Originally posted by: CplexUser1453

Hello

I am dealing with a problem where one of the decision variables looks like a fairly sparse 3D array. Thanks to tuples I was able to write a fairly elegant OPL model. However now I have trouble accessing my representation through the JAVA API.

Here is my OPL representation :

tuple TripleIndex{
int t;
int u;
int g;
}

{TripleIndex} Index = {<t,u,g> | t in Time, u in Urange, g in 1..Garray[t][u]};

dvar boolean Xs in Index;

Accesing my X variable in OPL scripting is easy :

Xhttp://Index.find(t,u,g) (for a given t,u,g)

Using the Java API I am capable of getting X and TripleIndex :

IloTupleSchema Tindex = opl.getElement("TripleIndex").asTupleSchema();
IloNumMap Xmap = opl.getElement("X").asNumMap();

Here are my questions :
1) How do I get my "Index" tuple from opl?
2) Once I have all three elements, how do I do the equivalent of Xhttp://Index.find(t,u,g)?

Thanks a lot!
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: CplexUser1453

sorry the formatting is not right. Here is the message with good formatting :

I am dealing with a problem where one of the decision variables looks like a fairly sparse 3D array. Thanks to tuples I was able to write a fairly elegant OPL model. However now I have trouble accessing my representation through the JAVA API.

Here is my OPL representation :

tuple TripleIndex
{ 

int t; 

int u; 

int g; 
}   
{TripleIndex
} Index = 
{<t,u,g> | t in Time, u in Urange, g in 1..Garray[t][u]
};   dvar 

boolean X[s in Index];

Accesing my X variable in OPL scripting is easy :

X[Index.find(t,u,g)] (

for a given t,u,g)

Using the Java API I am capable of getting X and TripleIndex :

IloTupleSchema Tindex = opl.getElement("TripleIndex").asTupleSchema();
IloNumMap Xmap = opl.getElement("X").asNumMap();

Here are my questions :
1) How do I get my "Index" tuple from opl?
2) Once I have all three elements, how do I do the equivalent of

X[Index.find(t,u,g)] (

for a given t,u,g)

Thanks a lot!
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: CplexUser1453

Found the answer so let me repond to my question myself (as I have not found a similar topic on the forum and I think this is an example)

So, it seems to me that with my OPL model defined as

tuple TripleIndex{
int t;
int u;
int g;
}
 
{TripleIndex} Index = {<t,u,g> | t in Time, u in Urange, g in 1..Garray[t][u]};
 
dvar boolean X[s in Index];

the equivalent of

X[Index.find(t,u,g)]

is :

IloNumMap Xmap = opl.getElement("X").asNumMap();
IloTupleBuffer buf = opl.getElement("Index").asTupleSet().makeTupleBuffer(-1);
buf.setIntValue("t", tt);
buf.setIntValue("u", uu);
buf.setIntValue("g", gg);
buf.commit();
System.out.println(Xmap.get(index.find(buf)));

This seems to work for me...
#DecisionOptimization
#OPLusingCPLEXOptimizer