Originally posted by: SarangJagdale

We are working on a complex production planning & scheduling problem with sequence dependent setups in discrete manufacturing industry. Its a capacity constrained situation with demand and min/max inventory requirements.

In order to deal with capacity constrained situation we had to introduce demand shortage and min inventory requirement breach variables. Logical bounds have been introduced in the formulation to limit these variables.

CP solution search process starts with boundary conditions on these variables. For example initial solution is shortage equals full demand. Subsequently CP keeps cutting down on shortage but unable to reach a solution in which shortage equals 0 after longer runs, even in cases where sufficient capacity is available.

Kindly suggest ways to efficiently treat deviation variables like shortage and inventory breach.

One approach we are trying is to provide starting values of such deviation variables, for example shortage equals 0. Are there any other alternatives within CP?

Thanks

Originally posted by: Philippe_Refalo

More information about the model formulation would ne needed to help you proprly. However, how do you integrate the demand shortage and min inventory requirement breach variables in the objective function ? This can make a difference.

Something that you could look at are search phases. This permit to tell the search to start with some variables and not others.

Regards,

Philippe