Originally posted by: SystemAdmin

Hi
I am trying to solve a MIP by using Lagrangian decomposition. To implement the lagrangian decomposition I have to create four model. Part of the main Pseodocode could be look like this:
try
{
Create model1; //LP relaxation of generic model
Solve.model1;
Create model2; // model based on generic model to calculate upper bound
Solve.model2;
Create model3.// subproblem1
Solve.model3;
Create.model4// subproblem 2
}
catch(IloException& e) {
cerr << " ERROR: " << e << endl;
}
catch(...) {
cerr << " ERROR" << endl;
}

env.end();

return 0;
}

I can solve each model individually within a minuet for my dataset but I can't solve all these model together. Since I have to use output of model as input to another model, I need to run all these models together. The cplex error says “out of memory”.
Can anyone please tell me what could be reason behind it?
Thanks in advance.

Roni
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

How much memory does each IloModel consume when you construct it? You can check free memory before and immediately after constructing a model to figure out the memory footprint. If you don't have enough memory to construct all four at once, there's probably not much you can do about it.

You can try applying the end() method to each IloCplex instance before extracting the next model (after getting the solution, of course), while keeping the IloModels alive. If this is in a loop, the cost is that you will have to spend time extracting each model in each pass of the loop (so you'll be trading time for memory).

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)
#CPLEXOptimizers
#DecisionOptimization