Abstract:In this paper, we develop two fast implicit difference schemes for solving a class of variable-coefficient time-space fractional diffusion equations with integral fractional Laplacian (IFL). The proposed schemes utilize the graded $L1$ formula for the Caputo fractional derivative and a special finite difference discretization for IFL, where the graded mesh can capture the model problem with a weak singularity at initial time. The stability and convergence are rigorously proved via the $M$-matrix analysis, which is from the spatial discretized matrix of IFL. Moreover, the proposed schemes use the fast sum-of-exponential approximation and Toeplitz matrix algorithms to reduce the computational cost for the nonlocal property of time and space fractional derivatives, respectively. The fast schemes greatly reduce the computational work of solving the discretized linear systems from $\mathcal{O}(MN^3 + M^2N)$ by a direct solver to $\mathcal{O}(MN(\log N + N_{exp}))$ per preconditioned Krylov subspace iteration and a memory requirement from $O(MN^2)$ to $O(NN_{exp})$, where $N$ and $(N_{exp} \ll)~M$ are the number of spatial and temporal grid nodes. The spectrum of preconditioned matrix is also given for ensuring the acceleration benefit of circulant preconditioners. Finally, numerical results are presented to show the utility of the proposed methods.
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper aims to solve the numerical solution problem of a class of time - space fractional - order diffusion equations (TSFDEs) with integral fractional - order Laplace operators (IFL). Specifically, the paper proposes two fast implicit difference schemes (IDSs) for solving time - space fractional - order diffusion equations with variable coefficients, spatio - temporal fractional - order derivatives and integral fractional - order Laplace operators.
#### Main problem descriptions
1. **Model complexity**: Traditional numerical methods have limitations in dealing with high - dimensional time - space fractional - order diffusion equations, especially when involving integral fractional - order Laplace operators, due to their non - locality and high computational complexity.
2. **Initial singularity**: The kernel function $(t - s)^{-\gamma}$ of the Caputo fractional - order derivative causes a weak singularity of the solution of the equation at the initial time, making approximation methods (such as the L1 formula) on a uniform grid converge slowly and have high computational costs.
3. **Computational efficiency**: Traditional methods (such as direct solvers) for directly solving discrete linear systems have a large amount of computation and high memory requirements, and are difficult to be applied to large - scale problems.
### Solutions
To overcome the above problems, the paper proposes the following solutions:
1. **Fast implicit difference scheme (FIDS)**:
- Use the graded L1 formula to approximate the Caputo fractional - order derivative, and combine special finite - difference discretization of the integral fractional - order Laplace operator to capture the initial singularity.
- Reduce the computational cost through the sum - of - exponentials approximation (SOE) and Toeplitz matrix algorithm, reducing the computational complexity from $O(MN^{3}+M^{2}N)$ to $O(MN(\log N + N_{\text{exp}}))$, and the memory requirement from $O(MN^{2})$ to $O(NN_{\text{exp}})$.
2. **Stability and convergence analysis**:
- Strictly prove the stability and convergence of the proposed fast implicit difference scheme through M - matrix analysis, ensuring the reliability and accuracy of the numerical method.
- Provide detailed error estimates, proving that under certain regularity assumptions, the error of the numerical solution is $O(M^{-\min\{r\gamma, 2-\gamma\}}+h^{2})$.
3. **Efficient implementation**:
- Use the preconditioned Krylov subspace iteration method and the circulant preconditioner to accelerate the solution of the discrete linear system, further improving the computational efficiency.
- Theoretically ensure the effectiveness of the circulant preconditioner, and verify its acceleration effect through numerical experiments.
### Summary
By proposing the fast implicit difference scheme and its efficient implementation method, the paper successfully solves the numerical solution problem of the diffusion equation with spatio - temporal fractional - order derivatives and integral fractional - order Laplace operators, significantly improving the computational efficiency and ensuring the stability and convergence of the numerical method.