Relationships among some classes of implicit Runge-Kutta methods and their stability functions (Q1092625)

Implicit Runge-Kutta (IRK) methods were introduced by \textit{J. C. Butcher} [Math. Comput. 18, 50-64 (1964; Zbl 0123.117)] for the numerical solution of initial value problems in ordinary differential equations. Furthermore Butcher established a theoretical framework that has lead to a deep understanding of the properties of IRK methods. However in the last years some modified implicit Runge-Kutta methods have been proposed by several authors to solve initial and boundary value problems. The aim of this long paper is to give a theoretical analysis for most of this modified RK methods that have been proposed recently. To do that the authors first introduce the so called parametrized implicit Runge Kutta (PIRK) methods given by the equations \[ y_{i+1}=y_ i+h\sum^{s}_{r=1}b_ rf(Y_ r),\quad Y_ r-(1-\gamma_ r)y_ i- \gamma_ ry_{i+1}-h\sum^{s}_{j=1}a_{rj}f(Y_ j)=0 \] where \(\gamma_ r\) are parameters and \(Y_ r\) are the approximations of the solution at the intermediate stages. In this context the authors generalize the algebraic theory developed for IRK methods to the PIRK methods. Thus they analyze within a theoretical framework the efficiency of several subclasses of PIRK and the relationships between them.