fixanddive_c++.cpp


/* Copyright 2019, Gurobi Optimization, LLC */

/* Implement a simple MIP heuristic.  Relax the model,
   sort variables based on fractionality, and fix the 25% of
   the fractional variables that are closest to integer variables.
   Repeat until either the relaxation is integer feasible or
   linearly infeasible. */

#include "gurobi_c++.h"
#include <algorithm>
#include <cmath>
#include <deque>
using namespace std;

bool vcomp(GRBVar*, GRBVar*) throw(GRBException);

int
main(int argc,
     char *argv[])
{
  if (argc < 2)
  {
    cout << "Usage: fixanddive_c++ filename" << endl;
    return 1;
  }

  GRBEnv* env = 0;
  GRBVar* x = 0;
  try
  {
    // Read model
    env = new GRBEnv();
    GRBModel model = GRBModel(*env, argv[1]);

    // Collect integer variables and relax them
    // Note that we use GRBVar* to copy variables
    deque<GRBVar*> intvars;
    x = model.getVars();
    for (int j = 0; j < model.get(GRB_IntAttr_NumVars); ++j)
    {
      if (x[j].get(GRB_CharAttr_VType) != GRB_CONTINUOUS)
      {
        intvars.push_back(&x[j]);
        x[j].set(GRB_CharAttr_VType, GRB_CONTINUOUS);
      }
    }

    model.set(GRB_IntParam_OutputFlag, 0);
    model.optimize();

    // Perform multiple iterations. In each iteration, identify the first
    // quartile of integer variables that are closest to an integer value
    // in the relaxation, fix them to the nearest integer, and repeat.

    for (int iter = 0; iter < 1000; ++iter)
    {

      // create a list of fractional variables, sorted in order of
      // increasing distance from the relaxation solution to the nearest
      // integer value

      deque<GRBVar*> fractional;
      for (size_t j = 0; j < intvars.size(); ++j)
      {
        double sol = fabs(intvars[j]->get(GRB_DoubleAttr_X));
        if (fabs(sol - floor(sol + 0.5)) > 1e-5)
        {
          fractional.push_back(intvars[j]);
        }
      }

      cout << "Iteration " << iter << ", obj " <<
      model.get(GRB_DoubleAttr_ObjVal) << ", fractional " <<
      fractional.size() << endl;

      if (fractional.size() == 0)
      {
        cout << "Found feasible solution - objective " <<
        model.get(GRB_DoubleAttr_ObjVal) << endl;
        break;
      }

      // Fix the first quartile to the nearest integer value
      sort(fractional.begin(), fractional.end(), vcomp);
      int nfix = fractional.size() / 4;
      nfix = (nfix > 1) ? nfix : 1;
      for (int i = 0; i < nfix; ++i)
      {
        GRBVar* v = fractional[i];
        double fixval = floor(v->get(GRB_DoubleAttr_X) + 0.5);
        v->set(GRB_DoubleAttr_LB, fixval);
        v->set(GRB_DoubleAttr_UB, fixval);
        cout << "  Fix " << v->get(GRB_StringAttr_VarName) << " to " <<
        fixval << " ( rel " << v->get(GRB_DoubleAttr_X) << " )" <<
        endl;
      }

      model.optimize();

      // Check optimization result

      if (model.get(GRB_IntAttr_Status) != GRB_OPTIMAL)
      {
        cout << "Relaxation is infeasible" << endl;
        break;
      }
    }

  }
  catch (GRBException e)
  {
    cout << "Error code = " << e.getErrorCode() << endl;
    cout << e.getMessage() << endl;
  }
  catch (...)
  {
    cout << "Error during optimization" << endl;
  }

  delete[] x;
  delete env;
  return 0;
}


bool vcomp(GRBVar* v1,
           GRBVar* v2) throw(GRBException)
{
  double sol1 = fabs(v1->get(GRB_DoubleAttr_X));
  double sol2 = fabs(v2->get(GRB_DoubleAttr_X));
  double frac1 = fabs(sol1 - floor(sol1 + 0.5));
  double frac2 = fabs(sol2 - floor(sol2 + 0.5));
  return (frac1 < frac2);
}

Try Gurobi for Free

Choose the evaluation license that fits you best, and start working with our Expert Team for technical guidance and support.

Evaluation License
Get a free, full-featured license of the Gurobi Optimizer to experience the performance, support, benchmarking and tuning services we provide as part of our product offering.
Academic License
Gurobi supports the teaching and use of optimization within academic institutions. We offer free, full-featured copies of Gurobi for use in class, and for research.
Cloud Trial

Request free trial hours, so you can see how quickly and easily a model can be solved on the cloud.

Search