[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Help-gsl] Odeiv bsimp?

From: Alexei Halpin
Subject: [Help-gsl] Odeiv bsimp?
Date: Tue, 5 Jul 2005 14:44:41 -0400

I have been using the gsl ode solving routines for a few months now,
and recently have tried to use the bsimp stepping routine to solve a
system of stiff coupled ODEs in a thermal physics context.
I have been running into difficulty time and time again with this
stepper, and was wondering if anyone has anyone advice pertaining to
its general use. It seems as though before taking a timestep forward
my solver gets confused and tries negative temperature solutions
before giving up and trying a trial solution of "nan". Changing the
precision has not helped, only integrating for very short time periods
I have tried solving the system using the Gear and other implicit
methods and they appear to be outputting the correct solution, but I
would like to make use of the computational economy that I have heard
the Burlisch Stoer method provides as speed will be an issue in my
projects future.
I have been using the standard control object of a_y = 1 and a_dydt =
0 . Could my steppers confusion have to do with a scaling problem? I
have read that one must scale variables carefully when stepping with
this routine.

Thanks for any advice or ideas,

reply via email to

[Prev in Thread] Current Thread [Next in Thread]