Strong polynomiality of resource constraint propagation

Constraint-based schedulers have been widely successful to tackle complex, disjunctive and cumulative, scheduling applications by combining tree search and constraint propagation. The constraint-propagation step is a fixpoint algorithm that applies pruning operators to tighten the release and due dates of activities using precedence or resource constraints. A variety of pruning operators for resource constraints have been proposed; they are based on edge finding or energetic reasoning and handle a single resource. Complexity results in this area are only available for a single application of these pruning operators, which is problematic for at least two reasons. On the one hand, the operators are not idempotent so a single application is rarely sufficient. On the other hand, the operators are not used in isolation but interact with each other. Existing results thus provide a very partial picture of the complexity of propagating resource constraints in constraint-based scheduling. Th...

• Type: Journal
• Year: 2007
• Where: DISOPT
• Authors: Luc Mercier, Pascal Van Hentenryck
• Keywords: DISOPT 2007, Operators, Pruning, Resource Constraints