GAC Via Unit Propagation

In this paper we argue that an attractive and potentially very general way of achieving generalized arc consistency (GAC) on a constraint is by using unit propagation (UP) over a clausal theory encoding the constraint. This approach to GAC offers a number of advantages over the traditional approach of designing constraint specific algorithms (propagators). Most importantly, it is easier to implement (highly efficient UP engines are publicly available), it automatically provides an incremental algorithm for achieving GAC, it provides almost free decrementality, and it can easily provide clausal reasons to support learning and non-chronological backtracking. We give a new way of achieving GAC via UP for generic constraints that also sheds some additional light on the inferential power of GAC. Furthermore, we give structure specific ways of achieving GAC via UP for the regular, among, and gen-sequence constraints. For these constraints UP is able to achieve the same run time complexity bounds as previously presented propagators. Finally, we explain how a UP engine can be added to a CSP solver to achieve a fairly seamless integration of constraints propagated via UP and those propagated via more traditional constraint specific