Numerical smoothing of Runge-Kutta schemes

Abstract

First we give an intuitive explanation of the general idea of Sun (2005) [1]: consistency and numerical smoothing implies convergence and, in addition, enables error estimates. Then, we briefly discuss some of the advantages of numerical smoothing over numerical stability in error analysis. The main aim of this paper is to introduce a smoothing function and use it to investigate the smoothing properties of some familiar schemes. © 2009 Elsevier B.V. All rights reserved.