Re: [Help-glpk] different solutions for same code

[xypron]

Hello,

have a look a  glp_main() in src/glpapi19.c (of glpk4.39). 
Here you find how the defaults of glpsol are set:

      glp_init_iocp(&csa->iocp);
      csa->iocp.presolve = GLP_ON;
      csa->tran = NULL;
      csa->graph = NULL;
      csa->format = FMT_MPS_FILE;
      csa->in_file = NULL;
      csa->ndf = 0;
      csa->out_dpy = NULL;
      csa->solution = SOL_BASIC;
      csa->in_res = NULL;
      csa->dir = 0;
      csa->scale = 1;
      csa->out_sol = NULL;
      csa->out_res = NULL;
      csa->out_bnds = NULL;
      csa->check = 0;
      csa->new_name = NULL;
      csa->out_mps = NULL;
      csa->out_freemps = NULL;
      csa->out_cpxlp = NULL;
      csa->out_pb = NULL;
      csa->out_npb = NULL;
      csa->log_file = NULL;
      csa->crash = USE_ADV_BASIS;
      csa->exact = 0;
      csa->xcheck = 0;
      csa->nomip = 0;

If you set a time limit I guess the parameters scale and crash could make
some difference.

Best regards

Xypron


husalihas wrote:
> 
> Hello ,
> 
> I wrote a code in "mathprog" to solve MILP problem. The code is running
> fine with GLPK stand alone solver with the following options:
> --dfs --first --mir --gomory --cover --tmlim 10
> the minimum price is : 928173
> But after solving the same code (and same data) using API routines as
> shown below , I got differen price of : 949072 (and different solution).
> Can any body explain why ? is there any thing missing in my C++ code shown
> below?
> 
> #include <stdio.h>
> #include <stdlib.h>
> #include <glpk.h>
> 
> int main(void)
> {     glp_prob *mip;
>       glp_iocp parm;
>       glp_tran *tran;
>       int ret;
>       mip = glp_create_prob();
>       tran = glp_mpl_alloc_wksp();
>       ret = glp_mpl_read_model(tran, "mycode.mod", 1);
>       if (ret != 0)
>       {  fprintf(stderr, "Error on translating model\n");
>          goto skip;
>       }
>       ret = glp_mpl_read_data(tran, "mycode.dat");
>       if (ret != 0)
>       {  fprintf(stderr, "Error on translating data\n");
>          goto skip;
>       }
>       ret = glp_mpl_generate(tran, NULL);
>       if (ret != 0)
>       {  fprintf(stderr, "Error on generating model\n");
>          goto skip;
>       }
>       glp_mpl_build_prob(tran, mip);
>       glp_simplex(mip, NULL);
>       glp_init_iocp (&parm) ;
>       parm.br_tech = GLP_BR_FFV ;
>       parm.bt_tech = GLP_BT_DFS ;
>       parm.gmi_cuts = GLP_ON ;
>       parm.mir_cuts = GLP_ON ;
>       parm.cov_cuts = GLP_ON ;
>       parm.tm_lim = 10000 ;
>       ret = glp_intopt(mip, &parm);
>       ret = glp_mpl_postsolve(tran, mip, GLP_MIP); 
>       if (ret != 0)
>          fprintf(stderr, "Error on postsolving model\n");
>       glp_print_mip(mip, "mycode.txt"); 
> skip: glp_mpl_free_wksp(tran);
>       glp_delete_prob(mip);
>       return 0;
> }
> 

-- 
View this message in context: 
http://www.nabble.com/different-solutions-for-same-code-tp25245249p25269699.html
Sent from the Gnu - GLPK - Help mailing list archive at Nabble.com.