Originally posted by: SystemAdmin

Hi,
I try to model an “Exist” constraint that we haven’t in opl. For example how to implement in opl the following constraint?

dvar 

int x[1..10] in 1..10; subject to
{ 
//(exists i in 1..10, exists j in i+1..10) 
//    x[i]==x[j]; 
}

Thanks in advance

NL
#ConstraintProgramming-General
#DecisionOptimization

Originally posted by: Didier Vidal

Hi,

I think that an aggregated or, available in CP, does exactly what you need.
Do other users have a different approach to implement this constraint ?

using CP;

dvar int x[1..10] in 1..10;
maximize sum(i in 1..10) (x[i] == i);
subject to{
or (ordered i,j in 1..10) (x[i]==x[j]) == 1;
//(exists i in 1..10, exists j in i+1..10)
// x[i]==x[j];
}

Didier.
#ConstraintProgramming-General
#DecisionOptimization

Originally posted by: SystemAdmin

thanks Diddier,
it's what I searched, thank you again.

NL
#ConstraintProgramming-General
#DecisionOptimization

Originally posted by: SystemAdmin

A slightly simpler, and probably more efficient model is found below.

{code]
using CP;

dvar int xhttp://1..10 in 1..10;
maximize sum(i in 1..10) (x[i] == i);
subject to{
or(ordered i,j in 1..10) (x[i]==x[j]);
}
code
#ConstraintProgramming-General
#DecisionOptimization

Originally posted by: SystemAdmin

The previous post is unreadable. here it is again.

using CP;   dvar 

int x[1..10] in 1..10; maximize sum(i in 1..10) (x[i] == i); subject to
{ or(ordered i,j in 1..10) (x[i]==x[j]); 
}

#ConstraintProgramming-General
#DecisionOptimization

Originally posted by: SystemAdmin

Hi every body,
can we have a discussion about the difference between:

or(ordered i,j in 1..10) (x[i]==x[j]);

and:

or (ordered i,j in 1..10) (x[i]==x[j]) == 1;

is it really more efficient ?

NL
#ConstraintProgramming-General
#DecisionOptimization