An analysis of the order of Runge-Kutta methods that use an iterative scheme to compute their internal stage values (Q2565272)

This paper uses \(B\)-series to study the order of implicit Runge-Kutta methods in which the internal stage approximations are solved by a variety of iterative techniques such as fixed point, modified Newton or full Newton iteration. The basic result is that if the initial predictor is of order \(q\) and the underlying method has order \(p\) with stage order \(w\), then after \(k\) iterations the order is \(\min \{p,q+k\}\), \(\min \{p,q+2k-1\}\), \(\min \{p,2^{k-1} (r+2) + q-r-1\}\) with \(r= \min \{q,w\}\) in these three cases. Some numerical results illustrate these theorems.