Decision procedures using model building techniques

Abstract

Few year ago we have developed an Automated Deduction approach to model building. The method, called RAMC looks simultaneously for inconsistencies and models for a given formula. The capabilities of RAMC have been extended both for model building and for unsatisfiability detection by including in it the use of semantic strategies. In the present work we go further in this direction and define more general and powerful semantic rules. These rules are an extension of Slagle's semantic resolution. The robustness of our approach is evidenced by proving that the method is also a decision procedure for a wide range of classes decidable by semantic resolution and in particular by hyper-resolution. Moreover, the method builds models for satisfiable formulae in these classes, in particular, for satisfiable formulae that do not have any finite model.