Originally posted by: jiaodaxiaozi

Hi

I wrote the following indicator constraints in OPL :

indicatorconstraint = forall (l in L, t in T)     t - 1 - delta_a[l] == 0 => ind[l][t] == 1;

Here the results is:

The value of delta_a[l] = 1 and delta[2] = 2, so I expecting the value of ind is:

ind =  [[0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
         [0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]

so how cplex deals with the indicator constraints, I don't understand.

Hi,

do not you mix imply and equivalent ?

range L=1..2;
range T=1..4;

dvar boolean ind[L][T];
dvar int delta_a[L];

maximize sum(l in L,t in T) ind[l][t];
subject to
{
indicatorconstraint :forall (l in L, t in T)     (t - 1 - delta_a[l] == 0) => (ind[l][t] == 1);
delta_a[1] == 1;
delta_a[2] == 2;
}

execute
{
writeln(ind);
}

gives

[[1 1 1 1]
         [1 1 1 1]]

but if we change

indicatorconstraint :forall (l in L, t in T)     (t - 1 - delta_a[l] == 0) => (ind[l][t] == 1);

into

indicatorconstraint :forall (l in L, t in T)     (t - 1 - delta_a[l] == 0) == (ind[l][t] == 1);

then we get

[[0 1 0 0]
         [0 0 1 0]]

regards

Originally posted by: jiaodaxiaozi

Hi Alex,

I tried " indicatorconstraint :forall (l in L, t in T)     (t - 1 - delta_a[l] == 0) == (ind[l][t] == 1); " , and it works, thanks for your kind help!