|
Science Forum Index » Math - Numerical Analysis Forum » Why does DDASKR have problem with these eqns.?
Page 1 of 1
|
| Author |
Message |
| gurry |
 Posted: Thu Mar 22, 2007 8:43 am |
|
|
|
Guest
|
Hi,
I'm using DDASKR to solve these two equations:
residue[0] = yPrime[0] - 1.0;
residue[1] = y[1] - y[0] + 1;
And the initial values I've set are these:
y[0] = 1.0;
y[1] = 0.0;
yPrime[0] = 1.0;
yPrime[1] = 1.0;
(I'm calling DDASKR from C code. That's why I have these square
brackets instead of round ones.)
As I understand, this is a index 1 DAE system and the initial values
are consistent too. But DDASKR throws up the following error when I
try to run the integration:
DASKR-- AT T (=R1) AND STEPSIZE H (=R2) THE
In above, R1 = 0.0000000000000D+00 R2 = 0.3510833468577D-14
DASKR-- ITERATION MATRIX IS SINGULAR.
My relative and absolute tolerance values are both 1.0e-14.
I just want to know if there's any problem with the way I've
formulated my equations or set my initial values. If there isn't any
such problem, then I can be sure that the fault is with the rest of my
code.
Kindly help.
Thanks in advance.
Gurinder |
|
|
| Back to top |
|
| Torsten Hennig |
| Posted: Thu Mar 22, 2007 8:43 am |
|
|
|
Guest
|
Quote: Hi,
I'm using DDASKR to solve these two equations:
residue[0] = yPrime[0] - 1.0;
residue[1] = y[1] - y[0] + 1;
And the initial values I've set are these:
y[0] = 1.0;
y[1] = 0.0;
yPrime[0] = 1.0;
yPrime[1] = 1.0;
(I'm calling DDASKR from C code. That's why I have these >square
brackets instead of round ones.)
As I understand, this is a index 1 DAE system and the >initial values
are consistent too. But DDASKR throws up the following >error when I
try to run the integration:
DASKR-- AT T (=R1) AND STEPSIZE H (=R2) THE
In above, R1 = 0.0000000000000D+00 R2 = 0.3510833468577D->14
DASKR-- ITERATION MATRIX IS SINGULAR.
My relative and absolute tolerance values are both 1.0e->14.
The error tolerances you set are quite stringent.
With a relative and absolute error tolerance of
1e-8, integration of your test problem with DASPK
is successful on my PC.
For lower error tolerances, I get the same
error message as you did.
Quote:
I just want to know if there's any problem with the way >I've
formulated my equations or set my initial values. If >there isn't any
such problem, then I can be sure that the fault is with >the rest of my
code.
Kindly help.
Thanks in advance.
Gurinder
Best wishes
Torsten. |
|
|
| Back to top |
|
| gurry |
| Posted: Fri Mar 23, 2007 12:34 am |
|
|
|
Guest
|
Thanks a lot Torsten. I've now also consulted the Petzold book and the
thing about tolerances you said makes sense because too strict
tolerances screw up jacobian approximation.
Your help is much appreciated.
~Gurinder |
|
|
| Back to top |
|
| |
