Towards RIF-OWL combination : an effective rasoning technique in integrating OWL and negation-free rules

Abstract

For the development of Semantic Web the the World Wide Web Consortium has focussed on the future Web Ontology Language standard OWL 2 and rule languages in Rule Interchange Format which are required to interoperate. A lot of researches have been carried out towards the integration of various Description Logics and rules and RIF-OWL combination is a step forward in this direction. Most of the Description Logics are a decidable subset of classical First-Order logic while rules originate from logic programming. In this thesis, we discuss the combination of Description Logics and negation-free rules, both of which are expressed in the standard First- Order logic semantics. We propose an algorithm for reasoning which is sound but not complete for these combinations. Our algorithm uses existing standard reasoning tool for retrieving facts from the Description Logic Knowledge Base with which rules are put together to form a logic program. Finally, rule reasoner is used for answering conjunctive queries on this logic program. We identify the reasons behind the incompleteness and chose a subset of these combinations which combines restricted Description Logic and rules. This restricted subset consists of Description Logic Horn-SHIQ and rules such that Description Logic predicates are allowed only in the rule bodies and all the rules are DL-safe. A prototype implementation of the reasoning process is also presented which uses Pellet Description Logic reasoner and XSB rule reasoner showing the e

For the development of Semantic Web the the World Wide Web Consortium has focussed on the future Web Ontology Language standard OWL 2 and rule languages in Rule Interchange Format which are required to interoperate. A lot of researches have been carried out towards the integration of various Description Logics and rules and RIF-OWL combination is a step forward in this direction. Most of the Description Logics are a decidable subset of classical First-Order logic while rules originate from logic programming. In this thesis, we discuss the combination of Description Logics and negation-free rules, both of which are expressed in the standard First- Order logic semantics. We propose an algorithm for reasoning which is sound but not complete for these combinations. Our algorithm uses existing standard reasoning tool for retrieving facts from the Description Logic Knowledge Base with which rules are put together to form a logic program. Finally, rule reasoner is used for answering conjunctive queries on this logic program. We identify the reasons behind the incompleteness and chose a subset of these combinations which combines restricted Description Logic and rules. This restricted subset consists of Description Logic Horn-SHIQ and rules such that Description Logic predicates are allowed only in the rule bodies and all the rules are DL-safe. A prototype implementation of the reasoning process is also presented which uses Pellet Description Logic reasoner and XSB rule reasoner showing the e