Originally posted by: DavidGravot

Hi

Running the attached cpo file, there is this weird conflict:

alwaysIn(Plus_27, 844, 2284, 1, 1);
alwaysIn(Plus_27, 2284, 2344, 2, 2);

This corresponds to two interval variables that do overlap in the period [2284,234) so their cumul indeed jumps from 1 to 2 at time 2284

How come CPO considers this as a conflict ?

Thanks 

David

Originally posted by: PhilippeLaborie

Hum,

You have the following 4 interval variables:

IntervalVar_1 = intervalVar(optional, start=844..19564, end=844..19564, size=1440);
IntervalVar_5 = intervalVar(optional, start=2284..19564, end=2284..19564, size=840);
IntervalVar_9 = intervalVar(optional, start=3375..19564, end=3375..19564, size=1440);
IntervalVar_13 = intervalVar(optional, start=5164..19564, end=5164..19564, size=1380);

Because of alwaysIn(Plus_27, 844, 2284, 1, 1), the first interval which is the only one that can increase the function level on [844, 2284) (all the others start after 2284) will have to be fixed start=844, end=2284=844+1440.

Then constraint alwaysIn(Plus_27, 2284, 2344, 2, 2); requires two additional interval variable increasing the level over [2284,2344). But only interval IntervalVar_5 can be used for that, as the two other ones: IntervalVar_9 and IntervalVar_13 start later than 2284.

So the problem with these two constraints is indeed infeasible.

Philippe

Originally posted by: DavidGravot

Thanks Philippe

I indeed found a bug in my code... I forgot to add some intervals to my pulse function 

regards

David