Suitable Graphs for Answer Set Programming

Often graphs are used to investigate properties of logic programs. In general, different graphs represent different kinds of information of the corresponding programs. Sometimes this information is not sufficient for solving a certain problem. In this paper we define graphs which are suitable for computing answer sets of different classes of logic programs. Intuitively, a graph associated to a logic program is suitable for answer set semantics if its structure is sufficient to compute the answer sets of the corresponding program. That is, algorithms that use suitable graphs do not have to consider the original logic program any longer. We investigate different classes of graphs which are suitable for answer set computation of normal nested logic programs, normal logic programs and normal programs with at most on positive body atom.