Originally posted by: bindasmariner

Dear all

I am experimenting with a Column generation assignment in ILOG OPL. I want to terminate the sub-problem which is an MIP, when it reaches a certain MIP tolerance gap (say 10%). I tried following code in the body of the sub-problem in the pre-processing part, but it is not working.Although it works when I run the sub-problem separately.

execute PARAM{
cplex.epagap = 1;
}

Please suggest the workable way.

Best Regards
Saurabh
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: harshvnit

Are you loading and calling your sub-problem model file into the scripting code. If yes, then you should keep the 'cplex.epgap' statement in your subproblem model file. That should serve the purpose of solving the subproblem and stopping it with a 10% gap and giving you a best integer solution.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: bindasmariner

Thanks for the reply. This thing works the way you said. Although is it possible to change the CPLEX parameters from the main block for a model, rather than by writing within the MIP model itself?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

To set a CPLEX parameter by script (in a main block for example), you would use a statement like:

cplex.theParam = theValue;

In the documentation, check "IDE and OPL > Optimization Programming Language (OPL) > Parameters and Settings in OPL > Mathematical programming options > CPLEX parameters and OPL parameters" to get the list of parameters that cannot be set by scripting.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: harshvnit

If you have your sub problem defined like this, you can change the parameter by m2Cplex.epagap, m2Cplex.tilim etc.

var m2Source = new IloOplModelSource("uPMSP_LowerBound3.mod");
  var m2Cplex = new IloCplex();
  var m2Def = new IloOplModelDefinition(m2Source);
   
 var m2_init = new IloOplModel(m2Def,m2Cplex);
  m2_init.addDataSource(data);
 
m2.generate();
m2Cplex.tilim=30;
m2Cplex.solve();

In this example it will start solving the uPMSP_LowerBound3.mod and stop after 30 seconds
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

Dear all,
I have a problem to stop my OPL running after it gives the solution. the same model stops after finding the optimal but for the large size model although it finds the optimal and also gives the solution, it is still running.
Any help or idea this kind of problem.
Thanks, in advance.
Erhan
#DecisionOptimization
#OPLusingCPLEXOptimizer