Originally posted by: SystemAdmin

code
IloNumVarArray _vars(env);
IloNumArray _vals(env);
for(unsigned int id=0;id<C.size();id++)
{

z.add(IloIntVar(env,0,1,var_name));

if ((id >= _first) && (id < _last))
{
_vars.add(zid);
_vals.add(1);
}

.....
}
cplex.addMIPStart(_vars,_vals);
_vals.end();
_vars.end();
[/code]

It displays "CPLEX Error 3003: Not a mixed-integer problem" when I add cplex.addMIPStart(_vars,_vals); (CPLEX12.1)
otherwise, code works. z is an IloIntVarArray.

Thanks for your comments
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

Does it help to do

cplex.add(IloIntVar(env, 0, 0, 
"dummy"));

before trying to add a MIP start?
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

When I added this line, then it displayed "segmentation fault".
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

Sorry, that should of course have been

model.add(IloIntVar(env, 0, 0, 
"dummy"));

Does this help?

Another question is: It seems like you don't have any integral variables in your model. So why do you want to add a MIP start?
Or will you add constraints with integer variables after the MIP start? If so then adding the constraints before adding the MIP starts should also solve the problem.
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

I put addMIPStart in the final line of the procedure. Before i already used this function.

z contains integral variables : z is an IloIntVarArray
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

I have no idea what might be going wrong.
Could you export your model to a SAV file (cplex.exportModel("model.sav")) right before you invoke cplex.addMIPStart() and attach the model here? If you cannot do that, what do cplex.getNbinVars() and cplex.getNintVars() return right before cplex.addMIPstart()?
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

Perhaps a silly question, but does z appear in the constraints or objective of your model?

Paul

Mathematicians are like Frenchmen: whenever you say something to them, they translate it into their own language, and at once it is something entirely different. (Goethe)
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: SystemAdmin

z appears in the objective function and constraints.
#DecisionOptimization
#MathematicalProgramming-General

Originally posted by: HamidMelb

 

Hi,

I have a problem with warm start of an LP using CPLEX 12.7 (C++). I think it is related to this forum topic. I have a linear programming network model. I want to start with a solution. CPLEX throws an error: "CPLEX 3003: not a mixed integer problem." Cplex recognise my model correctly, but why an LP cannot have a start solution?

To somehow fix it, I added a dummy integer variable and a dummy  constraint. Then I receive this error: "CPLEX 1017: not available for mixed integer problems." Can you give me hint please?

 

   IloNumVar dummy(env_subp,0,0,ILOINT)
   model.add( dummy ==0 );

...

    for(int k=0; k<N; k++) {
        startVar.add(R[k]);
        startVal.add(Rsoln[k]);
        }
    cpx_subp.addMIPStart(startVar, startVal);

 

#DecisionOptimization
#MathematicalProgramming-General

As the name of the function suggests, a MIP start is only for MIP.

If you want to provide a starting point for a continuous problem use any of IloCplex::setBasisStatuses() or IloCplex::setStart().

#DecisionOptimization
#MathematicalProgramming-General