Table of Contents
Fetching ...

Schwarz Methods for Nonlocal Problems

Matthias Schuster, Christian Vollmann, Volker Schulz

Abstract

The first domain decomposition methods for partial differential equations were already developed in 1870 by H. A. Schwarz. Here we consider a nonlocal Dirichlet problem with variable coefficients, where a nonlocal diffusion operator is used. We find that domain decomposition methods like the so-called Schwarz methods seem to be a natural way to solve these nonlocal problems. In this work we show the convergence for nonlocal problems, where specific symmetric kernels are employed, and present the implementation of the multiplicative and additive Schwarz algorithms in the above mentioned nonlocal setting.

Schwarz Methods for Nonlocal Problems

Abstract

The first domain decomposition methods for partial differential equations were already developed in 1870 by H. A. Schwarz. Here we consider a nonlocal Dirichlet problem with variable coefficients, where a nonlocal diffusion operator is used. We find that domain decomposition methods like the so-called Schwarz methods seem to be a natural way to solve these nonlocal problems. In this work we show the convergence for nonlocal problems, where specific symmetric kernels are employed, and present the implementation of the multiplicative and additive Schwarz algorithms in the above mentioned nonlocal setting.
Paper Structure (21 sections, 11 theorems, 86 equations, 8 figures, 3 tables, 5 algorithms)

This paper contains 21 sections, 11 theorems, 86 equations, 8 figures, 3 tables, 5 algorithms.

Key Result

Theorem 2.4

The space $\mathcal{V}_c(\Omega, \mathcal{I}^N, \mathcal{I}^D)$ is complete regarding $||\cdot||_{\mathcal{V}(\Omega, \mathcal{I}^N, \mathcal{I}^D)}$. As a consequence, $V_c(\Omega, \mathcal{I}^N, \mathcal{I}^D)$ is a Hilbert space with respect to $||\cdot||_{V(\Omega, \mathcal{I}^N, \mathcal{I}^D)}

Figures (8)

  • Figure 7.1: The domain $\Omega = (0,1)^2$ is divided into $\Omega_1$, $\Omega_2$ and $\Omega_3$. The nonlocal boundary $\mathcal{I}$ is depicted in red.
  • Figure 7.2: Here, we can see the residual error of the multiplicative and additive Schwarz method for the nonlocal Dirichlet problem with a singular symmetric kernel regarding different choices for the mesh resolution $h$. In all cases the error decreases in a linear fashion, which we expected at least for the multiplicative Schwarz version due to Remark \ref{['remark:conv_discrete_schwarz']}. Moreover, the additive version needs roughly twice as many iterations as the multiplicative Schwarz algorithm.
  • Figure 7.3: In this example $\Omega$ is divided in two subdomains $\Omega_1$ and $\Omega_2$ and the nonlocal boundary $\mathcal{I}$ is decomposed in $\mathcal{I}_1^N$ and $\mathcal{I}_2^N$.
  • Figure 7.4: In this picture the residual error regarding the multiplicative Schwarz method for the nonlocal Problem with Neumann boundary condition w.r.t. a selection of mesh parameter $h$ is depicted. Again, we can observe a linear decrease in the residual error.
  • Figure 7.5: Here, we see the decomposition of $\Omega$ in a turquoise area $\Omega_1$ and in a gray domain $\Omega_2$ that we use for the patch test and for testing the (preconditioned) GMRES in Chapter \ref{['chap:preconditioned_gmres']}. The nonlocal boundary $\mathcal{I}$ is again depicted in red.
  • ...and 3 more figures

Theorems & Definitions (30)

  • Definition 2.1
  • Remark 2.2
  • Definition 2.3
  • Theorem 2.4
  • proof
  • Corollary 2.5
  • Remark 3.1
  • Remark 3.2
  • Theorem 4.1
  • Remark 4.2
  • ...and 20 more