By binaries, do you mean this http://sourceforge.net/projects/winglpk/? I
tried it and I installed glpsol using this.

As far as I understand, glpk binaries do not create libglpk.a or include
file. GLPKMex requires those files and one can get that by installing
through http://en.wikibooks.org/wiki/GLPK/Windows. Hence, I could not
install GLPKMex through the binaries. Please let me know if I am wrong.

The major problem is that I don't have vcvars64.bat file in Program
Files\Visual Studio\VC. If I want to create it manually with notepad, do
you know what I should put in there?

Thanks,

Nazmul


On Sun, Dec 29, 2013 at 3:39 PM, Jeffrey Kantor <[email protected]> wrote:

> Did you try using the binaries?
>
>
> On Sunday, December 29, 2013, Nazmul Islam wrote:
>
>> Hello Jeff,
>>
>> Thanks for your email. I am using windows and Matlab 2013. GLPKMex
>> requires GLPK installation through
>> http://en.wikibooks.org/wiki/GLPK/Windows. I was able to install GLPK
>> through windows executables (
>> http://en.wikibooks.org/wiki/GLPK/Windows_executables) but could not
>> install GLPK through http://en.wikibooks.org/wiki/GLPK/Windows. GLPKMex
>> is asking for glpk.h and libglpk.a file and I think that I can only get
>> them from GLPK Windows batch file based installation.
>>
>> Although, my current description will deviate from the subject topic, I
>> will describe the steps that I performed in GLPK installation.
>>
>> 1. I downloaded and installed all packages that are described in
>> http://en.wikibooks.org/wiki/GLPK/Windows (Visual C++, Windows SDK,
>> Visual Studio 2010, Visual C++ 2010 Service Pack).
>>
>> 2, I downloaded glpk-4.48 from http://ftp.gnu.org/gnu/glpk/.
>>
>> 3. I go to the /glpk-4.48/w64/ folder and run 'Build_GLPK_with_VC10_DLL'
>> as an administrator and it shows me the following message:
>>
>> "The specified configuration type is missing. The tools for the
>> configuration might not be installed" !!
>>
>> 4. I tried "CALL "C:\Program Files\Microsoft
>> SDKs\Windows\v7.1\Bin\SetEnv.cmd" /x64" in my cmd. But it did not create
>> the vcvars64.bat file in C:\Program Files (x86)\Microsoft Visual Studio
>> 10.0\VC\ directory. I don't know why this might be happening :S.
>>
>> 5. Someone of this forum suggested me to generated vcvars64.bat file by
>> notepad. But I don't know what I should put there. I see vcvarsall.bat file
>> in my C:/Program Files/Visual Studio 10.0/vc/ subfolder directory but I
>> don't know how I can change the contents to create vcvars64.bat file.
>>
>> My current option is to uninstall all packages of
>> http://en.wikibooks.org/wiki/GLPK/Windows (Visual C++, Windows SDK,
>> Visual Studio 2010, Visual C++ 2010 Service Pack) and install them again. I
>> wanted to check with you if I missed any step before I went through that
>> route.
>>
>> Any feedback regarding this installation will be really helpful. At
>> present, I don't have much clue why this is happening.
>>
>> Thanks,
>>
>> Nazmul
>>
>>
>>
>>
>> On Sun, Dec 29, 2013 at 11:25 AM, Jeffrey Kantor <[email protected]> wrote:
>>
>>
>> I may not be understanding your intended use. If the problem has already
>> been reduced to canonical form with A, b, c, lb, and ub as Matlab matrices,
>> then there is little value to introducing a modeling layer. You could just
>> call a glpk through glpkmex (see http://en.wikibooks.org/wiki/GLPK/Matlab)
>> and avoid the overhead of converting your data to a text format, building
>> the model, and reading the results.
>>
>> If you do wish to use a modeling language, such as Mathprog, then you
>> might consider using AWK to manipulate and generate instances of data
>> inputs to your model.  But from your note, it sounds like you've already
>> got that worked out in Matlab.
>>
>> Also, if you have access to it, you may want to look into the prerelease
>> of Matlab 2014a.
>>
>> Jeff
>>
>>
>> On Saturday, December 28, 2013, Nazmul Islam wrote:
>>
>> Hello Reg (& others),
>>
>> Thanks a lot for your email.
>>
>> As a fast step, I want to solve a mixed integer linear program like
>> follows
>>
>> min c'x
>> s.t. Ax <= b,
>> lb <= x <= ub
>> Some variables of x are binary ones.
>>
>> I have generated A, b, c, lb, ub in Matlab and stored them in .mat
>> format. I have to convert them to GMPL readable format so that I can solve
>> this code using GLPK. I have attached a sample of those files with my
>> email. I will really appreciate if you could tell me how I can convert them
>> using AWK. Please let me know if you want me to send a .mat file with
>> smaller size.
>>
>> Your help will be very appreciated.
>>
>> Thanks,
>>
>> Nazmul
>>
>>
>> On Fri, Dec 27, 2013 at 3:30 PM, Reginald Beardsley 
>> <[email protected]>wrote:
>>
>> Nazmul,
>>
>> I suggest you consider using  awk to generate data files in GMPL format
>> from ASCII text (.txt) files if you have very many instances to solve.  You
>> put the problem formulation in one file and the data for each instance in
>>  another file and execute "gplsol -m model.dat -d data.dat".  I've found
>> this works very well solving large numbers of problem instances.  In
>> particular, it makes changing the model very simple.  I can't find out
>> anything about the Guesk editor so I can't comment on setting up GMPL
>> notation data files using that.
>>
>> I tried using the CSV import facility in GMPL, but found it not very
>> convenient because of restrictions on the CSV format glpk accepts.
>>
>> There is a MATLAB binding for GLPK, but I don't know how usable it is.
>>  The MATLAB clone, Octave, also has a built-in binding to GLPK.
>>
>> http://en.wikibooks.org/wiki/GLPK/Matlab
>>
>> http://en.wikibooks.org/wiki/GLPK/Octave
>>
>> http://en.wikibooks.org/wiki/GLPK/Scripting_plus_MathProg
>>
>> There is a Gnu implementation of awk, gawk, which is standard on Linux
>> systems.  It's also available on Windows.
>>
>> http://gnuwin32.sourceforge.net/packages/gawk.htm
>>
>> If you need help with writing an awk script to convert the MATLAB .txt
>> files to GMPL notation send me a small example of the MATLAB .txt file and
>> I'll add an example to the GLPK wiki.
>>
>> You'll need to figure out how to write a basic GMPL problem first.  A
>> good place to look is in the examples directory at cf12a.mod which
>> implements an L1 line fit.  I solve physics inverse  problems by basis
>> pursuit using GMPL notation so I can probably offer some help.  I do
>> problems with 30,000 columns by 50-300 rows.  The resulting files are huge,
>> so you want to get everything working using a small example before doing
>> something that large.  I did some little 10x10 problems to figure out how
>> to set it up.
>>
>> Reg
>>
>>
>>
>>
>> --------------------------------------------
>> On Thu, 12/26/13, Nazmul Islam <[email protected]> wrote:
>>
>>  Subject: [Help-glpk] Import data fr
>>
>>
_______________________________________________
Help-glpk mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/help-glpk

Reply via email to