Originally posted by: R9UA_Ishan_Mukherjee

I am very new to using CPLEX and am having trouble figuring out how to sum over columns of a variable array for a constraint.

This is what I did for rows, which appears to work:

for(int i=0; i<size; ++i)
{
model.add(IloSum(variables[i]) == 1);
}

Where variables is a 2 dimensional array. I want to do model.add(variables[0][0] + variables[1][0] + variables[2][0] + ... == 1)
model.add(variables[0][1] + variables[1][1] + variables[2][1] + ... == 1)

But I want it to loop so I can have something like the for loop above. How can I do that?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: R9UA_Ishan_Mukherjee

I came up with this but I want a cleaner way:

IloExpr onePerson(env);
for(int j=0; j<size; ++j)
{
for(int i=0; i<size; ++i)
{
onePerson += variables[i][j];
}
model.add(onePerson == j+1);
}

Specifically I don't like how I constantly add onePerson where and never set it back to 0. When I do onePerson = 0; at the top of the loop I run into problems. Is there another way to clear an expression?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Why not instantiate IloExpr inside the loop:

for(int j=0; j<size; ++j) {
   IloExpr onePerson(env);
   for(int i=0; i<size; ++i) {
      onePerson += variables[i][j];
   }
   model.add(onePerson == j+1);
}

Wouldn't that give you what you need?
I cannot think of a cleaner way to create your sum. Maybe you can do something similar to std::for_each but that sounds like overkill here.
#CPLEXOptimizers
#DecisionOptimization