A Generic Querying Algorithm for Greedy Sets of Existential Rules

Michaël Thomazo 1 Jean-François Baget 1 Marie-Laure Mugnier 1 Sebastian Rudolph 2
1 GRAPHIK - Graphs for Inferences on Knowledge
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier, CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : Answering queries in information systems that allow for ex- pressive inferencing is currently a field of intense research. This problem is often referred to as ontology-based data ac- cess (OBDA). We focus on conjunctive query entailment un- der logical rules known as tuple-generating dependencies, existential rules or Datalog+/-. One of the most expressive decidable classes of existential rules known today is that of greedy bounded treewidth sets (gbts). We propose an algo- rithm for this class, which is worst-case optimal for data and combined complexities, with or without bound on the pred- icate arity. A beneficial feature of this algorithm is that it allows for separation between offline and online processing steps: the knowledge base can be compiled independently from queries, which are evaluated against the compiled form. Moreover, very simple adaptations of the algorithm lead to worst-case-optimal complexities for specific subclasses of gbts which have lower complexities, such as guarded rules.
Document type :
Reports
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download

https://hal-lirmm.ccsd.cnrs.fr/lirmm-00675560
Contributor : Thomazo Michaël <>
Submitted on : Thursday, March 1, 2012 - 2:23:13 PM
Last modification on : Wednesday, December 12, 2018 - 2:38:02 PM
Long-term archiving on : Wednesday, December 14, 2016 - 9:47:47 AM

File

main-kr-long.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : lirmm-00675560, version 1

Collections

Citation

Michaël Thomazo, Jean-François Baget, Marie-Laure Mugnier, Sebastian Rudolph. A Generic Querying Algorithm for Greedy Sets of Existential Rules. RR-12006, 2012. ⟨lirmm-00675560⟩

Share

Metrics

Record views

534

Files downloads

352