TSINGHUA SCIENCE AND TECHNOLOGY
ISSNll1007-0214ll02/14llpp623-632
Volume
15, Number 6, December 2010
Finding Data Tractable Description Logics for Computing a Minimum
Cost Diagnosis Based on ABox Decomposition
*
DU Jianfeng (杜剑峰)
1,2,**
, QI Guilin (漆桂林)
3
, Jeff Z. Pan
4
1. Guangdong University of Foreign Studies, Guangzhou 510006, China;
2. State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China;
3. School of Computer Science and Engineering, Southeast University, Nanjing 211189, China;
4. Department of Computing Science, the University of Aberdeen, Aberdeen AB24 3UE, UK
Abstract: Ontology diagnosis, a well-known approach for handling inconsistencies in a description logic (DL)
based ontology, computes a diagnosis of the ontology, i.e., a minimal subset of axioms in the ontology
whose removal restores consistency. However, ontology diagnosis is computationally hard, especially com-
puting a minimum cost diagnosis (MCD) which is a diagnosis such that the sum of the removal costs at-
tached to its axioms is minimized. This paper addresses this problem by finding data tractable DLs for com-
puting an MCD which allow computing an MCD in time polynomial in the size of the ABox of a given ontol-
ogy. ABox decomposition is used to find a sufficient and necessary condition to identify data tractable DLs
for computing an MCD under the unique name assumption (UNA) among all fragments of SHIN that are
at least as expressive as
core
DL - Lite without inverse roles. The most expressive, data tractable DL identi-
fied is SHIN without inverse roles or qualified existential restrictions.
Key words: ontology diagnosis; minimum cost diagnosis; description logics; data tractability
Introduction
Ontologies play a core role in the semantic web (SW) as
they provide formal representations of shared knowl-
edge. Ontologies need to be logically constructed to
enable users to understand the knowledge in a consis-
tent way. To this end, the W3C organization proposed
the web ontology language (OWL)
[1]
for modeling on-
tologies based on description logics (DLs)
[2]
. Most
DLs are decidable fragments of first-order logic (FOL)
and they inherit from FOL the property ex falso sequi-
tur quodlibet (from falsehood/contradiction follows
anything), thus inconsistency in a DL-based ontology
is fatal. Unfortunately, inconsistencies frequently occur
during the construction, evolution, and integration of
DL-based ontologies. Thus the handling of inconsis-
tencies is crucial in real-life applications.
Ontology diagnosis is a well-known approach for
handling inconsistencies in a DL-based ontology
[3]
,
originating from the field of model-based diagnosis
(MBD). In a seminal paper on MBD, Reiter
[4]
intro-
duced diagnoses as minimal subsets of components in
a logical system whose removal corrects the system.
When applied to DL-based ontologies, a diagnosis of
an inconsistent ontology is a minimal subset of axioms
whose removal restores consistency. This notion of
diagnosis has been adapted by Schlobach
[3]
to handle
incoherent TBoxes and by Kalyanpur et al.
[5]
to handle
unsatisfiable atomic concepts. A DL-based ontology is
made up of a TBox and an ABox. A TBox is incoherent
iff it has some unsatisfiable atomic concepts.
Received: 2010-09-16; revised: 2010-10-13
*
Supported by the National Natural Science Foundation of Chin
(Nos. 61005043 and 60970045)
**
To whom correspondence should be addressed.
E-mail: jfdu@mail.gdufs.edu.cn; Tel: 86-20-39328876