Originally posted by: orange2005

Dear all,

I have a difficulty for coding a constraint.
I have IloNumExpr A, IloNumVarArray x, and IloBoolVarArray y.

I want to code the below constraint:

if y[i][j]==1 then A+=x[i][j].

anyone knows how to code it in Cplex with c++?

Thanks in advance!
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: RafaelMartinelli

Notice that += is not a constraint.
What do you want to do with the expression A?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: orange2005

the final constraint is
model.add(A<=1)
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Do I understand correctly that you want to state this:

if y[i][j]==1
  then A + x[i][j] <= 1
  else A <= 1

where x and y are variables.
It should be possible to formulate that using IloIfThen.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: orange2005

Hi, Daniel,

I want
for (i=0; i<I; i++)
{
for (j=0; j<J; j++)
{
if (y[i][j]==1)
A+=x[i][j];
}
}
model.add(A<=1);

It seems that I cannot use IloIfThen(env,(y[i][j]==1),(A+=x[i][j]));

Could you please help me out?

Thank you!
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Is y a parameter or a variable?
What is the range of +x+? Is it a boolean variable? Is it non-negative?
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: orange2005

y is Bool variable
x is real number variable, its range is 0,1
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

I can think of different ways to implement that.
First:

model.add(A <= 1); // This must always be satisfied.
model.add(IloIfThen(env, y[i][j] == 1, A + x[i][j] <= 1)); // More restrictive if y[i][j] == 1

Second (using a helper variable):

IloNumVar b(env, 0, 1, "b");
model.add(A <= b);
model.add(IloIfThen(env, y[i][j] == 1, b <= 1 - x[i][j]));

Third (turning your problem into a QCP):

model.add(A + y[i][j] * x[i][j] <= 1);

#CPLEXOptimizers
#DecisionOptimization

Originally posted by: orange2005

Thanks a lot!
#CPLEXOptimizers
#DecisionOptimization