예제 #1
0
static PyObject* Bar_getkind(BarObject *self, void *closure) {
  PyObject *retval=NULL;
  int kind = GLP_CV;
  if (!Bar_Valid(self, 1)) return NULL;
  if (!Bar_Row(self)) {
    kind = glp_get_col_kind(LP, Bar_Index(self)+1);
  }
  switch (kind) {
  case GLP_CV: retval = (PyObject*)&PyFloat_Type; break;
  case GLP_IV: retval = (PyObject*)&PyInt_Type;   break;
  case GLP_BV: retval = (PyObject*)&PyBool_Type;  break;
  default:
    PyErr_SetString(PyExc_RuntimeError,
		    "unexpected variable kind encountered");
    return NULL;
  }
  Py_INCREF(retval);
  return retval;
}
예제 #2
0
파일: ppl_lpsol.c 프로젝트: hnxiao/ppl
static void
solve(char* file_name) {
  ppl_Constraint_System_t ppl_cs;
#ifndef NDEBUG
  ppl_Constraint_System_t ppl_cs_copy;
#endif
  ppl_Generator_t optimum_location;
  ppl_Linear_Expression_t ppl_le;
  int dimension, row, num_rows, column, nz, i, j, type;
  int* coefficient_index;
  double lb, ub;
  double* coefficient_value;
  mpq_t rational_lb, rational_ub;
  mpq_t* rational_coefficient;
  mpq_t* objective;
  ppl_Linear_Expression_t ppl_objective_le;
  ppl_Coefficient_t optimum_n;
  ppl_Coefficient_t optimum_d;
  mpq_t optimum;
  mpz_t den_lcm;
  int optimum_found;
  glp_mpscp glpk_mpscp;

  glpk_lp = glp_create_prob();
  glp_init_mpscp(&glpk_mpscp);

  if (verbosity == 0) {
    /* FIXME: find a way to suppress output from glp_read_mps. */
  }

#ifdef PPL_LPSOL_SUPPORTS_TIMINGS

  if (print_timings)
    start_clock();

#endif /* defined(PPL_LPSOL_SUPPORTS_TIMINGS) */

  if (glp_read_mps(glpk_lp, GLP_MPS_FILE, &glpk_mpscp, file_name) != 0)
    fatal("cannot read MPS file `%s'", file_name);

#ifdef PPL_LPSOL_SUPPORTS_TIMINGS

  if (print_timings) {
    fprintf(stderr, "Time to read the input file: ");
    print_clock(stderr);
    fprintf(stderr, " s\n");
    start_clock();
  }

#endif /* defined(PPL_LPSOL_SUPPORTS_TIMINGS) */

  glpk_lp_num_int = glp_get_num_int(glpk_lp);

  if (glpk_lp_num_int > 0 && !no_mip && !use_simplex)
     fatal("the enumeration solving method can not handle MIP problems");

  dimension = glp_get_num_cols(glpk_lp);

  /* Read variables constrained to be integer. */
  if (glpk_lp_num_int > 0 && !no_mip && use_simplex) {
    if (verbosity >= 4)
      fprintf(output_file, "Integer variables:\n");
    integer_variables = (ppl_dimension_type*)
      malloc((glpk_lp_num_int + 1)*sizeof(ppl_dimension_type));
    for (i = 0, j = 0; i < dimension; ++i) {
      int col_kind = glp_get_col_kind(glpk_lp, i+1);
      if (col_kind == GLP_IV || col_kind == GLP_BV) {
        integer_variables[j] = i;
        if (verbosity >= 4) {
          ppl_io_fprint_variable(output_file, i);
          fprintf(output_file, " ");
        }
        ++j;
      }
    }
  }
  coefficient_index = (int*) malloc((dimension+1)*sizeof(int));
  coefficient_value = (double*) malloc((dimension+1)*sizeof(double));
  rational_coefficient = (mpq_t*) malloc((dimension+1)*sizeof(mpq_t));


  ppl_new_Constraint_System(&ppl_cs);

  mpq_init(rational_lb);
  mpq_init(rational_ub);
  for (i = 1; i <= dimension; ++i)
    mpq_init(rational_coefficient[i]);

  mpz_init(den_lcm);

  if (verbosity >= 4)
    fprintf(output_file, "\nConstraints:\n");

  /* Set up the row (ordinary) constraints. */
  num_rows = glp_get_num_rows(glpk_lp);
  for (row = 1; row <= num_rows; ++row) {
    /* Initialize the least common multiple computation. */
    mpz_set_si(den_lcm, 1);
    /* Set `nz' to the number of non-zero coefficients. */
    nz = glp_get_mat_row(glpk_lp, row, coefficient_index, coefficient_value);
    for (i = 1; i <= nz; ++i) {
      set_mpq_t_from_double(rational_coefficient[i], coefficient_value[i]);
      /* Update den_lcm. */
      mpz_lcm(den_lcm, den_lcm, mpq_denref(rational_coefficient[i]));
    }

    lb = glp_get_row_lb(glpk_lp, row);
    ub = glp_get_row_ub(glpk_lp, row);

    set_mpq_t_from_double(rational_lb, lb);
    set_mpq_t_from_double(rational_ub, ub);

    mpz_lcm(den_lcm, den_lcm, mpq_denref(rational_lb));
    mpz_lcm(den_lcm, den_lcm, mpq_denref(rational_ub));

    ppl_new_Linear_Expression_with_dimension(&ppl_le, dimension);

    for (i = 1; i <= nz; ++i) {
      mpz_mul(tmp_z, den_lcm, mpq_numref(rational_coefficient[i]));
      mpz_divexact(tmp_z, tmp_z, mpq_denref(rational_coefficient[i]));
      ppl_assign_Coefficient_from_mpz_t(ppl_coeff, tmp_z);
      ppl_Linear_Expression_add_to_coefficient(ppl_le, coefficient_index[i]-1,
                                               ppl_coeff);
    }

    type = glp_get_row_type(glpk_lp, row);
    add_constraints(ppl_le, type, rational_lb, rational_ub, den_lcm, ppl_cs);

    ppl_delete_Linear_Expression(ppl_le);
  }

  free(coefficient_value);
  for (i = 1; i <= dimension; ++i)
    mpq_clear(rational_coefficient[i]);
  free(rational_coefficient);
  free(coefficient_index);

#ifndef NDEBUG
  ppl_new_Constraint_System_from_Constraint_System(&ppl_cs_copy, ppl_cs);
#endif

  /*
    FIXME: here we could build the polyhedron and minimize it before
    adding the variable bounds.
  */

  /* Set up the columns constraints, i.e., variable bounds. */
  for (column = 1; column <= dimension; ++column) {

    lb = glp_get_col_lb(glpk_lp, column);
    ub = glp_get_col_ub(glpk_lp, column);

    set_mpq_t_from_double(rational_lb, lb);
    set_mpq_t_from_double(rational_ub, ub);

    /* Initialize the least common multiple computation. */
    mpz_set_si(den_lcm, 1);
    mpz_lcm(den_lcm, den_lcm, mpq_denref(rational_lb));
    mpz_lcm(den_lcm, den_lcm, mpq_denref(rational_ub));

    ppl_new_Linear_Expression_with_dimension(&ppl_le, dimension);
    ppl_assign_Coefficient_from_mpz_t(ppl_coeff, den_lcm);
    ppl_Linear_Expression_add_to_coefficient(ppl_le, column-1, ppl_coeff);

    type = glp_get_col_type(glpk_lp, column);
    add_constraints(ppl_le, type, rational_lb, rational_ub, den_lcm, ppl_cs);

    ppl_delete_Linear_Expression(ppl_le);
  }

  mpq_clear(rational_ub);
  mpq_clear(rational_lb);

  /* Deal with the objective function. */
  objective = (mpq_t*) malloc((dimension+1)*sizeof(mpq_t));

  /* Initialize the least common multiple computation. */
  mpz_set_si(den_lcm, 1);

  mpq_init(objective[0]);
  set_mpq_t_from_double(objective[0], glp_get_obj_coef(glpk_lp, 0));
  for (i = 1; i <= dimension; ++i) {
    mpq_init(objective[i]);
    set_mpq_t_from_double(objective[i], glp_get_obj_coef(glpk_lp, i));
    /* Update den_lcm. */
    mpz_lcm(den_lcm, den_lcm, mpq_denref(objective[i]));
  }

  /* Set the ppl_objective_le to be the objective function. */
  ppl_new_Linear_Expression_with_dimension(&ppl_objective_le, dimension);
  /* Set value for objective function's inhomogeneous term. */
  mpz_mul(tmp_z, den_lcm, mpq_numref(objective[0]));
  mpz_divexact(tmp_z, tmp_z, mpq_denref(objective[0]));
  ppl_assign_Coefficient_from_mpz_t(ppl_coeff, tmp_z);
  ppl_Linear_Expression_add_to_inhomogeneous(ppl_objective_le, ppl_coeff);
  /* Set values for objective function's variable coefficients. */
  for (i = 1; i <= dimension; ++i) {
    mpz_mul(tmp_z, den_lcm, mpq_numref(objective[i]));
    mpz_divexact(tmp_z, tmp_z, mpq_denref(objective[i]));
    ppl_assign_Coefficient_from_mpz_t(ppl_coeff, tmp_z);
    ppl_Linear_Expression_add_to_coefficient(ppl_objective_le, i-1, ppl_coeff);
  }

  if (verbosity >= 4) {
    fprintf(output_file, "Objective function:\n");
    if (mpz_cmp_si(den_lcm, 1) != 0)
      fprintf(output_file, "(");
    ppl_io_fprint_Linear_Expression(output_file, ppl_objective_le);
  }

  for (i = 0; i <= dimension; ++i)
    mpq_clear(objective[i]);
  free(objective);

  if (verbosity >= 4) {
    if (mpz_cmp_si(den_lcm, 1) != 0) {
      fprintf(output_file, ")/");
      mpz_out_str(output_file, 10, den_lcm);
    }
    fprintf(output_file, "\n%s\n",
            (maximize ? "Maximizing." : "Minimizing."));
  }

  ppl_new_Coefficient(&optimum_n);
  ppl_new_Coefficient(&optimum_d);
  ppl_new_Generator_zero_dim_point(&optimum_location);

  optimum_found = use_simplex
    ? solve_with_simplex(ppl_cs,
                         ppl_objective_le,
                         optimum_n,
                         optimum_d,
                         optimum_location)
    : solve_with_generators(ppl_cs,
                            ppl_objective_le,
                            optimum_n,
                            optimum_d,
                            optimum_location);

  ppl_delete_Linear_Expression(ppl_objective_le);

  if (glpk_lp_num_int > 0)
      free(integer_variables);

  if (optimum_found) {
    mpq_init(optimum);
    ppl_Coefficient_to_mpz_t(optimum_n, tmp_z);
    mpq_set_num(optimum, tmp_z);
    ppl_Coefficient_to_mpz_t(optimum_d, tmp_z);
    mpz_mul(tmp_z, tmp_z, den_lcm);
    mpq_set_den(optimum, tmp_z);
    if (verbosity == 1)
      fprintf(output_file, "Optimized problem.\n");
    if (verbosity >= 2)
      fprintf(output_file, "Optimum value: %.10g\n", mpq_get_d(optimum));
    if (verbosity >= 3) {
      fprintf(output_file, "Optimum location:\n");
      ppl_Generator_divisor(optimum_location, ppl_coeff);
      ppl_Coefficient_to_mpz_t(ppl_coeff, tmp_z);
      for (i = 0; i < dimension; ++i) {
        mpz_set(mpq_denref(tmp1_q), tmp_z);
        ppl_Generator_coefficient(optimum_location, i, ppl_coeff);
        ppl_Coefficient_to_mpz_t(ppl_coeff, mpq_numref(tmp1_q));
        ppl_io_fprint_variable(output_file, i);
        fprintf(output_file, " = %.10g\n", mpq_get_d(tmp1_q));
      }
    }
#ifndef NDEBUG
    {
      ppl_Polyhedron_t ph;
      unsigned int relation;
      ppl_new_C_Polyhedron_recycle_Constraint_System(&ph, ppl_cs_copy);
      ppl_delete_Constraint_System(ppl_cs_copy);
      relation = ppl_Polyhedron_relation_with_Generator(ph, optimum_location);
      ppl_delete_Polyhedron(ph);
      assert(relation == PPL_POLY_GEN_RELATION_SUBSUMES);
    }
#endif
    maybe_check_results(PPL_MIP_PROBLEM_STATUS_OPTIMIZED,
                        mpq_get_d(optimum));
    mpq_clear(optimum);
  }

  ppl_delete_Constraint_System(ppl_cs);
  ppl_delete_Coefficient(optimum_d);
  ppl_delete_Coefficient(optimum_n);
  ppl_delete_Generator(optimum_location);

  glp_delete_prob(glpk_lp);
}
예제 #3
0
int lpx_get_col_kind(LPX *lp, int j)
{     /* retrieve column kind */
      return glp_get_col_kind(lp, j) == GLP_CV ? LPX_CV : LPX_IV;
}
예제 #4
0
static int branch_drtom(glp_tree *T, int *_next)
{     glp_prob *mip = T->mip;
      int m = mip->m;
      int n = mip->n;
      char *non_int = T->non_int;
      int j, jj, k, t, next, kase, len, stat, *ind;
      double x, dk, alfa, delta_j, delta_k, delta_z, dz_dn, dz_up,
         dd_dn, dd_up, degrad, *val;
      /* basic solution of LP relaxation must be optimal */
      xassert(glp_get_status(mip) == GLP_OPT);
      /* allocate working arrays */
      ind = xcalloc(1+n, sizeof(int));
      val = xcalloc(1+n, sizeof(double));
      /* nothing has been chosen so far */
      jj = 0, degrad = -1.0;
      /* walk through the list of columns (structural variables) */
      for (j = 1; j <= n; j++)
      {  /* if j-th column is not marked as fractional, skip it */
         if (!non_int[j]) continue;
         /* obtain (fractional) value of j-th column in basic solution
            of LP relaxation */
         x = glp_get_col_prim(mip, j);
         /* since the value of j-th column is fractional, the column is
            basic; compute corresponding row of the simplex table */
         len = glp_eval_tab_row(mip, m+j, ind, val);
         /* the following fragment computes a change in the objective
            function: delta Z = new Z - old Z, where old Z is the
            objective value in the current optimal basis, and new Z is
            the objective value in the adjacent basis, for two cases:
            1) if new upper bound ub' = floor(x[j]) is introduced for
               j-th column (down branch);
            2) if new lower bound lb' = ceil(x[j]) is introduced for
               j-th column (up branch);
            since in both cases the solution remaining dual feasible
            becomes primal infeasible, one implicit simplex iteration
            is performed to determine the change delta Z;
            it is obvious that new Z, which is never better than old Z,
            is a lower (minimization) or upper (maximization) bound of
            the objective function for down- and up-branches. */
         for (kase = -1; kase <= +1; kase += 2)
         {  /* if kase < 0, the new upper bound of x[j] is introduced;
               in this case x[j] should decrease in order to leave the
               basis and go to its new upper bound */
            /* if kase > 0, the new lower bound of x[j] is introduced;
               in this case x[j] should increase in order to leave the
               basis and go to its new lower bound */
            /* apply the dual ratio test in order to determine which
               auxiliary or structural variable should enter the basis
               to keep dual feasibility */
            k = glp_dual_rtest(mip, len, ind, val, kase, 1e-9);
            if (k != 0) k = ind[k];
            /* if no non-basic variable has been chosen, LP relaxation
               of corresponding branch being primal infeasible and dual
               unbounded has no primal feasible solution; in this case
               the change delta Z is formally set to infinity */
            if (k == 0)
            {  delta_z =
                  (T->mip->dir == GLP_MIN ? +DBL_MAX : -DBL_MAX);
               goto skip;
            }
            /* row of the simplex table that corresponds to non-basic
               variable x[k] choosen by the dual ratio test is:
                  x[j] = ... + alfa * x[k] + ...
               where alfa is the influence coefficient (an element of
               the simplex table row) */
            /* determine the coefficient alfa */
            for (t = 1; t <= len; t++) if (ind[t] == k) break;
            xassert(1 <= t && t <= len);
            alfa = val[t];
            /* since in the adjacent basis the variable x[j] becomes
               non-basic, knowing its value in the current basis we can
               determine its change delta x[j] = new x[j] - old x[j] */
            delta_j = (kase < 0 ? floor(x) : ceil(x)) - x;
            /* and knowing the coefficient alfa we can determine the
               corresponding change delta x[k] = new x[k] - old x[k],
               where old x[k] is a value of x[k] in the current basis,
               and new x[k] is a value of x[k] in the adjacent basis */
            delta_k = delta_j / alfa;
            /* Tomlin noticed that if the variable x[k] is of integer
               kind, its change cannot be less (eventually) than one in
               the magnitude */
            if (k > m && glp_get_col_kind(mip, k-m) != GLP_CV)
            {  /* x[k] is structural integer variable */
               if (fabs(delta_k - floor(delta_k + 0.5)) > 1e-3)
               {  if (delta_k > 0.0)
                     delta_k = ceil(delta_k);  /* +3.14 -> +4 */
                  else
                     delta_k = floor(delta_k); /* -3.14 -> -4 */
               }
            }
            /* now determine the status and reduced cost of x[k] in the
               current basis */
            if (k <= m)
            {  stat = glp_get_row_stat(mip, k);
               dk = glp_get_row_dual(mip, k);
            }
            else
            {  stat = glp_get_col_stat(mip, k-m);
               dk = glp_get_col_dual(mip, k-m);
            }
            /* if the current basis is dual degenerate, some reduced
               costs which are close to zero may have wrong sign due to
               round-off errors, so correct the sign of d[k] */
            switch (T->mip->dir)
            {  case GLP_MIN:
                  if (stat == GLP_NL && dk < 0.0 ||
                      stat == GLP_NU && dk > 0.0 ||
                      stat == GLP_NF) dk = 0.0;
                  break;
               case GLP_MAX:
                  if (stat == GLP_NL && dk > 0.0 ||
                      stat == GLP_NU && dk < 0.0 ||
                      stat == GLP_NF) dk = 0.0;
                  break;
               default:
                  xassert(T != T);
            }
            /* now knowing the change of x[k] and its reduced cost d[k]
               we can compute the corresponding change in the objective
               function delta Z = new Z - old Z = d[k] * delta x[k];
               note that due to Tomlin's modification new Z can be even
               worse than in the adjacent basis */
            delta_z = dk * delta_k;
skip:       /* new Z is never better than old Z, therefore the change
               delta Z is always non-negative (in case of minimization)
               or non-positive (in case of maximization) */
            switch (T->mip->dir)
            {  case GLP_MIN: xassert(delta_z >= 0.0); break;
               case GLP_MAX: xassert(delta_z <= 0.0); break;
               default: xassert(T != T);
            }
            /* save the change in the objective fnction for down- and
               up-branches, respectively */
            if (kase < 0) dz_dn = delta_z; else dz_up = delta_z;
         }
         /* thus, in down-branch no integer feasible solution can be
            better than Z + dz_dn, and in up-branch no integer feasible
            solution can be better than Z + dz_up, where Z is value of
            the objective function in the current basis */
         /* following the heuristic by Driebeck and Tomlin we choose a
            column (i.e. structural variable) which provides largest
            degradation of the objective function in some of branches;
            besides, we select the branch with smaller degradation to
            be solved next and keep other branch with larger degradation
            in the active list hoping to minimize the number of further
            backtrackings */
         if (degrad < fabs(dz_dn) || degrad < fabs(dz_up))
         {  jj = j;
            if (fabs(dz_dn) < fabs(dz_up))
            {  /* select down branch to be solved next */
               next = GLP_DN_BRNCH;
               degrad = fabs(dz_up);
            }
            else
            {  /* select up branch to be solved next */
               next = GLP_UP_BRNCH;
               degrad = fabs(dz_dn);
            }
            /* save the objective changes for printing */
            dd_dn = dz_dn, dd_up = dz_up;
            /* if down- or up-branch has no feasible solution, we does
               not need to consider other candidates (in principle, the
               corresponding branch could be pruned right now) */
            if (degrad == DBL_MAX) break;
         }
      }
      /* free working arrays */
      xfree(ind);
      xfree(val);
      /* something must be chosen */
      xassert(1 <= jj && jj <= n);
#if 1 /* 02/XI-2009 */
      if (degrad < 1e-6 * (1.0 + 0.001 * fabs(mip->obj_val)))
      {  jj = branch_mostf(T, &next);
         goto done;
      }
#endif
      if (T->parm->msg_lev >= GLP_MSG_DBG)
      {  xprintf("branch_drtom: column %d chosen to branch on\n", jj);
         if (fabs(dd_dn) == DBL_MAX)
            xprintf("branch_drtom: down-branch is infeasible\n");
         else
            xprintf("branch_drtom: down-branch bound is %.9e\n",
               lpx_get_obj_val(mip) + dd_dn);
         if (fabs(dd_up) == DBL_MAX)
            xprintf("branch_drtom: up-branch   is infeasible\n");
         else
            xprintf("branch_drtom: up-branch   bound is %.9e\n",
               lpx_get_obj_val(mip) + dd_up);
      }
done: *_next = next;
      return jj;
}
예제 #5
0
int c_glp_get_col_kind (glp_prob *lp, int i){
  	return glp_get_col_kind (lp, i);
}
예제 #6
0
// retrieve all missing values of LP/MILP
void Rglpk_retrieve_MP_from_file (char **file, int *type,
				  int *lp_n_constraints,
				  int *lp_n_objective_vars,
				  double *lp_objective_coefficients,
				  int *lp_constraint_matrix_i,
				  int *lp_constraint_matrix_j,
				  double *lp_constraint_matrix_values,
				  int *lp_direction_of_constraints,
				  double *lp_right_hand_side,
				  double *lp_left_hand_side,
				  int *lp_objective_var_is_integer,
				  int *lp_objective_var_is_binary,
				  int *lp_bounds_type,
				  double *lp_bounds_lower,
				  double *lp_bounds_upper,
				  int *lp_ignore_first_row,
				  int *lp_verbosity,
				  char **lp_constraint_names,
				  char **lp_objective_vars_names
				  ) {
  extern glp_prob *lp;
  glp_tran *tran;
  const char *str; 
  
  int i, j, lp_column_kind, tmp;
  int ind_offset, status;

  // Turn on/off Terminal Output
  if (*lp_verbosity==1)
    glp_term_out(GLP_ON);
  else
    glp_term_out(GLP_OFF);

  // create problem object
  if (lp)
    glp_delete_prob(lp);
  lp = glp_create_prob();

  // read file -> gets stored as an GLPK problem object 'lp'
  // which file type do we have?
  switch (*type){
  case 1: 
    // Fixed (ancient) MPS Format, param argument currently NULL
    status = glp_read_mps(lp, GLP_MPS_DECK, NULL, *file);
    break;
  case 2:
    // Free (modern) MPS format, param argument currently NULL
    status = glp_read_mps(lp, GLP_MPS_FILE, NULL, *file);
    break;
  case 3:
    // CPLEX LP Format
    status = glp_read_lp(lp, NULL, *file);
    break;
  case 4:
    // MATHPROG Format (based on lpx_read_model function)
    tran = glp_mpl_alloc_wksp();

    status = glp_mpl_read_model(tran, *file, 0);

    if (!status) {
        status = glp_mpl_generate(tran, NULL);
        if (!status) {
            glp_mpl_build_prob(tran, lp);
        }
    }
    glp_mpl_free_wksp(tran);
    break;    
  } 

  // if file read successfully glp_read_* returns zero
  if ( status != 0 ) {
    glp_delete_prob(lp);
    lp = NULL;
    error("Reading file %c failed.", *file);
  }
  
  if(*lp_verbosity==1)
    Rprintf("Retrieve column specific data ...\n");

  if(glp_get_num_cols(lp) != *lp_n_objective_vars) {
    glp_delete_prob(lp);
    lp = NULL;
    error("The number of columns is not as specified");
  }

  // retrieve column specific data (values, bounds and type)
  for (i = 0; i < *lp_n_objective_vars; i++) {
    lp_objective_coefficients[i] = glp_get_obj_coef(lp, i+1);
    
    // Note that str must not be freed befor we have returned
    // from the .C call in R! 
    str = glp_get_col_name(lp, i+1);    
    if (str){
      lp_objective_vars_names[i] = (char *) str;
    }
    
    lp_bounds_type[i]            = glp_get_col_type(lp, i+1);
    lp_bounds_lower[i]           = glp_get_col_lb  (lp, i+1);
    lp_bounds_upper[i]           = glp_get_col_ub  (lp, i+1);
    lp_column_kind               = glp_get_col_kind(lp, i+1);
    // set to TRUE if objective variable is integer or binary  
    switch (lp_column_kind){
    case GLP_IV: 
      lp_objective_var_is_integer[i] = 1;
      break;
    case GLP_BV:
      lp_objective_var_is_binary[i] = 1;
      break;
    }
  }
  
  ind_offset = 0;

  if(*lp_verbosity==1)
    Rprintf("Retrieve row specific data ...\n");

  if(glp_get_num_rows(lp) != *lp_n_constraints) {
    glp_delete_prob(lp);
    lp = NULL;
    error("The number of rows is not as specified");
  }

  // retrieve row specific data (right hand side, direction of constraints)
  for (i = *lp_ignore_first_row; i < *lp_n_constraints; i++) {
    lp_direction_of_constraints[i] = glp_get_row_type(lp, i+1);
    
    str = glp_get_row_name(lp, i + 1);
    if (str) { 
      lp_constraint_names[i] = (char *) str;
    }
    
    // the right hand side. Note we don't allow for double bounded or
    // free auxiliary variables 
    if( lp_direction_of_constraints[i] == GLP_LO )
      lp_right_hand_side[i] = glp_get_row_lb(lp, i+1);
    if( lp_direction_of_constraints[i] == GLP_UP )
      lp_right_hand_side[i] = glp_get_row_ub(lp, i+1);
    if( lp_direction_of_constraints[i] == GLP_FX )
      lp_right_hand_side[i] = glp_get_row_lb(lp, i+1);
    if( lp_direction_of_constraints[i] == GLP_DB  ){
      lp_right_hand_side[i] = glp_get_row_ub(lp, i+1);
      lp_left_hand_side[i] =  glp_get_row_lb(lp, i+1);
    }

    tmp = glp_get_mat_row(lp, i+1, &lp_constraint_matrix_j[ind_offset-1],
			           &lp_constraint_matrix_values[ind_offset-1]);
    if (tmp > 0)
      for (j = 0; j < tmp; j++)
	lp_constraint_matrix_i[ind_offset+j] = i+1;
	ind_offset += tmp; 
  }
  
  if(*lp_verbosity==1)
    Rprintf("Done.\n");

}