TITLE: ``Minimal Model Reasoning and its Efficient Implementation''

ABSTRACT:

The talk deals with the minimal model semantics MIN for positive disjunctive logic programs. Under this semantics a disjunction of ground literals is considered true if and only if it holds in all minimal Herbrand models of the program.

The positive disjunctions in MIN form the so-called minimal model state. We show how they can be derived efficiently by using delta-iteration together with a tree representation, called clause tree, for sets of disjunctions. The negative disjunctions in MIN form the so-called extended generalized closed-world-assumption. They can be characterized in terms of two types of characteristic sets for disjunctions: the support-for-negation sets and the support-for-minimality sets. We present techniques for the efficient derivation of the collection of all such characteristic sets. Finally, we show that the mixed disjunctions in MIN can be derived by logical deduction from the positive and the negative ones.

The minimal model semantics is one of the semantics that is provided by the system DisLog which has been developed at the University of Tuebingen. DisLog is a toolkit for non-monotonic reasoning with disjunctive logic programs. It integrates many of the common semantics for logic programs (minimal model, stable model, possible model and well-founded semantics) and allows for the fast implementation of further semantics.

Everyone is welcome to attend.

Web Accessibility