Mechanizing Proofs of Integrity Constraints in the Situation Calculus

Abstract

We address the problem of proving automatically integrity constraints in databases represented within the situation calculus. Integrity constraints specify what counts as a legal database state; it is a property that every database state must satisfy. Reiter ([REI, 91] and [REI, 93]) has presented an approach to database integrity constraints in which the concept of a database satisfying its constraints is defined in terms of inductive entailment from the database, induction axioms on database states, and other axioms of the situation calculus. In this paper we develop and implement a framework for proving integrity constraints in databases. We show how our induction theorem prover that includes strategies for finding suitable induction schemes and inference rules for automatically generalizing conjectures can be used to prove many integrity constraints completely automatically from an update database specification alone. We provide computer applications to a toy database (education) to illustrate our framework.KeywordsInference RuleIntegrity ConstraintDatabase StateInduction SchemeInduction VariableThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.