Regarding problem in GAMS error 13 modelstat and solvestat

GAMS
1

Respected Sir…

Thanks a ton…I finally got rid of 13 error…
ya system call works…

But sir I am still stuck at unbounded solutions. My problem MIN c’x subject to Ax>=b with x unrestricted and A,b,c randomly generated mostly gives unbounded result as dimension of m and n increases.

Usually it works for m>n but not wen n>m … Sir is it because of random data that i am generating?
Pl Guide me.

It would be great if you could help me in this regard.


Thanks and regards
deepti

On Mon, Sep 12, 2011 at 6:25 PM, Steven Dirkse wrote:

Deepti,

I suggest you have a second look at how you are making the call from Matlab into GAMS. If you are doing solves in a look you should not use the gams() function but call gams via Matlab’s system() function instead. Details are here:

http://support.gams-software.com/doku.php?id=matlab_and_gams:interfacing_optimization_and_visualization_software_via_the_gdxmrw_utilities#error_running_gams_function_in_a_matlab_loop

HTH,

-Steve


On Fri, Sep 9, 2011 at 2:51 AM, deepti singhal wrote:

Respected Sir

I am Deepti Chadha(Singhal) a student of M.Tech(Comp app) , IIT delhi
Sir I am working on a project on Privacy LPP and I am facing problem while solving LPP in GAMS using MAtlab

I need urgently help on my problem. I am really new to GAMS world.

The problem is

MIN c’x subject to Ax>= b
x’s are unrestricted

c, A, b all are randomly generated vectors (uniform in the interval
[1,10] [-10,10] )
now I am solving the problem till I get an optimal solution . However, my problem is generally unbounded or infeasible( because of random

data generated). So I solve it in a loop (again generate new vectors) till i get some solution. Now while solving the GAMS give mostly an
error 13(ERRoR NO SOLUTION) and after that even if i execute it afresh, the matlab gets stuck at this odd no. 13 and i have to
restart matlab to get the answers. and after a fresh start, matlab usually gives optimal solution.

Pl try to help me on this 13… where i am making a mistake…

Thanks and regards

deepti

–
Deepti Chadha (Singhal)



\

Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com



\

Deepti Chadha (Singhal)

2

Deepti,

If you have n variables, you will also typically have n linearly independent constraints active at a solution. Since you have no variable bounds to become active, if you have n > m then you won’t have enough constraints to be active and you’ll get an unbounded problem.

I’d guess you need to adjust your scheme so that you are generating solvable LPs, or else figure out a way to proceed in the presence of unbounded LPs.

-Steve

On Mon, Sep 12, 2011 at 10:15 AM, deepti singhal wrote:

Respected Sir…

Thanks a ton…I finally got rid of 13 error…
ya system call works…

But sir I am still stuck at unbounded solutions. My problem MIN c’x subject to Ax>=b with x unrestricted and A,b,c randomly generated mostly gives unbounded result as dimension of m and n increases.

Usually it works for m>n but not wen n>m … Sir is it because of random data that i am generating?
Pl Guide me.

It would be great if you could help me in this regard.


Thanks and regards
deepti

On Mon, Sep 12, 2011 at 6:25 PM, Steven Dirkse wrote:

Deepti,

I suggest you have a second look at how you are making the call from Matlab into GAMS. If you are doing solves in a look you should not use the gams() function but call gams via Matlab’s system() function instead. Details are here:

http://support.gams-software.com/doku.php?id=matlab_and_gams:interfacing_optimization_and_visualization_software_via_the_gdxmrw_utilities#error_running_gams_function_in_a_matlab_loop

HTH,

-Steve


On Fri, Sep 9, 2011 at 2:51 AM, deepti singhal wrote:

Respected Sir

I am Deepti Chadha(Singhal) a student of M.Tech(Comp app) , IIT delhi
Sir I am working on a project on Privacy LPP and I am facing problem while solving LPP in GAMS using MAtlab

I need urgently help on my problem. I am really new to GAMS world.

The problem is

MIN c’x subject to Ax>= b
x’s are unrestricted

c, A, b all are randomly generated vectors (uniform in the interval
[1,10] [-10,10] )
now I am solving the problem till I get an optimal solution . However, my problem is generally unbounded or infeasible( because of random

data generated). So I solve it in a loop (again generate new vectors) till i get some solution. Now while solving the GAMS give mostly an
error 13(ERRoR NO SOLUTION) and after that even if i execute it afresh, the matlab gets stuck at this odd no. 13 and i have to
restart matlab to get the answers. and after a fresh start, matlab usually gives optimal solution.

Pl try to help me on this 13… where i am making a mistake…

Thanks and regards

deepti

–
Deepti Chadha (Singhal)



\

Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com



\

Deepti Chadha (Singhal)



\

Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com

3

Respected Steven Sir

Thanks a ton. Thanks a lot sir for such a quick reply.you are absolutely right… I read carefully the research paper on Privacy LPP that I am trying to implement after reading ur response and observed exactly what you had described to me.
They have taken A in [-50,50]
also the components of primal(x) were uniformly distributed in [-5,5] with about half of the primal constraints active with corresponding duals uniformly distributed in [0,10].

With these conditions imposed, they had executed program successfully on m=100, n=1000 and m=1000 and n=100

Sir , I have understood how to put bounds on decision variables (by up and lo) . But how can I check the active constraints in GAMS and that too that there are exactly 50 percent of them being active.
It would be really great sir if you could guide me in this regard.


Thanks and regards

Deepti

On Tue, Sep 13, 2011 at 8:28 PM, Steven Dirkse wrote:

Deepti,

If you have n variables, you will also typically have n linearly independent constraints active at a solution. Since you have no variable bounds to become active, if you have n > m then you won’t have enough constraints to be active and you’ll get an unbounded problem.

I’d guess you need to adjust your scheme so that you are generating solvable LPs, or else figure out a way to proceed in the presence of unbounded LPs.

-Steve


On Mon, Sep 12, 2011 at 10:15 AM, deepti singhal wrote:

Respected Sir…

Thanks a ton…I finally got rid of 13 error…
ya system call works…

But sir I am still stuck at unbounded solutions. My problem MIN c’x subject to Ax>=b with x unrestricted and A,b,c randomly generated mostly gives unbounded result as dimension of m and n increases.

Usually it works for m>n but not wen n>m … Sir is it because of random data that i am generating?
Pl Guide me.

It would be great if you could help me in this regard.


Thanks and regards
deepti

On Mon, Sep 12, 2011 at 6:25 PM, Steven Dirkse wrote:

Deepti,

I suggest you have a second look at how you are making the call from Matlab into GAMS. If you are doing solves in a look you should not use the gams() function but call gams via Matlab’s system() function instead. Details are here:

http://support.gams-software.com/doku.php?id=matlab_and_gams:interfacing_optimization_and_visualization_software_via_the_gdxmrw_utilities#error_running_gams_function_in_a_matlab_loop

HTH,

-Steve


On Fri, Sep 9, 2011 at 2:51 AM, deepti singhal wrote:

Respected Sir

I am Deepti Chadha(Singhal) a student of M.Tech(Comp app) , IIT delhi
Sir I am working on a project on Privacy LPP and I am facing problem while solving LPP in GAMS using MAtlab

I need urgently help on my problem. I am really new to GAMS world.

The problem is

MIN c’x subject to Ax>= b
x’s are unrestricted

c, A, b all are randomly generated vectors (uniform in the interval
[1,10] [-10,10] )
now I am solving the problem till I get an optimal solution . However, my problem is generally unbounded or infeasible( because of random

data generated). So I solve it in a loop (again generate new vectors) till i get some solution. Now while solving the GAMS give mostly an
error 13(ERRoR NO SOLUTION) and after that even if i execute it afresh, the matlab gets stuck at this odd no. 13 and i have to
restart matlab to get the answers. and after a fresh start, matlab usually gives optimal solution.

Pl try to help me on this 13… where i am making a mistake…

Thanks and regards

deepti

–
Deepti Chadha (Singhal)



\

Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com



\

Deepti Chadha (Singhal)



\

Steven Dirkse, Ph.D.
GAMS Development Corp., Washington DC
Voice: (202)342-0180 Fax: (202)342-0181
sdirkse@gams.com
http://www.gams.com



\

Deepti Chadha (Singhal)