Re: [Help-glpk] instability problem

[Huy Nhiem Nguyen]

Hi Cheng,

Thanks so much. Your reply is very helpful. I go ahead and find that the 
smallest number that GLPK can deal with is 1E-07. Then, I change my data files 
so that there is no longer any number higher than 1E+07 or lower than 1E-07. It 
runs!!!!

Thanks again,

Best,

Neo Nguyen

On Wed, Nov 12, 2008 at 3:36 PM, Cheng,Jen-Min <address@hidden> wrote:
glpspx01.c  spx_primal


        /* make sure that the current basic solution remains primal
           feasible (or pseudo feasible on phase I) */
        if (check_stab(csa, parm->tol_bnd))
           {
           /* there are excessive bound violations due to round-off
              errors */
           if (parm->msg_lev >= GLP_MSG_ERR)
              xprintf("Warning: numerical instability (primal simplex,"
                 " phase %s)\n", csa->phase == 1 ? "I" : "II");
           /* restart the search */
           csa->phase = 0;
           binv_st = 0;
           rigorous = 5;
           goto loop;


-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of Huy
Nhiem Nguyen
Sent: Wednesday, November 12, 2008 12:48 PM
To: address@hidden
Subject: [Help-glpk] instability problem

Hi all,

Could someone please tell me what problem I encounter with?

I run glpsol and receive this message printed out like forever:

spx_simplex: numerical instability (primal simplex, phase II)

If GLPK cannot solve my model, why doesn #39;t it stop but continues
trying to solve?





Thanks a lot,

Best,

Neo Nguyen










Hi Cheng,

Thanks so much. Your reply is very helpful. I go ahead and find that the smallest number that GLPK can deal with is 1E-07. Then, I change my data files so that there is no longer any number higher than 1E+07 or lower than 1E-07. It runs!!!!

Thanks again,

Best,

Neo Nguyen

On Wed, Nov 12, 2008 at 3:36 PM, Cheng,Jen-Min <address@hidden> wrote:

glpspx01.c  spx_primal


        /* make sure that the current basic solution remains primal
           feasible (or pseudo feasible on phase I) */
        if (check_stab(csa, parm->tol_bnd))
           {
           /* there are excessive bound violations due to round-off
              errors */
           if (parm->msg_lev >= GLP_MSG_ERR)
              xprintf("Warning: numerical instability (primal simplex,"
                 " phase %s)\n", csa->phase == 1 ? "I" : "II");
           /* restart the search */
           csa->phase = 0;
           binv_st = 0;
           rigorous = 5;
           goto loop;

-----Original Message-----
From: help-glpk-bounces+jcheng=mwdh2o.com@gnu.org
[mailto:help-glpk-bounces+jcheng=mwdh2o.com@gnu.org] On Behalf Of Huy
Nhiem Nguyen
Sent: Wednesday, November 12, 2008 12:48 PM
To: address@hidden
Subject: [Help-glpk] instability problem

Hi all,

Could someone please tell me what problem I encounter with?

I run glpsol and receive this message printed out like forever:

spx_simplex: numerical instability (primal simplex, phase II)

If GLPK cannot solve my model, why doesn #39;t it stop but continues
trying to solve?

Thanks a lot,

Best,

Neo Nguyen