Semi-implicit extrapolation method


vseipel
07-01-2004, 03:22 AM
Hello !

I`m having a problem with the semi-implicit extrapolation method
for ODEs which is described in section 16.6.

I want to solve a set of 5 ODEs.
The recommended stepsize sequence is n = 2,6,10,14,...
The thing is that for some initial conditions (that are not
strange at all) the method does not lead to sensible results.

If I use the stepsize sequence n = 2,6,19,14 (which does not
seem reasonable) I do not run into that problem.

Actually, I thought the method was possible with almost any
stepsize sequence, just affecting the performance.

Did I forgot any aspect ?
Or is it not enough just to change the stepsize sequence in the
data instruction ( "data nseq /2,6,......" ) ?

I shall be grateful for your reply.

Thank you in advance.

Best wishes


Volker Seipel