Abstract

. Database and Artificial Intelligence applications are briefly discussed and it is argued that they need deduction methods that are not only refutation complete but also complete for finite satisfiability. A novel deduction method is introduced for such applications. Instead of relying on Skolemization, as most refutation methods do, the proposed method processes existential quantifiers in a special manner which makes it complete not only for refutation, but also for finite satisfiability. A main contribution of this paper is the proof of these results. Keywords: Artificial Intelligence, Expert Systems, Databases, Automated Reasoning, Finite Satisfiability. 1 Introduction For many applications of automated reasoning, the tableaux methods [32, 16, 34, 18] have the following advantages: They not only detect unsatisfiability but also generate models; they are close to common sense reasoning, hence easy to enhance with an explanation tool; and they are quite easy to adapt to the special...

Citations