Scheduling, limited capacity and constraint logic programming

Solving scheduling problems with constraint logic programming has been a well studied subject in research projects over the last few years. The emphasis has been put on how to model and solve scheduling problems in CLP. The main contribution of this paper is the generalization of the notion of limited capacity in the context of scheduling. In former approaches, resources were assumed to offer a capacity of one unit, and no choice among different resources was possible, or only one resource with a capacity greater than one was considered. The author proposes a more abstract view of resources called abstract resource groups, ARGs, allowing one to model and solve resource group planning problems. An algorithm to decompose scheduling problems on ARGs into independent subproblems is presented. The author describes a translation of the constraints used to model scheduling problems into CLP with emphasis on the capacity conditions. Finally, the author briefly reports his experiences with the application of a system based on this approach to a large sized industrial scheduling problem