Spreading excellence and disseminating the cutting edge results of our research and development efforts is crucial to our institute. Check for our educational offers for Bachelor, Master and PhD studies at the University of Innsbruck!

ELLY - An ELP Reasoner

Type: 
Master
Supervisor: 
Student name: 
Daniel Winkler
Completion Date: 
January 29, 2010
Download: 
AttachmentSize
PDF icon masterthesis-danielwinkler.pdf1.56 MB

Description Logics (DL) have been investigated for over three decades as a decidable formalism for modeling knowledge. Since even very inexpressive DL languages showed high worst-case complexity results, research has concentrated on extending existing formalisms and thereby retaining decidability. However, recent work has investigated subsets of existing formalisms that boast polynomial reasoning characteristics and ELP has been presented as a relatively expressive representative for the category of tractable DLs. ELP captures not only many existing tractable DLs but also serves as a formalism for a tractable subset of the Logic Programming paradigm.

Contemporary knowledge representation formalisms for the Semantic Web need to be expressive enough to be useful while at the same time having reasoning characteristics that scale to the size of the Web. This is the reason why major Web ontology languages like OWL and WSML provide variants that are restricted in their semantic expressivity for the sake of better reasoning behavior. The DL ELP captures the semantics of the OWL 2 profiles EL and RL as well as the WSML-DL v2.0 variant, thus a reasoner implementation of ELP is capable of supporting all of these languages. This thesis discusses the implementation of such an ELP reasoner and evaluates its scalability characteristics. Furthermore, a description is provided of the integration of Elly into frameworks that are used for reasoning with Semantic Web data.