Using algencan for a subproblem

43 views
Skip to first unread message

gabarosky

unread,
Mar 8, 2012, 7:38:55 AM3/8/12
to tango-...@googlegroups.com
Hello

I have an algorithm and I need solve a NLP, like subproblem. I had tried to use ALGENCAN for this subproblem but I couldn't refresh the subproblem between iterations. There is a way for use algencan in this case?

Ernesto G. Birgin

unread,
Mar 8, 2012, 8:06:22 AM3/8/12
to tango-...@googlegroups.com
Dear Gabarosky,

Thank you very much for your interest in Algencan. I am not sure about the meaning of "refreshing the subproblem". If, by refreshing, you mean to set up the subproblem properties like number of variables, constraints, initial guess, bounds, etc, you must call subroutine inip to do that.

I hope it helps. Do not hesitate in contact us in case yoiu have further questions or comments.

Regards,
Ernesto.


On Thursday, March 8, 2012 9:38:55 AM UTC-3, gabarosky wrote:
Hello

I have an algorithm and I need solve a NLP, like subproblem. I had tried to use ALGENCAN for this subproblem but I couldn't refresh the subproblem between iterations. There is a way for use algencan in this case?

gabarosky

unread,
Mar 9, 2012, 2:26:45 PM3/9/12
to tango-...@googlegroups.com
Hi Ernesto

Thank you for your atention. I need change the functions for the contraints and the change depends on the result of the previous solution. 
For a few iterations I changed the functions: evalc, evaljac and evalhc; and compile again the program; but I'm loking for a way for make this changes automatically.

Regards,
gAbriel

Ernesto G. Birgin

unread,
Mar 9, 2012, 5:22:55 PM3/9/12
to tango-...@googlegroups.com
Hi! Do you mean that you need to change a few parameters or do you need to change the expression of the constraints and the number of constraints? If your case is the first one, a common block (global variables) may solve your problem. Otherwise, if you need to code new constraints, you should code all possibilities in advance, within a big if statement. If your case is none of them, plesae, describe it in more detail. Regards, Ernesto.

Jose Mario Martinez

unread,
Mar 9, 2012, 8:37:26 PM3/9/12
to gabarosky, tango-...@googlegroups.com
  A common probably solves.
  common/gabarosky/nprob
  if(nprob.eq.1) then
  ...
  endif

   Mario

--
You received this message because you are subscribed to the Google Groups "TANGO Project - ALGENCAN" group.
To view this discussion on the web visit https://groups.google.com/d/msg/tango-project/-/QdnM0qYx-0UJ.

To post to this group, send email to tango-...@googlegroups.com.
To unsubscribe from this group, send email to tango-projec...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/tango-project?hl=en.

Reply all
Reply to author
Forward
0 new messages