Originally posted by: PreranaJ

Hi,

We are trying to solve a scheduling problem using CP. However, there are more soft constraints in the problem than hard, making the feasible search space very large (no further reduction of domain is possible  using constraint propagation).  Because of a large feasible search space CP takes very long time to solve and is not making smart decisions quickly. Is there any way we can make the search smart in reduced feasible search space?

Originally posted by: PhilippeLaborie

Hello,

Unlike many classical CP systems, CPO does not necessarily assume that the constraint propagation is strong in order to find good quality solutions to your problem, and the large size of the search space does not mean that it will have difficulties to find good solutions. Here also, the first thing to do is to try better formulations of the problem. We would be interested in understanding your current model and advise better ways to model it (if any of course).

Philippe

Originally posted by: PreranaJ

Hi Philipe,

We tried formulating the problem in number of different ways but the one I posted in the other post gives the best performance. Any help on how I can improve CP's performance will be very helpful. 

Regards

Prerana