Strong stability preserving explicit Runge-Kutta methods of maximal effective order

Yiannis Hadjimichael,Colin B. Macdonald,David I. Ketcheson,James H. Verner

DOI: https://doi.org/10.1137/120884201

2013-03-11

Abstract:We apply the concept of effective order to strong stability preserving (SSP) explicit Runge-Kutta methods. Relative to classical Runge-Kutta methods, methods with an effective order of accuracy are designed to satisfy a relaxed set of order conditions, but yield higher order accuracy when composed with special starting and stopping methods. We show that this allows the construction of four-stage SSP methods with effective order four (such methods cannot have classical order four). However, we also prove that effective order five methods - like classical order five methods - require the use of non-positive weights and so cannot be SSP. By numerical optimization, we construct explicit SSP Runge-Kutta methods up to effective order four and establish the optimality of many of them. Numerical experiments demonstrate the validity of these methods in practice.

Numerical Analysis

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper mainly explores the design and optimization of Strong Stability Preserving (SSP) explicit Runge - Kutta (RK) methods with the maximum effective order. Specifically, the authors attempt to answer the following key questions:

1. **Can SSP Runge - Kutta methods with an effective order greater than four be constructed?**
- The paper proves that methods with an effective order of five cannot be achieved because they require the use of non - positive weights, which violates the requirements of SSP methods.

2. **If only the effective order of the method is required to be \( q \), can a larger SSP coefficient be achieved than that of the method with a classical order of \( q \)?**
- Through numerical optimization, the authors found some optimal SSP methods and verified the effectiveness of these methods.

3. **Can SSP methods with an effective order of four have fewer stages?**
- The authors successfully constructed SSP methods with an effective order of four and four stages, while classical fourth - order methods require at least five stages.

### Main contributions of the paper

- **Theoretical analysis**: Proved that SSP methods with an effective order of five cannot exist.
- **Numerical optimization**: Constructed multiple optimal SSP Runge - Kutta methods with effective orders of three and four through numerical optimization techniques.
- **Practical application**: Verified the effectiveness and superiority of the constructed methods through numerical experiments.

### Key formulas

- **Effective order condition**:
\[
(βαβ^{-1})(t)=E(t), \quad \text{for all trees satisfying } r(t)\leq q
\]
where \( β \) and \( β^{-1} \) are the elements corresponding to the starting and ending methods respectively, \( α \) is the element corresponding to the main method, and \( E \) is the exact evolution operator.

- **Convex combination and Jensen's inequality**:
\[
b^T v^2\leq(b^T v)^2
\]
here \( v = \frac{1}{2} c^2 - Ac \), and the equality holds if and only if all components of \( v \) are equal.

### Summary

By introducing the concept of effective order, this paper expands the design space of traditional SSP Runge - Kutta methods, overcomes some limitations in constructing high - order SSP methods, and verifies the effectiveness of the new methods through numerical optimization and experiments.