Consider linear systems of ordinary differential equations.;which contain two different time scales, a slow one and a fast one. In many applications one is only interested in computing the solutions which vary on the slow time scale. There are different methods to compute these smooth solutions. One is to choose the proper initial data. The other is to filter the numerical solution. In this thesis we make a thorough investigation of a large class of filtering techniques.;(DIAGRAM, TABLE OR GRAPHIC OMITTED...PLEASE SEE DAI). |