Aditya K. Ghose, University of Wollongong and Srinivas Padmanabhuni, University of Alberta, Edmonton
Recent research on the integration of the abductive and constraint logic programming paradigms has led to systems which are both expressive and computationally efficient. This paper investigates therole of constraints in integrity checking in the context of such systems. Providing support for constraints in this role leads to a framework that is significantly more expressive, without sacrificing efficiency. We augment the Abductive Constraint Logic Programming framework with constraint assumptions and provide model- and proof-theoretic accounts of two variants: one which involves commitment to such assumptions, and one which does not. We also show that such accounts extend easily to a constraint logic programming framework which supports both negation and constraint assumptions. The gains in expressivity in these frameworks turn out to be particularly useful in a variety of application domains, including scheduling and constraint database updates.