Suitable graphs for answer set programming

Abstract

Often graphs are used to investigate properties of logic programs. In general, different graphs represent different kinds of information of the underlying 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 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. We identify different classes of graphs to be suitable for different classes of programs. We also give first experimental results showing the impact of the used graph type to one particular graph based algorithm for answer set computation. © Springer-Verlag Berlin Heidelberg 2005.