Originally posted by: SystemAdmin

Hello everyone,

I'm coding a problem in CP Optimizer, and it returns the error: "empty handle"!:

IloCP: empty handle
testCplex: /opt/ibm/ILOG/CPLEX_Studio125/cpoptimizer/include/ilcp/cpext.h:852: void IlcCPORevInt::setValue(IloCP, IlcInt): Assertion `(cp.getImpl() != 0) || (IlcBacktrace(), std:: cerr << ("IloCP: empty handle") << std:: endl, ilc_stop_assert())' failed.

I debugged the code, and it returned me this error http://at if (cp.solve()) :
Single stepping until exit from function siglongjmp,
which has no line number information.
Warning:
Cannot insert breakpoint 0.
Error accessing memory address 0x55a6f7f7: Input/output error.

and in the next line:

IloCP: empty handle
testCplex: /opt/ibm/ILOG/CPLEX_Studio125/cpoptimizer/include/ilcp/cpext.h:852: void IlcCPORevInt::setValue(IloCP, IlcInt): Assertion `(cp.getImpl() != 0) || (IlcBacktrace(), std:: cerr << ("IloCP: empty handle") << std:: endl, ilc_stop_assert())' failed.

IloCP: empty handle
testCplex: /opt/ibm/ILOG/CPLEX_Studio125/cpoptimizer/include/ilcp/cpext.h:852: void IlcCPORevInt::setValue(IloCP, IlcInt): Assertion `(cp.getImpl() != 0) || (IlcBacktrace(), std:: cerr << ("IloCP: empty handle") << std:: endl, ilc_stop_assert())' failed.
Could you help me figure out what this error means and how I can correct it?

Thank you.
#CPOptimizer
#DecisionOptimization

Originally posted by: ChrisBr

Hello,

The "empty handle" error means that you are trying to use a handle with no implementation.

see: http://pic.dhe.ibm.com/infocenter/cosinfoc/v12r5/index.jsp
CP Optimizer > CP Optimizer C++ API Reference Manual > Concepts > Handle class

For example:

IloCP cp;

declares a empty IloCP handle

whereas

IloCP cp(env);

creates the right implementation attached to the cp handle.
Nevertheless, if you don't use any feature of CP Optimizer extensions (CP Optimizer > CP Optimizer C++ API Reference Manual > optim.cpoptimizer.extensions) it wouldn't be so obvious for you to find the guilty call (because the error message mentions IlcCPORevInt which is one of these extensions).
In this case, it could help if you post your model, as we would be able to help you better.

Regards,

Chris.
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

Thanks for the reply.

I do use IloCP cp as an object's argument, and

In the code, I had defined some constant to which no value had been assigned. I corrected this mistake, and now I returns no empty class error. However, it reports infeasible! Surprisingly, it says "0 Variables", while I do have some variables.

Please see what it returns in the console:

!

---

! Satisfiability problem - 0 variables, 20 constraints
! Problem found infeasible at the root node
!

---

! Search terminated normally, model has no solution.
! Number of branches : 0
! Number of fails : 3
! Total memory usage : 327.2 kB (316.4 kB CP Optimizer + 10.8 kB Concert)
! Time spent in solve : 0.00s (0.00s engine + 0.00s extraction)
! Search speed (br. / s) : 0
Please also see the propagate() as a public member function of the constraint class NoSubtour, where I use the empty handle IloCP cp:

void NoSubtourI::propagate(){
IloInt FirstNode;
IloInt EndNode;
IloCP cp;
for (IloInt i=0; i < successor.getSize()-2; i++){

if (successor[i].isFixed()){
FirstNode=first[i];
first[successorhttp://i].getValue().setValue(cp, FirstNode);

EndNode=end[successorhttp://i].getValue();
end[i].setValue(cp, EndNode);

successorEndNode.removeValue(FirstNode);

}
}

}

Well, I had to use IloCP cp there because otherwise it could not identify getValue() and setValue().

Thanks.
#CPOptimizer
#DecisionOptimization

Originally posted by: ChrisBr

Hello,

I'm not sure of what you want to do...

Nevertheless, I'll try to fix the problem in your "propagate" method.
Assuming that "first" and "end" are 2 arrays of IlcRevInt (is this right?), I think your code should look like:

void NoSubtourI::propagate()
{ IloInt FirstNode; IloInt EndNode; 

for (IloInt i=0; i < successor.getSize()-2; i++)
{ 

if (successor[i].isFixed())
{ FirstNode=first[i].getValue(); first[successor[i].getValue()].setValue(getCP(), FirstNode); EndNode=end[successor[i].getValue()].getValue(); end[i].setValue(getCP(), EndNode); successor[EndNode].removeValue(FirstNode); 
} 
} 
}

The message

! Satisfiability problem - 0 variables, 20 constraints
! Problem found infeasible at the root node

indicates that the infeasibility has been found during the presolve
(don't worry about the "0 variables" report in this case).

The fact that your model has been found infeasible might be related to other mistakes.

I hope this helps,

Chris.
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

Thanks a lot.

I did as you said. Also, Id' made some mistakes in data extraction which made the problem infeasible. Vow it perfectly works! ;)
#CPOptimizer
#DecisionOptimization