H. De Sterck,R. D. Falgout,O. A. Krzysik,J. B. Schroder
Abstract:We consider the parallel-in-time solution of scalar nonlinear conservation laws in one spatial dimension. The equations are discretized in space with a conservative finite-volume method using weighted essentially non-oscillatory (WENO) reconstructions, and in time with high-order explicit Runge-Kutta methods. The solution of the global, discretized space-time problem is sought via a nonlinear iteration that uses a novel linearization strategy in cases of non-differentiable equations. Under certain choices of discretization and algorithmic parameters, the nonlinear iteration coincides with Newton's method, although, more generally, it is a preconditioned residual correction scheme. At each nonlinear iteration, the linearized problem takes the form of a certain discretization of a linear conservation law over the space-time domain in question. An approximate parallel-in-time solution of the linearized problem is computed with a single multigrid reduction-in-time (MGRIT) iteration. The MGRIT iteration employs a novel coarse-grid operator that is a modified conservative semi-Lagrangian discretization and generalizes those we have developed previously for non-conservative scalar linear hyperbolic problems. Numerical tests are performed for the inviscid Burgers and Buckley--Leverett equations. For many test problems, the solver converges in just a handful of iterations with convergence rate independent of mesh resolution, including problems with (interacting) shocks and rarefactions.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to solve one - dimensional nonlinear conservation laws in parallel in time. Specifically, the paper focuses on how to efficiently solve these equations through multilevel - in - time methods, especially the multigrid reduction - in - time (MGRIT). Traditional parallel - in - time methods usually perform poorly when dealing with convection - dominated partial differential equations (PDEs), especially when dealing with nonlinear problems containing shock waves and rarefaction waves. Therefore, the paper proposes a new linearization strategy and combines it with the MGRIT method to improve the robustness and efficiency of these methods in solving nonlinear hyperbolic PDEs.
### Main problems
1. **Parallel - in - time solution of nonlinear hyperbolic PDEs**: Traditional parallel - in - time methods (such as MGRIT and Parareal) have slow convergence and poor robustness when dealing with convection - dominated PDEs, especially when dealing with nonlinear problems containing shock waves and rarefaction waves. The paper aims to develop a new method that can show better performance on these problems.
2. **Linearization strategy**: In order to overcome the limitations of traditional methods, the paper proposes a new linearization strategy. This strategy can not only handle non - differentiable equations, but also accurately propagate information on characteristic curves on coarse grids, thereby improving the convergence speed and robustness of the solver.
3. **Design of coarse - grid operators**: The paper designs a new coarse - grid operator, which is based on the modified conservative semi - Lagrangian discretization method. This operator can effectively approximate the solution of the fine - grid problem on the coarse grid, thereby accelerating the overall solution process.
### Solutions
- **Global linearization**: The paper proposes a global linearization method, which transforms nonlinear problems into a series of linear problems. These linear problems are solved by the MGRIT method, thereby achieving parallel computation in time.
- **Modified coarse - grid operator**: The paper designs a new coarse - grid operator, which is based on the modified semi - Lagrangian discretization method. This operator can accurately propagate information on characteristic curves on the coarse grid, thereby improving the convergence speed and robustness of the solver.
- **Numerical experiments**: The paper verifies the effectiveness of the proposed method through numerical experiments. The experimental results show that when solving the Burgers equation and the Buckley - Leverett equation, this method can reach convergence within a small number of iterations, and the convergence rate is independent of the grid resolution.
### Formula summary
- **Nonlinear conservation law equations**:
\[
\frac{\partial u}{\partial t}+\frac{\partial f(u)}{\partial x} = 0, \quad(x, t)\in(- 1,1)\times(0, T], \quad u(x,0)=u_0(x)
\]
- **Linearization problems**:
\[
\frac{\partial e}{\partial t}+\frac{\partial(\alpha(x, t)e)}{\partial x} = 0, \quad(x, t)\in(- 1,1)\times(0, T], \quad e(x,0)=e_0(x)
\]
- **Lax - Friedrichs flux**:
\[
\hat{f}_{i + 1/2}=\frac{1}{2}\left(f(u^-_{i + 1/2})+f(u^+_{i + 1/2})\right)+\nu_{i + 1/2}\left(u^-_{i + 1/2}-u^+_{i + 1/2}\right)
\]
- **Linearized flux**:
\[
\hat{f}_{\text{lin}}(\alpha^-_{i + 1/2},e^-_{i + 1/2},\alpha^+_{i + 1/2},