BIROn - Birkbeck Institutional Research Online

    Boolean Role Inclusions in DL-Lite with and without time

    Kontchakov, Roman and Ryzhikov, Vladislav and Wolter, F. and Zakharyaschev, Michael (2020) Boolean Role Inclusions in DL-Lite with and without time. In: Calvanese, D. and Erdem, E. and Thielscher, M. (eds.) Proceedings of the 17th International Conference on Principles of Knowledge Representation and Reasoning (KR 2020). California, U.S.: IJCAI Organization, pp. 582-591. ISBN 9780999241172.

    [img] Text
    KR.pdf - Published Version of Record
    Restricted to Repository staff only

    Download (449kB)

    Abstract

    Traditionally, description logic has focused on representing and reasoning about classes rather than relations (roles), which has been justified by the deterioration of the computational properties if expressive role inclusions are added. The situation is even worse in the temporalised setting, where monodicity is viewed as an almost necessary condition for decidability. We take a fresh look at the description logic DL-Lite with expressive role inclusions, both with and without a temporal dimension. While we confirm that full Boolean expressive power on roles leads to FO^2-like behaviour in the atemporal case and undecidability in the temporal case, we show that, rather surprisingly, the restriction to Krom and Horn role inclusions leads to much lower complexity in the atemporal case and to decidability (and ExpSpace-completeness) in the temporal case, even if one admits full Booleans on concepts. The latter result is one of very few instances breaking the monodicity barrier in temporal FO. This is also reflected on the data complexity level, where we obtain new rewritability results into FO with relational primitive recursion and FO with unary divisibility predicates.

    Metadata

    Item Type: Book Section
    Additional Information: KR 2020, Rhodes, Greece, September 12-18, 2020. Series ISSN: 2334-1033.
    School: School of Business, Economics & Informatics > Computer Science and Information Systems
    Depositing User: Roman Kontchakov
    Date Deposited: 20 Jan 2021 12:01
    Last Modified: 29 Jun 2021 01:22
    URI: https://eprints.bbk.ac.uk/id/eprint/41537

    Statistics

    Downloads
    Activity Overview
    0Downloads
    85Hits

    Additional statistics are available via IRStats2.

    Archive Staff Only (login required)

    Edit/View Item Edit/View Item