Hello,

I am currently migrating the optimization model from CPLEX OPL to CPLEX Concert Technology + Java.

In Java, I would like to define the decision variables based on tuple structure. Would you please help me how to define it?

For example, I have a decision variable in OPL as below.

tuple INDEX{
      int employeeID;
      int offeringID;
}
{INDEX} indexes {<e, o> | e in employees, o in offerings}
dvar int+ schedule [indexes] in 0..1;

In order to define the decision variable (X) in Java, I first defined class data structure.
class INDEX{
       int employeeID;
       int offeringID;
}
List<INDEX> indexes = new ArrayList<INDEX>();

IloNumVar[] schedule = cplex.numVarArray(indexes.size(), 0, 1, IloNumVarType.Int);

I would like to have the constraint ( cplex.addEQ(cplex.sum(schedule[indexes]), 1) ), sum of schedule over indexes has to be equal to 1.

The above cplex.addEQ function does not work. Would you please help me how to defined tuple-based decision variables and constraints in Java?

Thank you for your help!

------------------------------
Sung Hwang
------------------------------

#DecisionOptimization

If you are summing overall index values, you can just use cplex.sum(schedule). If you are summing only for certain indices, you probably will want to use a loop.

IloLinearNumExpr expr = cplex.linearNumExpr();
for (int i = 0; i < indexes.size(); i++) {
  if (...) {
    expr.addTerm(1.0, schedule[i]);
  }
}
cplex.addEq(expr, 1.0);

where ... is a logical test to determine if the index is included in the summation.

As a side note, to facilitate moving between the integer index of schedule[] and your INDEX tuples, you might want to use either a Map<INDEX, Integer> or a Map<INDEX, IloNumVar> (and maybe a Map<IloNumVar, INDEX> in the opposite direction).

------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
------------------------------

Original Message:
Sent: Thu June 03, 2021 07:53 AM
From: Sung Hwang
Subject: Tuple Index in Java & CPLEX Concert

Hello,

I am currently migrating the optimization model from CPLEX OPL to CPLEX Concert Technology + Java.

In Java, I would like to define the decision variables based on tuple structure. Would you please help me how to define it?

For example, I have a decision variable in OPL as below.

tuple INDEX{
      int employeeID;
      int offeringID;
}
{INDEX} indexes {<e, o> | e in employees, o in offerings}
dvar int+ schedule [indexes] in 0..1;

In order to define the decision variable (X) in Java, I first defined class data structure.
class INDEX{
       int employeeID;
       int offeringID;
}
List<INDEX> indexes = new ArrayList<INDEX>();

IloNumVar[] schedule = cplex.numVarArray(indexes.size(), 0, 1, IloNumVarType.Int);

I would like to have the constraint ( cplex.addEQ(cplex.sum(schedule[indexes]), 1) ), sum of schedule over indexes has to be equal to 1.

The above cplex.addEQ function does not work. Would you please help me how to defined tuple-based decision variables and constraints in Java?

Thank you for your help!

------------------------------
Sung Hwang
------------------------------

#DecisionOptimization

So I can't define a decision variable using tuple (class structure data) in Java.

Thank you Dr. Rubin for your comment!


------------------------------
Sung Hwang
------------------------------

Original Message:
Sent: Thu June 03, 2021 03:19 PM
From: Paul Rubin
Subject: Tuple Index in Java & CPLEX Concert

If you are summing overall index values, you can just use cplex.sum(schedule). If you are summing only for certain indices, you probably will want to use a loop.

IloLinearNumExpr expr = cplex.linearNumExpr();
for (int i = 0; i < indexes.size(); i++) {
  if (...) {
    expr.addTerm(1.0, schedule[i]);
  }
}
cplex.addEq(expr, 1.0);

where ... is a logical test to determine if the index is included in the summation.

As a side note, to facilitate moving between the integer index of schedule[] and your INDEX tuples, you might want to use either a Map<INDEX, Integer> or a Map<INDEX, IloNumVar> (and maybe a Map<IloNumVar, INDEX> in the opposite direction).

------------------------------
Paul Rubin
Professor Emeritus
Michigan State University

Original Message:
Sent: Thu June 03, 2021 07:53 AM
From: Sung Hwang
Subject: Tuple Index in Java & CPLEX Concert

Hello,

I am currently migrating the optimization model from CPLEX OPL to CPLEX Concert Technology + Java.

In Java, I would like to define the decision variables based on tuple structure. Would you please help me how to define it?

For example, I have a decision variable in OPL as below.

tuple INDEX{
      int employeeID;
      int offeringID;
}
{INDEX} indexes {<e, o> | e in employees, o in offerings}
dvar int+ schedule [indexes] in 0..1;

In order to define the decision variable (X) in Java, I first defined class data structure.
class INDEX{
       int employeeID;
       int offeringID;
}
List<INDEX> indexes = new ArrayList<INDEX>();

IloNumVar[] schedule = cplex.numVarArray(indexes.size(), 0, 1, IloNumVarType.Int);

I would like to have the constraint ( cplex.addEQ(cplex.sum(schedule[indexes]), 1) ), sum of schedule over indexes has to be equal to 1.

The above cplex.addEQ function does not work. Would you please help me how to defined tuple-based decision variables and constraints in Java?

Thank you for your help!

------------------------------
Sung Hwang
------------------------------

#DecisionOptimization