Originally posted by: harrynick

Hello,

Can anyone tell me how to implement 'complement sets' in IBM OPL CPLEX?

Suppose, I have a set named 'Set' as :

{int} Set = {1,2};

Now I have a variable say Apple[Set][Setd].

Here Setd is the complement set of set 'Set'. In other words, if 'Set' takes value 1 then 'Setd' should take value 2(the complement). Also when the value in 'Set' is 2 value in 'Setd' is 1. 

I hope someone shade some light on this problem.

Thanks and Regards.

Hi,

have you had a look at diff ?

{int} Set = {1,2};

{int} a={1};
{int} b=Set diff a;

execute
{
writeln(b);
}

gives

{2}

regards

Originally posted by: harrynick

Thanks for shedding light on the problem.

But suppose I do not know about 'set a = {1}' as mentioned while implementing in a constraint. So how will I implement it?

As for an example:

{int} Set = {1,2};

forall( i in Set)  sum(j in Setd) Apple[Set][Setd]<=Mango[Any1][Any2]; (Here Mango is any other variable with Any1 and Any2 as two random sets)

Here Setd is the complement set of set 'Set'.

Thanks and Regards.

Hi,

{int} Set = {1,2};

{int} a[i in 1..10]=(rand(2)+i div i-1==0)?{1}:{2};;
{int} b[i in 1..10]=Set diff a[i];

execute
{
writeln(b);
}
 
 
 int su=sum(i in 1..10) sum(j in Set diff a[i]) j;
 
 execute
 {
 su;
 }
 
 assert su==sum(i in 1..10) sum(j in b[i]) j;

regards

Originally posted by: harrynick

Thanks and Regards.

Originally posted by: harrynick

Thanks and Regards.