DL-Lite with Attributes and Sub-Roles (Extended Abstract)

The DL-Lite family of description logics has recently been proposed and investigated in [5–7] and later extended in [1, 8, 3]. The relevance of the DL-Lite family is witnessed by the fact that it forms the basis of OWL 2 QL, one of the three profiles of OWL 2 (http://www.w3.org/TR/owl2-profiles/). According to the official W3C profiles document, the purpose of OWL 2 QL is to be the language of choice for applications that use very large amounts of data. This paper extends the DL-Lite languages of [3] by relaxing the restriction on the interaction between cardinality constraints (N ) and role inclusions (or hierarchies, H). We also introduce a new family of languages, DL-LiteHNA α , α ∈ {core, krom, horn, bool}, extending DL-Lite with attributes (A). The notion of attributes, borrowed from conceptual modeling formalisms, introduces a distinction between (abstract) objects and application domain values, and consequently, between concepts (sets of objects) and datatypes (sets of values), and between roles (relating objects to objects) and attributes (relating objects to values). The advantage of the presented languages over DL-LiteA [8] is that the range restrictions for attributes can be local (and not only global as in DL-LiteA). Indeed, DL-Lite HNA α has a possibility to express concept inclusion axioms of the form C v ∀U.T , for an attribute U and its datatype T . In this way, we allow re-use of the very same attribute associated to different concepts with different range restrictions. For example, we can say that employees’ salary is of type Integer, researchers’ salary is in the range 35,000–70,000 (enumeration type) and professors’ salary in the range 55,000–100,000—while both researchers and professors are employees. Note that local attributes are strictly more expressive than global ones. For example, concept disjointness (or unsatisfiability) can be inferred just from datatype disjointness for the same (existentially qualified) attribute. Since DL-Lite languages have been proved useful in capturing conceptual data models [8, 4, 2], the extension with attributes, as presented here, will generalize their use even further. We aim at establishing computational complexity of knowledge base satisfiability in these new DLs. In particular we prove the following results: