The application is a schedule generator for non-preemptive real-time distributed systems.
It ends up as a fairly large collection of global serialization constraints and arithmetic constraints (both equality and inequality). There is also a little bit of disjunction in some cases, but the case I was working with here didn't have any disjunctive constraints. Before building the model there was a procedure to try and infer a good level of discretization for the problem, but today I replaced that with a resolution parameter specified by the user. Obviously the discretization was too aggressive :-( Thanks again for all of your help, -Joe
_______________________________________________ Gecode users mailing list [EMAIL PROTECTED] https://www.gecode.org/mailman/listinfo/gecode-users
