Digest for gamsworld@googlegroups.com - 2 Messages in 2 Topics

GAMS
1

hi. I have a linear MIP model. solving it through Cplex from mip
solver it takes a few seconds. but when I use Solver Baron from Mip
solver, it takes several minutes.I wonder why this happens, because
Baron use Cplex for linear models!

On 11/30/11, gamsworld@googlegroups.com wrote:

=============================================================================
Today’s Topic Summary

Group: gamsworld@googlegroups.com
Url: http://groups.google.com/group/gamsworld/topics

=============================================================================
Topic: energy subsidies reform model
Url: http://groups.google.com/group/gamsworld/t/400ce1642125b92c

---------- 1 of 1 ----------
From: sundeep chowdary
Date: Dec 01 12:05AM +0530
Url: http://groups.google.com/group/gamsworld/msg/795060dc9219a592

hi,
can you share the literature for the same, so that I can try to help
regarding the same.

On Tue, Nov 22, 2011 at 3:52 PM, hatef hazeri niri

=============================================================================
Topic: hi any one
Url: http://groups.google.com/group/gamsworld/t/60da2302a80f3994

---------- 1 of 1 ----------
From: Mehdi Amini
Date: Nov 29 11:44AM -0800
Url: http://groups.google.com/group/gamsworld/msg/1dcfe11d460fcda6

hi
i Appreciate your attemption for the time that has been elapsed.
i want to employ GAMS from matlam during a “For” loop( for i=1:100 …).
the file that you had sent just call the exquted results from GAMS to
matlab.
i will be ashamed if you send a simple example of that.
thanks very much

From: Muhajir
To: gamsworld@googlegroups.com
Sent: Tuesday, November 29, 2011 1:24 AM
Subject: Re: hi any one

Hi there,

I am new in this topic, but this works for me on the new GAM’s
version(23.7.3) and matlab 2010

First add matlab path to the GAMS folder

addpath ‘C:\gams’; savepath;

After the solve statement in gams, crate result.gdx with values of v1 v2 …
you want to export:

execute_unload ‘reuslt’ v1 v2 OF;

Then you can read this result in matlab from the gdx file using:

res.field = ’ l ';
r=rgdx(‘result’, res);
v2=r.val;
…

I’ve referred to the document:
http://www.gams.com/dd/docs/tools/gdxmrw.pdf
and an example can be found
http://www.gams.com/~steve/gdxmrw.html

You can also find a similar thread in this group:
http://groups.google.com/group/gamsworld/browse_thread/thread/721274c4dbf76464/5bca6ba9337171dd?hl=en&lnk=gst&q=gams+matlab#5bca6ba9337171dd

Hope this helps!

On Mon, Nov 28, 2011 at 4:05 PM, Mehdi Amini
wrote:

Hi

–
“gamsworld” group.
To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to
gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/gamsworld?hl=en.

–
“gamsworld” group.
To post to this group, send email to gamsworld@googlegroups.com.
To unsubscribe from this group, send email to
gamsworld+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/gamsworld?hl=en.


\

2

some1 help please!
I have a linear MIP model. solving it through Cplex from mip
solver it takes a few seconds. but when I use Baron from Mip
solver, it takes several minutes.I wonder why this happens, because
Baron use Cplex for linear models! I’m about to compare the
computational time between Cplex and the GA code of my model, I
thought Baron and Cplex would do the same for milp, facing this
problem I don’t no what to do…

\

3

HI
After reading numbers 1 and 2 of the documentation about Baron found in
gams, I see different possible explanations for your problem, though I have
never used Baron myself and I don’t know which is the correct answer, these
are just ideas I developed:

  1. It seems that Baron was created as an heuristic algorithm that uses other
    solvers as a subroutine, so that it might get to global optimal solutions
    instead of local optima. So, is not surprising that it takes more time to
    get the solution with Baron than with Cplex, because that means that
    whenever Baron gets a solution from Cplex, it takes it as a local optimal
    solution, then processes the equations in some way and sends new problems to
    CPlex because that’s exactly the objective of Baron: being sure that it is
    the global optima. So, it is highly likely that Baron will make cplex
    iterate again and again through other zones of the feasible region until is
    sure that it has reached the global optima. This is possible because even
    Cplex does not always give you the optimal solution with MIP. Cplex will
    manage certain tolerances and give you a solution unless you specify
    Optca=0. So, maybe you are running Cplex with no options and it gives you a
    solution within tolerances but when you Baron, is highly probable that Baron
    takes away that tolerance. However, you might review that and change the
    tolerance in Baron as it is specified in the documentation.
  2. The documentation of baron says that “in order to achieve convergence to
    global optimality, additional model constraints may be required. The
    additional constraints may speed up solver solution time and increase the
    probability of success”. The documentation speaks of giving lower and upper
    values for the values of the variables if you know them…that might help
    also, though again, I have never worked with Baron.


    So, some suggestions or questions for you
    a) You might study the options for branching, range, tolerances and so forth
    at the documentation of Baron and maybe using some of the options you can
    improve the response time of Baron.
    b) You might add the upper and lower values of the variables.
    C) If you get the results you want with Cplex, why to run with Baron?


    Best Regards
    Jose

    \
4

I see!thanks for your answer. your ideas make sense. I used Baron
because it gives better solution with lower gap for my model.by
changing some Cplex options due to your ideas I think I can fix my
problem.
best regards
banafshe

\

5

that looks like a RAM problem,may be you should check the hardware and operation system

hope it helps!

regards!


On Mon, Apr 1, 2013 at 5:11 AM, wrote:

Today’s Topic Summary

Group: http://groups.google.com/group/gamsworld/topics

Error Messages [1 Update]
Error reading .lst file [1 Update]

Error Messages

“Özgür YaÅŸar ÇaÄŸlar” Mar 30 02:33PM -0700

Hi,

I solved my problem with defining new variables for the expressions in the
logs and gave lower bounds for the new variables.

Regards,

On Wednesday, March 27, 2013 6:29:05 PM UTC+2, Özgür Yaşar Çağlar wrote:



Error reading .lst file

“Thanh Tung” Mar 30 05:50AM -0700

Dear GAMS experts,

After running a large set of data, I have got the error “Error = Error 0:
The operation completed successfully” and can’t open .lst file. Then, from
the new window of Gamside, I have tried to open that .lst file again and got
error: “Access violation at address 006E5D38 in module ‘gamside.exe’. Read
of address 00000356”.

My .lst file is more than 2.9Gb.

Please give me your advices. Thank you very much.

Diep Thanh Tung.



You received this message because you are subscribed to the Google Group gamsworld.
You can post via email.
To unsubscribe from this group, send an empty message.
For more options, visit this group.

–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.



\

To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

\