I have a problem about determining $a_1, a_2, k_1, k_2, \ldots a_n, k_n,\ldots$

In the general form of the Higher Order Runge-Kutta below :


For the convenient, i'll write it down The Runge Kutta $2^{\text{nd}}$ in my book just in case if the formula is different from your views:

$$ \begin{align} k_1&=hf(x_r,y_r)\\ k_2&=hf(x_r+p_1h,y_r+q_{11}k_1)\\ y_{r+1}&=y_r+(a_1k_1+a_2k_2) \end{align} $$

I'm not really sure, but some sources talk about Butcher Tableau, slope for each $k_n$, Taylor expansion, and rooted trees, for determining the $a_n$. And i don't really understand what their relation is, especially for Butcher Tableau, and rooted trees.

Could you explain all of these for me? Or just give me a good reference that discusses about derivation of the higher order Runge-Kutta method in detail, please?

Because, some books just skip over the derivation part.


A good source on Butcher tableaus and B-trees would, almost canonically, be the book by Butcher himself. For a short overview, see the three sets of slides with an introduction to B-trees, demonstration of the method for up to order 4 and outlook to implicit methods, or a historical overview

You could of course also read the most original source, the introduction of Runge-Kutta methods by M. Wilhelm Kutta (1901), "Beitrag zur näherungsweisen Lösung totaler Differentialgleichungen", where the order conditions are established and solved up to order 4, with an order 5 method that turned out not to be correct for systems.

  • $\begingroup$ Is this source containing how to derive the RK formula? $\endgroup$ – user516076 Nov 15 '19 at 3:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.