Solving Practical Reasoning Problems with Extended Disjunctive Logic Programming

Herre, Heinrich
Wagner, Gerd
We present a definition of stable generated models for extended generalized logic programs (EGLP) which a) subsumes the definition of the answer set semantics for extended normal logic programs [GL91]; and b) does not refer to negation-as-failure by allowing for arbitrary quantifier free formulas in the body and in the head of as rule (i.e. does not depend on the presence of any specific connective, nor any specific syntax of rules). We show how to solve classical ATP problems in the framework of extended disjunctive logic programming (EDLP) where neither Contraposition nor the Law of the Excluded Middle are admitted principles of inference. Besides being able to solve classical ATP problems in a monotonic reasoning mode, EDLP can as well treat commonsense reasoning problems by employing its intrinsic nonmonotonic inference capabilities based on stable generated models. EDLP thus proves itself as a general-purpose AI reasoning system.
Appeared / Erschienen in: 
Report Nr. 8/1996 des Institutes für Informatik der Universität Leipzig
Pubdate / Erscheinungsdatum: 
Pages / Seitenanzahl: 
1996-7.pdf195.57 KB