Runge-Kutta Method
Runge-Kutta Method for Numerical Integration · Also known as: RK4, RK method, explicit Runge-Kutta
The Runge-Kutta Method is a family of explicit numerical techniques for solving ordinary differential equations (ODEs) developed independently by Carl Runge in 1895 and Martin Kutta in 1901. The fourth-order variant (RK4) is one of the most widely used algorithms in computational science and engineering for time-stepping problems.
Read the full method
Sign in with a free account to read this section.
When to use it
Use Runge-Kutta for non-stiff ODEs requiring moderate accuracy with reasonable computational cost. It is ideal for problems in physics, chemistry, and engineering where the solution is smooth and the time scale is well-defined. Avoid for stiff systems (where eigenvalues differ greatly) or when very high accuracy is required; consider implicit methods or higher-order schemes in those cases.
Strengths & limitations
- RK4 achieves O(h^4) accuracy with only four function evaluations per step, making it highly efficient
- Explicit formulation is easy to implement and requires no matrix inversion
- Stable over a wide range of time step sizes for non-stiff problems
- Adaptive step-size variants allow automatic error control
- Not suitable for stiff differential equations; implicit methods perform better
- Accuracy degrades rapidly as time step h increases; stability region is bounded
- Higher-order variants (RK5, RK7) offer only marginal gains for their added complexity
Frequently asked
Why does RK4 use those specific weights (1:2:2:1)?
The weights emerge from Taylor series analysis: matching the first four terms of the Taylor expansion of the exact solution requires these particular coefficients. They ensure the local truncation error is O(h^5), making the global error O(h^4).
When should I use adaptive step-size RK instead of fixed-step RK?
Use adaptive methods (Runge-Kutta-Fehlberg, Dormand-Prince) when the solution has regions of rapid change where small steps are needed alongside smooth regions where larger steps suffice. This saves computation while maintaining error control.
What is the stability region of RK4, and why does it matter?
RK4 is stable only for complex numbers z = λh within a specific region of the complex plane (roughly |λh| < 2.8). For stiff systems with large |λ|, even tiny h values may violate this, causing instability; implicit methods like backward Euler have larger stability regions.
How do I choose the time step h?
Start with h ≈ (desired_accuracy)^(1/4) times the characteristic time scale. Run two simulations with h and h/2; if results differ less than tolerance, h is adequate. Otherwise halve h and repeat.
Sources
- Runge, C. (1895). Ueber die numerische Auflösung von Differentialgleichungen. Mathematische Annalen, 46(2), 167–178. DOI: 10.1007/BF01446807 ↗
- Kutta, M. W. (1901). Beitrag zur näherungsweisen Integration totaler Differentialgleichungen. Zeitschrift für Mathematik und Physik, 46, 435–453. link ↗
- Butcher, J. C. (2008). Numerical Methods for Ordinary Differential Equations (2nd ed.). Wiley. DOI: 10.1002/9780470753767 ↗
How to cite this page
ScholarGate. (2026, June 3). Runge-Kutta Method for Numerical Integration. ScholarGate. https://scholargate.app/en/numerical-methods/runge-kutta-method