Abstract:Can graded meshes yield more accurate numerical solution than uniform meshes? A time-dependent nonlocal diffusion problem with a weakly singular kernel is considered using collocation method. For its steady-state counterpart, under the sufficiently smooth solution, we first clarify that the standard graded meshes are worse than uniform meshes and may even lead to divergence; instead, an optimal convergence rate arises in so-called anomalous graded meshes. Furthermore, under low regularity solutions, it may suffer from a severe order reduction in (Chen, Qi, Shi and Wu, IMA J. Numer. Anal., 41 (2021) 3145--3174). In this case, conversely, a sharp error estimates appears in standard graded meshes, but offering far less than first-order accuracy. For the time-dependent case, however, second-order convergence can be achieved on graded meshes. The related analysis are easily extended for certain multidimensional problems. Numerical results are provided that confirm the sharpness of the error estimates.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to explore whether graded meshes can provide more accurate numerical solutions than uniform meshes when solving time - dependent non - local diffusion problems with weakly singular kernels. Specifically, the paper focuses on the following points:
1. **For steady - state problems**: When the solution is smooth enough, the performance of standard graded meshes is not as good as that of uniform meshes and may even lead to divergence; on the contrary, the optimal convergence rate can be obtained in the so - called anomalous graded meshes. However, when the regularity of the solution is low, the standard graded meshes can provide a sharp error estimate, although its accuracy is far below the first - order.
2. **For time - dependent problems**: In the case of low - regularity solutions, second - order convergence can be achieved on graded meshes.
To answer these questions, the author verifies the influence of different types of meshes on non - local diffusion problems through detailed theoretical analysis and numerical experiments, and provides specific error estimation formulas. The following is a summary of several key conclusions:
- For steady - state problems (Equation 1.2), if the solution is smooth enough:
- Standard graded meshes (\(r\geq1\)) perform worse than uniform meshes and may even lead to divergence.
- Anomalous graded meshes (\(0 < r < 1\)) can obtain the optimal convergence rate, and the convergence rate is:
\[
\max_{1\leq i\leq2N - 1}|u(x_i)-u_i|\leq
\begin{cases}
CN^{r - 2}, &\text{if }r > \frac{2}{3},\\
CN^{r - 2}\ln N, &\text{if }r=\frac{2}{3},\\
CN^{- 2r}, &\text{if }0 < r < \frac{2}{3}.
\end{cases}
\]
- For low - regularity solutions (Equation 1.4), standard graded meshes can provide a sharp error estimate, but the accuracy is far below the first - order:
\[
\max_{1\leq i\leq2N - 1}|u(x_i)-u_i|\leq
\begin{cases}
CN^{r - 2}, &\text{if }r(1+\sigma)>2,\\
CN^{r - 2}\ln N, &\text{if }r(1+\sigma)=2,\\
CN^{-r\sigma}, &\text{if }r(1+\sigma)<2.
\end{cases}
\]
- For time - dependent problems (Equation 1.1), in the case of low - regularity solutions, second - order convergence can be achieved on graded meshes:
\[
\max_{1\leq i\leq2N - 1,1\leq k\leq M}|u(x_i,t_k)-u_i^k|\leq
\begin{cases}
C\left(N^{-r(1+\sigma-\alpha)}+M^{-2}\right), &\text{if }r(1+\sigma-\alpha)<2,\\
C\left(N^{-r(1+\sigma-\alpha)}\ln N+M^{-2}\right), &\text{if }r(1+\sigma-\alpha)=2,\\
C\left(N^{-2}+M^{-2}\right), &\text{if }r(1+\sigma-\alpha)>2.
\end{cases}
\]
These results indicate that choosing the appropriate mesh type is crucial for improving the accuracy of numerical solutions to non - local diffusion problems.