Many-Stage Optimal Stabilized Runge-Kutta Methods for Hyperbolic Partial Differential Equations

Daniel Doehring,Gregor J. Gassner,Manuel Torrilhon

DOI: https://doi.org/10.1007/s10915-024-02478-5

2024-03-18

Abstract:A novel optimization procedure for the generation of stability polynomials of stabilized explicit Runge-Kutta methods is devised. Intended for semidiscretizations of hyperbolic partial differential equations, the herein developed approach allows the optimization of stability polynomials with more than hundred stages. A potential application of these high degree stability polynomials are problems with locally varying characteristic speeds as found in non-uniformly refined meshes and different wave speeds.

Numerical Analysis,Mathematical Physics,Classical Analysis and ODEs

What problem does this paper attempt to address?

### Problems the paper attempts to solve

The paper aims to develop a new optimization procedure for generating stability polynomials of stable explicit Runge - Kutta methods. Specifically, the goals of the paper are as follows:

1. **Improve stability**: Improve the stability characteristics of the scheme by adding additional stages, thus allowing for the use of larger time steps. This is particularly useful for dealing with problems with locally varying characteristic velocities, such as in the case of non - uniform grid refinement and different wave speeds.

2. **Optimize high - order stability polynomials**: The developed method can optimize high - order stability polynomials with more than 100 stages. These polynomials can be applied to a series of linear and nonlinear hyperbolic partial differential equations (PDEs) and can maintain the designed second - order accuracy when applied.

3. **Reduce rounding errors**: The constructed multi - stage Runge - Kutta method aims to reduce the amplification of rounding errors, which is an important concern in multi - stage methods.

4. **Adapt to different spectral types**: The paper is not only optimized for circular spectra but also discusses how to extend to non - convex spectra, enabling it to be more widely applied to different physical problems.

### Key technical means

- **Pseudo - extreme value parameterization**: Avoid the inherent ill - conditioned problem of polynomial coefficients by parameterizing the stability polynomial as its roots (i.e., pseudo - extreme values), thereby transforming the optimization problem into a nonlinear problem.
- **Initial guess**: Based on the results of known optimal stability polynomials, an excellent initial guess method is proposed, enabling the successful solution of the optimization problem.
- **Constraints**: For high - order consistent methods, necessary equality constraints are added to ensure the accuracy of the method.

### Potential applications

- **Local time stepping**: The optimized high - order stability polynomials can be applied to the paired explicit Runge - Kutta method (P - ERK) to achieve local time - stepping effects, thus more efficiently handling locally refined grids or varying characteristic velocities.
- **Multi - rate partitioned Runge - Kutta methods**: These methods can more effectively handle complex physical systems while maintaining consistency and conservation.

In summary, through the development of new optimization methods, the paper significantly improves the stability and efficiency of explicit Runge - Kutta methods in dealing with hyperbolic partial differential equations.