Data complexity of reasoning in very expressive description logics

  • Authors:
  • Ullrich Hustadt;Boris Motik;Ulrike Sattler

  • Affiliations:
  • Dept. of Computer Science, Univ. of Liverpool, UK;Forschungszentrum Informatik, Univ. of Karlsruhe, Germany;Dept. of Computer Science, Univ. of Manchester, UK

  • Venue:
  • IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Data complexity of reasoning in description logics (DLs) estimates the performance of reasoning algorithms measured in the size of the ABox only. We show that, even for the very expressive DL SHIQ, satisfiability checking is data complete for NP. For applications with large ABoxes, this can be a more accurate estimate than the usually considered combined complexity, which is EXPTIME-complete. Furthermore, we identify an expressive fragment, Horn-SHIQ, which is data complete for P, thus being very appealing for practical usage.