Model Representation over Finite and Infinite Signatures
Computationally adequate representation of models is a topic arising in various forms in logic and AI. Two fundamental decision problems in this area are: (1) to check whether a given clause is true in a represented model, and (2) to decide whether two representations of the same type represent the same model. ARMs, contexts and DIGs are three important examples of model representation formalisms. The complexity of the mentioned decision problems has been studied for ARMs only for finite signatures, and for contexts and DIGs only for infinite signatures, so far. We settle the remaining cases. Moreover we show that, similarly to the case for infinite signatures, contexts and DIGs allow one to represent the same classes of models also over finite signatures; however DIGs may be exponentially more succinct than all equivalent contexts.
KeywordsFunction Symbol Predicate Symbol Ground Atom Ground Instance Universal Variable
Unable to display preview. Download preview PDF.
- 3.Baumgartner, P., Fuchs, A., Tinelli, C.: Lemma Learning in the Model Evolution Calculus (submitted)Google Scholar
- 8.Eiter, T., Fink, M., Tompits, H., Traxler, P., Woltran, S.: Replacements in non-ground answer set programming. In: Proc. of WLP 2006, pp. 145–153 (2006)Google Scholar
- 9.Fermüller, C.G., Leitsch, A.: Model building by resolution. In: Martini, S., Börger, E., Kleine Büning, H., Jäger, G., Richter, M.M. (eds.) CSL 1992. LNCS, vol. 702, pp. 134–148. Springer, Heidelberg (1993)Google Scholar
- 14.Kunen, K.: Answer sets and negation as failure. In: Proceedings of ICLP 1987, pp. 219–228. MIT Press, Cambridge (1987)Google Scholar