Intensional Set Constraints

Existing approaches to deal with sets in a CLP framework generally assume the niteness of the subjacent sets: that is a reason why intensional sets are often forbidden. In this paper, we propose a new compilation method to treat intensional sets in a general setting. Our representation relies on a simply typed-calculus: sets are considered as Boolean-valued functions that may involve union and intersection. This set expression language can be manipulated through suitable rules added to a-calculus with explicit substitutions. Using the fact that we get a rst order mechanism derived from a calculus with explicit substitutions, eecient implementations are intended taking advantage of existing speciic abstract machines for such calculus. Our method is a way to overcome the standard niteness restrictions concerning sets in CLP languages.