Originally posted by: Bahman

Hello Everyone,

 

I have coded and ran my model in CPLEX OPL API, and It works fine. I am migrating to JAVA API. I had defined a tuple in CPLEX STUDIO (in OP:I) and I do not know if I need to define a tuple3 and List<Tuple>,, or a separate class to write the set of tuples (representing arcs).I do not know what is the best way to define the arcs that I had defined them in opl in tuples, and how to make use of that as a dimension of my decision variable X, in JAVA API. A  MWE  is presented as follows:

 

{int} Y=asSet(1..7);
{int} T=asSet(1..10);
{int} K=asSet(1..42);

int b[Y][K][G]=...;
{string} G=...; 

 

 

Then one of the decision Variable: 

dvar int+ X[Slegs][K][G][T];

 

and one of the constraints (flow conservation constraint in particular):

 

 I would appreciate if you could help me out to define the tuple, decision variable and the the set set of constraint written above as flow conservation constraints. 

 

Regards,
Bahman

#CPLEXOptimizers
#DecisionOptimization

One way to implement this would be

1. Represent your tuple as a class Sleg. This class would just have 3 public fields: i, j, l.
2. Represent the set 'Slegs' as an array of instances of Slegs (maybe first collect into a List<> during creation and then use toArray())
3. The dimension of the decision variable is then given by Slegs.length
4. An index 0 <= k < Sleg.length gives to you things: the tuple Slegs[k] as well as the index to find the decision variable indexed by <i,j,l>.

 

#CPLEXOptimizers
#DecisionOptimization

Originally posted by: Bahman

Thanks Daniel, 

 

Which is better: to define a Map like this: <Integer, Slegs> map = new HashMap<Integer, Slegs>;  and use the Integer as the key to use for iteration. or .ToArray as you suggested above?

 

I would really appreciate if you could show me how I would instantiate the Sleg and add tuples into that like this: Slegs {<1,2,500>, <2,3,450>, <3,4,375>, <2,5,560>, <5,7,448>}

int u[sL.Length][G]  = ... for instance: u[<1,2,500>][1] = 300; u[<1,2,500>][2] = 80; and etc.

 

Regards
Bahman

#CPLEXOptimizers
#DecisionOptimization

Using a map is only useful if the index set is sparse (which in your case I think it is not), otherwise using an array (or a java.util.Vector) is more efficient.

For the other question I am not sure I understand: are you asking how to instantiate a class in Java?

#CPLEXOptimizers
#DecisionOptimization