Originally posted by: elec

Hi,

Does there exist a nvalues global constraint (http://www.emn.fr/z-info/sdemasse/gccat/Cnvalues.html) in cp optimizer? I can't seem to find one in the documentation, but maybe i am overlooking something.

Best regards,

E.
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

Hello,
The nvalues global constraint does not exists in CP Optimizer. You can model it either with a sum of positive count expressions or a sum of or constraints as illustrated in the OPL model below:

range Values = 0..10; 

int n = 20;   dvar 

int x[i in 1..n] in Values; dvar 

int y in 1..n;   constraints 
{ y == sum(v in Values) (0<count(x,v)); 
// alternative model: 
// y == sum(v in Values) ( or(i in 1..n) x[i]==v ); 
}

Philippe
#CPOptimizer
#DecisionOptimization

Originally posted by: elec

Hi Phillipe,

Thanks for the fast response. However, i'm not sure i fully understand the alternatives you provided. I could be mistaken, but i don't think that those constraints do the same thing as the nvalues (i.e. count the number of distinct values assigned to a set of variables.) I believe the posted constraints do something like count how many times in total each distinct value is assigned.

For example, consider the following assignment:
x0 = 1;
x1 = 1;
x2 = 2;
x3 = 3;

The proposed constraints will have y=4, but nvalues will have y=3.

Am i correct to think this?
#CPOptimizer
#DecisionOptimization

Originally posted by: SystemAdmin

Hello,
The formulation I gave is correct, it just computes for each value a boolean expression saying whether or not this value is taken by at least one variable. And then it sums up these boolean expressions.

On your example:

x0 = 1; x1 = 1; x2 = 2; x3 = 3;   Values = 
{1, 2, 3
} x = [x0,x1,x2,x3]

With first formulation:

v     count(x,v)    (0<count(x,v)); 1     2              1 2     1              1 3     1              1 ------------------------------------ y=3

With second formulation:

v     (x0==v)  (x1==v)  (x2==v)  (x3==v)   (or(i in 0..3) xi==v) 1      1        1        0        0         1 2      0        0        1        0         1 3      0        0        0        1         1 ---------------------------------------------------------------- y=3

#CPOptimizer
#DecisionOptimization