Practical Querying of Temporal Data via OWL 2 QL and SQL:2011

Abstract

We develop a practical approach to querying temporal data stored in temporal SQL:2011 databases through the semantic layer of OWL 2 QL ontologies. An interval-based temporal query language (TQL), which we propose for this task, is defined via naturally characterizable combinations of temporal logic with conjunctive queries. This foundation warrants well-defined semantics and formal properties of TQL querying. In particular, we show that under certain mild restrictions the data complexity of query answering remains in AC 0 , i.e., as in the usual, nontemporal case. On the practical side, TQL is tailored specifically to offer maximum expressivity while preserving the possibility of reusing standard first-order rewriting techniques and tools for OWL 2 QL.

Tractable Interval Temporal Propositional and Description Logics

“…(In a nutshell, OWL 2 QL is DL-Lite H horn in which concept and role inclusions cannot have on the left-hand side.) In the context of temporal databases, we are interested in suitable ontology and query languages with temporal constructs (although some authors advocate the use of standard OWL 2 QL with temporal queries (Klarman 2014;Borgwardt, Lippmann, and Thost 2013)).…”

Temporal Conjunctive Query Answering in the Extended DL-Lite Family

“…Research in the second direction (temporal query answering with classical ontologies) has been first considered in a general way in [44] for expressive operators on both the temporal and the DL side. A particular result shows first-order rewritability for DL-Lite core , but this is achieved by considering epistemic semantics for CQs [49,52]. Temporal conjunctive queries have first been studied for ALC [14] and later for expressive extensions such as SHOQ [15,16], focusing on the complexity of the entailment problem under open-world semantics, resulting in very high combined complexities.…”

On coarser interval temporal logics

“…The coarser algebras IA 7 and IA 3 inspire the logics HS 7 and HS 3 proposed in this paper. These languages follow similar ideas to the standard SQL:2011 [Kla14], where interval relations are not necessarily Allen's ones (for example, later is interpreted as the disjunction of Allen's meets and later); therefore, they can be applied not only to classical areas of artificial intelligence, but also to temporal databases. We prove that the satisfiability problem for HS 3 is PSpace-complete in the finite/discrete case, and it is PSpace-hard when interpreted in any interesting class of linearly ordered sets, but that coarser relations do not guarantee the decidability of an interval temporal logic, because HS 7 remains undecidable over every interesting class of linearly ordered sets.…”