|
From: | Paul Schneider |
Subject: | Re: [Help-gsl] ODE solver gets stuck |
Date: | Wed, 16 Feb 2005 09:56:01 +0100 |
User-agent: | Mozilla Thunderbird 1.0 (X11/20041227) |
Paul Schneider wrote:
I'm trying to solve a system of ODE's with GSL. I have to say that I have no experience with ODE's and especially not with numerically nolving them. Therefore please don't hold back even with very basic hints.My problem: a 4 dimensional ODE can be solved rapidly, only for some parameter constellations GSL seems to get stuck in some loop and I have to manually kill the program. I looked at the output during such an incident and it seems that at some point t from below doesn't increase anymore, but the function values grow infinitely. I have tried all the solvers, with and without Jacobian. The following code is what I think is the relevant portion of my C++ program using GSL. TinyVector (PVec) is from the Blitz library.The function odeSolve below is called thousands of times for different (global) parameters KQd_ThQd_, KQd_. I am very grateful to all hints how I can get this to work. The ODE itself has not been reported to be especially difficult to handle.Thanks, Paul
[snip]I tried the same problem with matlab. Same thing. It doesn't seem to be GSL related. The only difference is that matlab doesn't get stuck, but returns with a message that it would have to choose a step size below the minimum step size 5.5e-15.
Can anybody give me a hint how I could tweak GSL to give me a solution. Or are the ODE's just not solvable for certain parametrizations? AS mentioned above, I am a beginner.
Thanks, Paul
[Prev in Thread] | Current Thread | [Next in Thread] |