[AMPL 2926] re-optimization

[Guney Petek]

Dear list readers,

I have a large MIP where I do a series of runs with a change of a
parameter array.

Solving problems are not taking long once they are fed into CPLEX,
around ten seconds..., however I think AMPL rebuilds each problem from
scratch although most of the model is the same as previous model.

Is it possible to ask AMPL not to rebuild the model from scratch, and
change only the parameters that were changed? So re-optimization would
be faster. Could you direct me to a document or a chapter in AMPL
book?

Thank you, \Guney

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---

[Robert Fourer-2]

After AMPL has built a data structure to represent your model and data in
memory, it does only a quick incremental revision each time a parameter
array is changed.  However, there is no way to avoid writing out a complete
problem file for CPLEX each time a solve is done.  Usually the time to write
the file is only a fraction of the time to solve; but you can ask AMPL to
tell you how long it takes by doing a test run with "option times 1;" set at
the beginning.

Bob Fourer
4er@...

> -----Original Message-----
> From: ampl@... [mailto:ampl@...]
> On Behalf Of Guney Petek [guneyp@...]
> Sent: Wednesday, October 28, 2009 2:18 PM
> To: AMPL Modeling Language
> Subject: [AMPL 2926] re-optimization
>
>
> Dear list readers,
>
> I have a large MIP where I do a series of runs with a change of a
> parameter array.
>
> Solving problems are not taking long once they are fed into CPLEX,
> around ten seconds..., however I think AMPL rebuilds each problem from
> scratch although most of the model is the same as previous model.
>
> Is it possible to ask AMPL not to rebuild the model from scratch, and
> change only the parameters that were changed? So re-optimization would
> be faster. Could you direct me to a document or a chapter in AMPL
> book?
>
> Thank you, \Guney
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---

[Guney Petek]

Thank you for your reply.  Also, we figured that it was presolve mechanism that was taking great deal of that extra time. Now that I use option presolve 0, AMPL processing time has reduced drastically.

Regards, \Guney

On Wed, Oct 28, 2009 at 8:14 PM, Robert Fourer <4er@...> wrote:

After AMPL has built a data structure to represent your model and data in
memory, it does only a quick incremental revision each time a parameter
array is changed.  However, there is no way to avoid writing out a complete
problem file for CPLEX each time a solve is done.  Usually the time to write
the file is only a fraction of the time to solve; but you can ask AMPL to
tell you how long it takes by doing a test run with "option times 1;" set at
the beginning.

Bob Fourer
4er@...

> -----Original Message-----
> From: ampl@... [mailto:ampl@...]
> On Behalf Of Guney Petek [guneyp@...]
> Sent: Wednesday, October 28, 2009 2:18 PM
> To: AMPL Modeling Language
> Subject: [AMPL 2926] re-optimization
>
>
> Dear list readers,
>
> I have a large MIP where I do a series of runs with a change of a
> parameter array.
>
> Solving problems are not taking long once they are fed into CPLEX,
> around ten seconds..., however I think AMPL rebuilds each problem from
> scratch although most of the model is the same as previous model.
>
> Is it possible to ask AMPL not to rebuild the model from scratch, and
> change only the parameters that were changed? So re-optimization would
> be faster. Could you direct me to a document or a chapter in AMPL
> book?
>
> Thank you, \Guney
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---