extensive numerical demonstrations up to about $135$ million unknowns showing dramatic reductions in iterations and total work. The approach provides a scalable, high-accuracy solver for complex 3D biharmonic problems with potential applications in elasticity and related fields, leveraging explicit multigrid extrapolation techniques with a robust iterative solver.

Abstract

In this paper, we propose an efficient extrapolation cascadic multigrid (EXCMG) method combined with 25-point difference approximation to solve the three-dimensional biharmonic equation. First, through applying Richardson extrapolation and quadratic interpolation on numerical solutions on current and previous grids, a third-order approximation to the finite difference solution can be obtained and used as the iterative initial guess on the next finer grid. Then we adopt the bi-conjugate gradient (Bi-CG) method to solve the large linear system resulting from the 25-point difference approximation. In addition, an extrapolation method based on midpoint extrapolation formula is used to achieve higher-order accuracy on the entire finest grid. Finally, some numerical experiments are performed to show that the EXCMG method is an efficient solver for the 3D biharmonic equation.

An efficient multigrid solver for 3D biharmonic equation with a discretization by 25-point difference scheme

TL;DR

The paper tackles solving the 3D biharmonic equation on with Dirichlet boundary conditions, a challenging large-scale non-symmetric linear system arising from a 25-point FD discretization. It introduces the EXCMG_bi-cg framework, which uses Richardson extrapolation and quadratic interpolation to generate a third-order accurate initial guess on progressively finer grids and solves the resulting systems with Bi-CG under a relative residual tolerance. The main contributions are the EXCMG algorithm, its 3D extension with extrapolation on embedded meshes, and extensive numerical demonstrations up to about million unknowns showing dramatic reductions in iterations and total work. The approach provides a scalable, high-accuracy solver for complex 3D biharmonic problems with potential applications in elasticity and related fields, leveraging explicit multigrid extrapolation techniques with a robust iterative solver.

Abstract

In this paper, we propose an efficient extrapolation cascadic multigrid (EXCMG) method combined with 25-point difference approximation to solve the three-dimensional biharmonic equation. First, through applying Richardson extrapolation and quadratic interpolation on numerical solutions on current and previous grids, a third-order approximation to the finite difference solution can be obtained and used as the iterative initial guess on the next finer grid. Then we adopt the bi-conjugate gradient (Bi-CG) method to solve the large linear system resulting from the 25-point difference approximation. In addition, an extrapolation method based on midpoint extrapolation formula is used to achieve higher-order accuracy on the entire finest grid. Finally, some numerical experiments are performed to show that the EXCMG method is an efficient solver for the 3D biharmonic equation.

Paper Structure

This paper contains 12 sections, 41 equations, 3 figures, 5 tables, 2 algorithms.

Figures (3)

  • Figure 1: The four level structure of the V- and W-cycles, CMG and EXCMG methods. In the diagram, $\bullet$ denotes pre-smoothing steps, $\circ$ denotes post-smoothing steps, $\uparrow$ denotes prolongation, $\downarrow$ denotes restriction, $\Uparrow$ denotes extrapolation and quadratic interpolation, and $\blacksquare$ denotes direct solver.
  • Figure 2: Three embedded mesh in 1D.
  • Figure 3: Three embedded hexahedral mesh

Theorems & Definitions (1)

  • Remark 1