Abstract:Runge-Kutta methods are a popular class of numerical methods for solving ordinary differential equations. Every Runge-Kutta method is characterized by two basic parameters: its order, which measures the accuracy of the solution it produces, and its number of stages, which measures the amount of computational work it requires. The primary goal in constructing Runge-Kutta methods is to maximize order using a minimum number of stages. However, high-order Runge-Kutta methods are difficult to construct because their parameters must satisfy an exponentially large system of polynomial equations. This paper presents the first known 10th-order Runge-Kutta method with only 16 stages, breaking a 40-year standing record for the number of stages required to achieve 10th-order accuracy. It also discusses the tools and techniques that enabled the discovery of this method using a straightforward numerical search.

What problem does this paper attempt to address?

This paper aims to solve the problem of constructing high - order Runge - Kutta methods. Specifically, the author's goal is to achieve high - order accuracy with as few stages as possible. The Runge - Kutta method is a class of numerical methods for solving ordinary differential equations (ODEs), and its performance is mainly determined by two parameters: the order, that is, the accuracy of the solution produced by the method; and the number of stages, that is, the computational effort required by the method. The construction of high - order Runge - Kutta methods is very difficult because its parameters must satisfy a system of polynomial equations that grows exponentially (called the order conditions). ### Main contributions of the paper 1. **Discovery of a new 10 - order Runge - Kutta method**: - The author discovered the first 16 - stage 10 - order Runge - Kutta method, breaking the previous record that required 17 stages to achieve 10 - order accuracy. 2. **Numerical search method**: - The author described in detail how to discover this new method through an unstructured numerical search method. This method takes advantage of the capabilities of modern computer hardware and avoids complex manual analysis or simplifying assumptions. 3. **Tools and techniques**: - The paper discussed a variety of algorithmic tools and numerical techniques that make this search possible on modern computer hardware. These techniques include, but are not limited to, the BFGS algorithm, parallelization and data layout, dual - number arithmetic, extended - precision machine arithmetic, etc. ### Background and challenges - **Historical background**: - The Runge - Kutta method has been studied for more than 120 years and is widely used in many fields such as physics, astronomy, chemistry, biology, pharmacology, and economics. - The construction of high - order Runge - Kutta methods is very difficult because its parameters must satisfy a system of polynomial equations that grows exponentially. For example, in 1965, John Butcher proved the minimum number of stages required to achieve 7 - order accuracy, but it was not until 1985 that the corresponding proof for 8 - order accuracy was completed. - **Limitations of traditional methods**: - Traditional methods usually rely on "simplifying assumptions" to reduce the number of equations that need to be considered simultaneously. However, this method may miss some methods because it forms sufficient rather than necessary conditions. ### Methods and results - **Numerical search method**: - The author used a method to directly solve the complete order - condition system, avoiding traditional simplifying assumptions. This method takes advantage of the powerful computing power of modern computers and found a new 10 - order 16 - stage Runge - Kutta method through numerical optimization techniques. - **Result comparison**: - The author compared the newly discovered method with existing Runge - Kutta methods and demonstrated its advantages in terms of accuracy and computational efficiency. ### Conclusion - The author successfully discovered a new 10 - order 16 - stage Runge - Kutta method, breaking the previous record. This result is not only of great theoretical significance but also provides a more efficient numerical method for practical applications. In conclusion, through an innovative numerical search method, this paper successfully solved the key problems in the construction of high - order Runge - Kutta methods and made important contributions to the field of numerical analysis.

Discovering New Runge-Kutta Methods Using Unstructured Numerical Search

An explicit 16-stage Runge–Kutta method of order 10 discovered by numerical search

Comparison of Computational Accuracy for Three Five-Stage Implicit Runge-Kutta Methods

A variable order Runge-Kutta method for initial value problems with rapidly varying right-hand sides

Very High-Order A-stable Stiffly Accurate Diagonally Implicit Runge-Kutta Methods with Error Estimators

Runge–Kutta methods for linear ordinary differential equations

An introduction to “Almost Runge-Kutta” methods

Two-step Runge-Kutta Methods of Order over Five with Reduced Function Evaluations.

Runge-Kutta Methods of Dynamic Iteration for Index-2 Differential-Algebraic Equations

Explicit Runge Kutta Methods that Alleviate Order Reduction

Nullspaces yield new explicit Runge–Kutta pairs

Runge-Kutta methods: some historical notes

Development of a new Runge‐Kutta method and its economical implementation

Probabilistic ODE Solvers with Runge-Kutta Means

Two new families of fourth-order explicit exponential Runge--Kutta methods with four stages for first-order differential systems

A Class of Runge-Kutta Formulae of Order Three and Four with Reduced Evaluations of Function

Randomized low-rank Runge-Kutta methods

A Phase-Fitted Runge-Kutta-Nyström method for the Numerical Solution of Initial Value Problems with Oscillating Solutions

An efficient ODE-solving method based on heuristic and statistical computations: αII-(2 + 3)P method

The Numerical Solution of Differential-Algebraic Systems by Runge-Kutta Methods

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