Solve an LP problem in C++ using Gurobi

Recently, I used Gurobi to solve an flow-based LP problem, in which the c++ interfaces were called in my project.

I packed the relevant c++ interfaces into a class ( class Cdtn_flow; ), where the key function is illustrated as follows:

void Cdtn_flow::try_to_solve_the_model( void )
{
    try
    {
        // ============================= 0. Find all the possible flows.
        find_all_possible_flows();

        // ============================= 1. Add variables into Model.
        // ---- Print-Test:
        cout << "\t--- Before adding vars into model, "
            << "\n\tthe size of Bus_set: " << m_BIDSet.size()
            << "\n\tthe size of Hlp_set: " << m_HIDSet.size()
            << "\n\tthe size of Sub_set: " << m_SIDSet.size()
            << "\n\tthe size of allArcs: " << m_allArcs.size()
            << endl;

        add_vars_into_model();

        // ============================= 2. Set the objective of the model.
        // The  objective  is to minimize the sum of flows sourced from S-node.
        setObjective_of_the_model( GRB_MINIMIZE );

        // ============================= 3. Add Constraints.
        addConstrs();

        // ============================= 4. Solve the model.
        solve_the_model();

        // ============================= 5. print Solution.
        print_solution();
    }
    catch (GRBException e)
    {
        cout << "Error code = " << e.getErrorCode() << endl;
        cout << e.getMessage() << endl;

        LogFile::instance()->m_of_opt_result
		<< "Error code = " << e.getErrorCode() << endl;
        LogFile::instance()->m_of_opt_result << e.getMessage() << endl;
    }
    catch (...)
    {
        cout << "Exception during optimization" << endl;
        LogFile::instance()->m_of_opt_result
	<< "Exception during optimization" << endl;
    }
}

From this function, we can see the logic to solve an optimization.

Within which, there is a big error I made in the function setObjective_of_the_model( GRB_MINIMIZE ).  This error is shown as:

void Cdtn_flow::setObjective_of_the_model( int nModel_sense )
{
    GRBLinExpr sumOptimal = 0;

    // ---ERROR-->| Index should be within both 1~10( Bus_IDs ) and 15~17( SUb_IDs ).
    //for ( int j = 1; j <= m_nCNT_Flow_su; j++ )
    //{
    //    sumOptimal += flow_su[j];
    //}

    // == RIGHT -->| Index is within both 1~10( Bus_IDs ) + 15~17( SUb_IDs ).
    for ( vector<int>::iterator Bit = m_BIDSet.begin(); Bit != m_BIDSet.end(); ++Bit )
    {
        int nBusID = *Bit;
        sumOptimal += flow_su[ nBusID ];
    }

    for ( vector<int>::iterator it = m_SIDSet.begin(); it != m_SIDSet.end(); ++it )
    {
        int nSubID = *it;
        sumOptimal += flow_su[ nSubID ];
    }

    model.setObjective( sumOptimal, nModel_sense );
    model.update();
}

 The mistake is caused by misunderstanding of the index of the obj_flow_sum. As a result, it always shows the Error hints in console window:

Error Code: 20001
Variable not in model.

 I find that, the exception is always thrown at the line 24 of the function setObjective_of_the_model(..).

That is because, the variable flow_su[idx] was calculated in a wrong way, it can not be added into the model of course.

// --- Davy_H

// --- 2013-02-14 :~