Analysis of error growth via stability regions in numerical initial value problems (Q1415406)

The paper concerns the theoretical analysis of numerical methods for solving initial value problems (one-step methods including Runge-Kutta methods and linear multistep methods). The investigations are focussing on the solution of a system of linear ordinary differential equations with initial conditions. The purpose of this paper is to address the natural question of whether the stability estimates for some numerical methods can be improved. The authors show that some of these estimates can be substantially sharpened and extended (for linear multistep formulas), whereas others are best possible (for Runge-Kutta and other one-step methods). The tool used systematically by the authors is the so called ``Kreiss resolvent condition with respect to the stability region''.