Maximal order for second derivative general linear methods with Runge-Kutta stability (Q719432)

The authors recall the class of second derivative general linear methods for the numerical solution of classical initial value problems for first order differential equations from the work of \textit{J. C. Butcher} and \textit{G. Hojjati} [Numer. Algorithms 40, No.~4, 415--429 (2005; Zbl 1084.65069)]. These methods are classified into one of four categories, depending on the stiffness of the given problem and on whether the algorithms are to be implemented on a sequential or a parallel computer. For one of these types (appropriate for stiff equations on sequential hardware), the authors determine the maximal order of the method under the assumption of Runge-Kutta stability. A similar result is then proved for a class of methods that generalizes the type useful for stiff problems in a parallel computing environment. Some special cases are explicitly constructed.