Attributed and n-ary relations in OWL for knowledge modeling

Abstract

Knowledge modeling is the basis for the whole process of conceptual design. We aim, in this paper, at representing knowledge through an efficient model that enables efficient reuse. The knowledge model must be both expressive, so as not to omit any pertinent knowledge, and computable to allow for reasoning and learning. The ontology model, along with its well-known description language OWL (Web Ontology Language), is commonly used for this purpose. However, despite the strong expressiveness of OWL, this ontology language still shows limitations with regard to some modeling needs. We aim at promoting the representation of such aspects. We especially deal with the lack of an adequate representation for attributed relations and n-ary relations. The attributed relation, which is a relation that possesses its own attributes, is introduced into OWL, while preserving the basic semantics. The n-ary relation is also added to OWL by introducing a supplementary representational layer that is defined with a sound semantics, inspired from the DLR logic.