A Parallel Deduction for Description Logics with ALC Language

The term Description Logics (DLs) is commonly accepted to indicate a certain class of formal logics for representing knowledge and reasoning about it in information systems. These logics are descendants of a formal calculus that was proposed by Brachman in the KL-ONE system [5]. To represent the important entities of a given domain in DLs one can use atomic concepts, roles and individuals (instances of the concepts). Additionally, a set of constructors for denoting complex concepts and roles is defined to obtain the adequate expressivity of this formalism. Description Logics can be classified by the languages they support; one of the basic languages in this area is called ALC (the acronym stands for Attribute Concept Description Language with Complements).

A possible real life application of this approach is the Semantic Web domain. To realize this new Web vision the community of knowledge engineers searches for efficient methods of knowledge representation and reasoning. Open standards are built for the representation of Web-based knowledge in a machine readable manner (RDF, OWL) and different tools for supporting reasoning in the Semantic Web are proposed (with SWRL submitted by W3C). A significant part of the tools is based on DLs, thus finding efficient inference methods in this area is a particularly important task.

The organization of this paper is as follows. The basic formalism of DLs is presented in section 2. In section 3, the essential constraint programming terms are depicted. A parallelization of an inference process for DL, performed in the constraint programming system Mozart, is defined in section 4. Section 5 contains some final remarks. An earlier version of this article was published as [6].