Elsevier

Journal of Web Semantics

Volumes 27–28, August–October 2014, Pages 78-85
Journal of Web Semantics

Konclude: System description

https://doi.org/10.1016/j.websem.2014.06.003Get rights and content

Abstract

This paper introduces Konclude, a high-performance reasoner for the Description Logic SROIQV. The supported ontology language is a superset of the logic underlying OWL 2 extended by nominal schemas, which allows for expressing arbitrary DL-safe rules. Konclude’s reasoning core is primarily based on the well-known tableau calculus for expressive Description Logics. In addition, Konclude also incorporates adaptations of more specialised procedures, such as consequence-based reasoning, in order to support the tableau algorithm. Konclude is designed for performance and uses well-known optimisations such as absorption or caching, but also implements several new optimisation techniques. The system can furthermore take advantage of multiple CPU’s at several levels of its processing architecture. This paper describes Konclude’s interface options, reasoner architecture, processing workflow, and key optimisations. Furthermore, we provide results of a comparison with other widely used OWL 2 reasoning systems, which show that Konclude performs eminently well on ontologies from any language fragment of OWL 2.

Introduction

The current version of the Web Ontology Language (OWL 2)  [1] is based on the very expressive Description Logic (DL) SROIQ (see  [2] for a DL introduction) and extends the first version of OWL with more expressive language features such as qualified cardinality restrictions and property chains.

Many existing reasoning systems have been adapted to OWL 2 and several new optimisations have been developed to deal with the latest language features or specific profiles. Despite all the progress, reasoning performance still shows up as a noticeable issue for users. Because of the N2EXPTIME-complete worst-case complexity for standard reasoning tasks in SROIQ   [3], this is expected at least for some ontologies. However, there are several clues which indicate that further improvements are possible. For instance, an effective coupling of fully-fledged OWL 2 reasoning procedures with tractable procedures for OWL 2 profiles could improve the overall performance. Moreover, multi-core computers are ubiquitous now, but state-of-the-art reasoners for expressive DLs do not yet implement an effective parallelised processing architecture.

In this system description, we introduce the novel reasoning system Konclude,1 which addresses both aforementioned issues. It incorporates different reasoning procedures and implements new as well as extensions of existing optimisations adapted to concurrent processing within a multi-core, shared memory architecture. This significantly improves the running time of reasoning tasks for many real-world ontologies.

As of now, Konclude handles the DL SROIQ and also supports nominal schemas  [4] which generalise arbitrary DL-safe rules  [5]. Konclude supports the most common reasoning services such as classification, realisation, queries for sub-classes, class instances or types of individuals and it can be used as server or via command line on various platforms.

The rest of this system description is organised as follows: we next introduce the system’s architecture; in Section  3 we give an overview of the integrated optimisations; in Section  4 we present the result of a comprehensive evaluation and comparison to other state-of-the-art reasoners before we conclude in Section  5.

Section snippets

System architecture

Konclude is implemented in C++ and makes use of the cross-platform application framework Qt.2 The reasoner runs on all Qt supported platforms including Windows, OS X, Linux, and Solaris.

Konclude offers two communication options: First, it is an OWLlink server that exposes ontology management and reasoner functionality to one or more clients (usually ontology-based applications) via the W3C OWLlink protocol  [6]. As OWLlink server, Konclude supports OWL 2 XML as content

Optimisations

A naive tableau algorithm is not suitable for handling typical real-world ontologies since completion graphs can easily become very large and may contain many sources of non-determinism. To tackle these challenges, Konclude applies a significant range of state-of-the-art and new optimisations. The main conceptual and technical challenge of the system was to extend important optimisations to work with more expressive languages as well as to effectively implement them within an inherently

Evaluation

In this section we provide an evaluation that compares the current version (0.5.0) of Konclude and the state-of-the-art reasoners FaCT++ 1.6.2  [18], HermiT 1.3.8  [19], Pellet 2.3.1  [20], and ELK 0.4.1  [21] (for EL ontologies). The evaluation uses a large test corpus of ontologies that have been obtained by collecting all downloadable and parseable ontologies from the Gardiner ontology suite  [22], the NCBO BioPortal,5 the National Cancer Institute thesaurus

Conclusions and future work

In this system description we have introduced Konclude, a new OWL 2 DL reasoner that supports the DL SROIQV. We have described Konclude’s parallel processing architecture and the integrated key optimisation techniques that result in remarkable overall reasoning performance. In order to support the latter, we have presented and discussed results of a comprehensive comparison between Konclude and other state-of-the-art reasoners for different reasoning tasks and many ontologies. The comparison

References (26)

  • B. Glimm et al.

    A novel approach to ontology classification

    J. Web Sem.

    (2012)
  • E. Sirin et al.

    Pellet: a practical OWL-DL reasoner

    J. Web Sem.

    (2007)
  • W3C OWL Working Group, OWL 2 web ontology language: document overview, W3C Recommendation, 27 October...
  • F. Baader et al.

    The Description Logic Handbook: Theory, Implementation, and Applications

    (2007)
  • Y. Kazakov

    RIQ and SROIQ are harder than SHOIQ

  • M. Krötzsch et al.

    A better uncle for OWL: nominal schemas for integrating rules and ontologies

  • I. Horrocks, P.F. Patel-Schneider, H. Boley, S. Tabet, B.N. Grosof, M. Dean, SWRL: a semantic web rule language, W3C...
  • T. Liebig et al.

    OWLlink

    Semant. Web - Interoperability, Usability, Applicability

    (2011)
  • I. Horrocks et al.

    The even more irresistible SROIQ

  • F. Baader et al.

    Pushing the EL envelope

  • Y. Kazakov

    Consequence-driven reasoning for horn SHIQ ontologies

  • A. Steigmiller et al.

    Nominal schema absorption

  • A.K. Hudek et al.

    Binary absorption in tableaux-based reasoning for description logics

  • Cited by (134)

    • A multiplatform reasoning engine for the Semantic Web of Everything

      2022, Journal of Web Semantics
      Citation Excerpt :

      Concerning automated reasoning at a Web scale, algorithm optimization for distributed execution in Cloud environments has been first proposed by WebPIE [37]. Tiny-ME aims to support cloud environments through an extension of the OWLlink protocol [9] for client–server interaction: the standard specification is supported by Konclude [10] and many more Semantic Web reasoners. Table 1 summarizes relevant features of related reasoning systems, highlighting Tiny-ME’s contributions.

    • Towards an Ontology for Robot Introspection and Metacognition

      2023, Frontiers in Artificial Intelligence and Applications
    View all citing articles on Scopus
    View full text