Background

The semantics of disjunctive logic programs is based on minimal models which make atoms appearing in a disjunctive program false by default. In many cases, this is highly desirable, but certain problems become awkward to formalize if all atoms are blindly subject to minimization. Parallel circumscription enables the use of varying and fixed atoms in addition to those minimized which eases the task of knowledge presentation in many cases. The idea is that the atoms in P=Hb(Π)\(V ∪ F) are subject to minimization, while the truth values of the atoms in V may vary freely and the truth values of the atoms in F are kept fixed. Atoms that are minimized remain implicit in this definition.

An extended notation Circ(Π,P_1>...>P_k,V,F) is introduced to represent the prioritized circumscription of disjunctive program Π which includes the parallel circumscription of Π as its special case, that is, when k=1. The idea is that atoms in P_1 are falsified with the highest priority, those in P_2 with the next highest priority, and so on. We say that models of Circ(Π,P_1>...>P_k,V,F) are the ⟨P_1>...>P_k,V,F⟩-minimal models of positive program Π. Lifschitz [Lifschitz, 1985] shows that a prioritized circumscription Circ(Π,P_1>...>P_k,V,F) corresponds to a conjunction

.

The translator circ2dlp translates the problem of finding models for Circ(Π,P_1>...>P_k,V,F) into the problem of finding stable models for the linear-size translation circ2dlp(Π,P_1>...>P_k,V,F) [OJ, 2008] and enables thus the use of efficient ASP-solvers such as GnT and dlv.

This software has been designed to be used with GnT and the front-end lparse. In addition, there is a support for producing an input file for dlv.