| |
 |
|
|
Science Forum Index » Math - Numerical Analysis Forum » DDASKR question
Page 1 of 1
|
| Author |
Message |
| gurry |
 Posted: Wed May 09, 2007 2:57 am |
|
|
|
Guest
|
Hi,
The documentation of DDASKR says that after the solver returns IDID =
5, one should do this to continue integration:
"call the code again to continue the integration in
the
direction of TOUT. You may change the functions
Ri defined by RT after a return with IDID = 5, but
the number of constraint functions NRT must remain
the same. If you wish to change the functions in
RES or in RT, then you must restart the code."
My question is, if I change functions in RT and I have to restart the
code afresh (by resetting info(1) = 1), why can't I change the number
of constraint equations in RT too? My understanding is that when you
restart with info(1) = 1, you are essentially solving a totally fresh
problem, and therefore, it should be perfectly okay to change the
number of constraint equations as well.
May someone please explain then what the above instruction from the
documentation means by "the number of constraint functions NRT must
remain the same"?
Thank you.
Gurinder. |
|
|
| Back to top |
|
| Torsten Hennig |
| Posted: Wed May 09, 2007 4:25 am |
|
|
|
Guest
|
Quote: Hi,
The documentation of DDASKR says that after the solver >returns IDID =
5, one should do this to continue integration:
"call the code again to continue the integration in
the
direction of TOUT. You may change the functions
Ri defined by RT after a return with IDID = 5, but
the number of constraint functions NRT must remain
the same. If you wish to change the functions in
RES or in RT, then you must restart the code."
My question is, if I change functions in RT and I have >to restart the
code afresh (by resetting info(1) = 1), why can't I >change the number
of constraint equations in RT too? My understanding is >that when you
restart with info(1) = 1, you are essentially solving a >totally fresh
problem, and therefore, it should be perfectly okay to >change the
number of constraint equations as well.
A totally new problem is started by setting info(1)=0.
By setting info(1)=1 and calling DASKR again
for continuation of the current task, DASSL uses
all relevant information from the last time step
to continue integration (e.g. size of the problem, last step size taken etc.).
So it is not possible to change the dimension of the
problem or the functions in RES when returning info(1)=1.
Quote: May someone please explain then what the above >instruction from the
documentation means by "the number of constraint >functions NRT must
remain the same"?
Thank you.
Gurinder. |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Thu Jan 08, 2009 12:41 am
|
|