Try our new documentation site (beta).


mip2_cs.cs


/* Copyright 2020, Gurobi Optimization, LLC */

/* This example reads a MIP model from a file, solves it and
   prints the objective values from all feasible solutions
   generated while solving the MIP. Then it creates the fixed
   model and solves that model. */

using System;
using Gurobi;

class mip2_cs
{
  static void Main(string[] args)
  {
    if (args.Length < 1) {
      Console.Out.WriteLine("Usage: mip2_cs filename");
      return;
    }

    try {
      GRBEnv    env   = new GRBEnv();
      GRBModel  model = new GRBModel(env, args[0]);
      if (model.IsMIP == 0) {
        Console.WriteLine("Model is not a MIP");
        return;
      }

      model.Optimize();

      int optimstatus = model.Status;
      double objval = 0;
      if (optimstatus == GRB.Status.OPTIMAL) {
        objval = model.ObjVal;
        Console.WriteLine("Optimal objective: " + objval);
      } else if (optimstatus == GRB.Status.INF_OR_UNBD) {
        Console.WriteLine("Model is infeasible or unbounded");
        return;
      } else if (optimstatus == GRB.Status.INFEASIBLE) {
        Console.WriteLine("Model is infeasible");
        return;
      } else if (optimstatus == GRB.Status.UNBOUNDED) {
        Console.WriteLine("Model is unbounded");
        return;
      } else {
        Console.WriteLine("Optimization was stopped with status = "
                           + optimstatus);
        return;
      }

      /* Iterate over the solutions and compute the objectives */

      GRBVar[] vars = model.GetVars();
      model.Parameters.OutputFlag = 0;

      Console.WriteLine();
      for (int k = 0; k < model.SolCount; ++k) {
        model.Parameters.SolutionNumber = k;
        double objn = 0.0;

        for (int j = 0; j < vars.Length; j++) {
          objn += vars[j].Obj * vars[j].Xn;
        }

        Console.WriteLine("Solution " + k + " has objective: " + objn);
      }
      Console.WriteLine();
      model.Parameters.OutputFlag = 1;

      /* Create a fixed model, turn off presolve and solve */

      GRBModel fixedmodel = model.FixedModel();

      fixedmodel.Parameters.Presolve = 0;

      fixedmodel.Optimize();

      int foptimstatus = fixedmodel.Status;

      if (foptimstatus != GRB.Status.OPTIMAL) {
        Console.WriteLine("Error: fixed model isn't optimal");
        return;
      }

      double fobjval = fixedmodel.ObjVal;

      if (Math.Abs(fobjval - objval) > 1.0e-6 * (1.0 + Math.Abs(objval))) {
        Console.WriteLine("Error: objective values are different");
        return;
      }

      GRBVar[] fvars  = fixedmodel.GetVars();
      double[] x      = fixedmodel.Get(GRB.DoubleAttr.X, fvars);
      string[] vnames = fixedmodel.Get(GRB.StringAttr.VarName, fvars);

      for (int j = 0; j < fvars.Length; j++) {
        if (x[j] != 0.0) Console.WriteLine(vnames[j] + " " + x[j]);
      }

      // Dispose of models and env
      fixedmodel.Dispose();
      model.Dispose();
      env.Dispose();

    } catch (GRBException e) {
      Console.WriteLine("Error code: " + e.ErrorCode + ". " + e.Message);
    }
  }
}

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

Gurobi Optimization