Abstract:In this paper, for solving a class of linear parabolic equations in rectangular domains, we have proposed an efficient Parareal exponential integrator finite element method. The proposed method first uses the finite element approximation with continuous multilinear rectangular basis function for spatial discretization, and then takes the Runge-Kutta approach accompanied with Parareal framework for time integration of the resulting semi-discrete system to produce parallel-in-time numerical solution. Under certain regularity assumptions, fully-discrete error estimates in $L^2$-norm are derived for the proposed schemes with random interpolation nodes. Moreover, a fast solver can be provided based on tensor product spectral decomposition and fast Fourier transform (FFT), since the mass and coefficient matrices of the proposed method can be simultaneously diagonalized with an orthogonal matrix. A series of numerical experiments in various dimensions are also presented to validate the theoretical results and demonstrate the excellent performance of the proposed method.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to develop an efficient method for solving linear parabolic equations in parallel. Specifically, the author proposes an exponential integration finite element method (PEIFE - linear) based on the Parareal framework for solving a class of linear parabolic equations in rectangular domains. This type of equations is widely used in scientific and engineering models.
### Problem Description
Consider the linear parabolic equation in the following form:
\[
\begin{cases}
u_t = D\Delta u + f(t), & x\in\Omega, t_0\leq t\leq t_0 + T, \\
u(t_0):=u(t_0, x)=u_0, & x\in\Omega, \\
u(t, x) = 0, & x\in\partial\Omega, t_0\leq t\leq t_0 + T,
\end{cases}
\]
where:
- $\Omega$ is an open rectangular region in a $d$-dimensional space,
- $T>0$ is the time duration,
- $D>0$ is the diffusion coefficient,
- $f(t)$ is the reaction term.
### Solution
To solve the above - mentioned problem, the paper proposes the following methods:
1. **Spatial Discretization**: Use continuous multilinear rectangular element basis functions for finite - element approximation.
2. **Temporal Discretization**: Combine the Runge - Kutta method and the Parareal framework to perform time integration on the semi - discrete system, thus generating a numerical solution parallel in time.
3. **Fast Solver**: Utilize tensor - product spectral decomposition and fast Fourier transform (FFT) techniques to diagonalize the mass matrix and the coefficient matrix simultaneously in order to improve computational efficiency.
### Main Contributions
1. **Efficient Parallel Computation**: Achieve time - parallel computation through the Parareal framework, significantly improving computational efficiency.
2. **Global Error Estimation**: Derive a fully - discrete error estimate in the $L^2$ norm and prove the unconditional stability of the method.
3. **Fast Implementation**: Use FFT and tensor - product operations to achieve efficient matrix - vector multiplication, which is suitable for Dirichlet and periodic boundary conditions.
### Conclusion
This method not only has a strict global error estimate in theory but also performs excellently in actual numerical experiments and can effectively solve large - scale linear parabolic equation problems.