Aus der Einleitung: Description Logics (DLs) are a family of knowledge representation formalisms that are successfully applied in many application domains. They provide the logical foundation for the Direct Semantics of the standard web ontology language OWL2. The light-weight DL EL, underlying the OWL2 EL profile, is of particular interest since all common reasoning problems are polynomial in this logic, and it is used in many prominent biomedical ontologies like SNOMEDCT and the Gene Ontology.

We introduce an extension of the lightweight Description Logic EL that allows us to de_ne concepts in an approximate way. For this purpose, we use a graded membership function, which for each individual and concept yields a number in the interval [0, 1] expressing the degree to which the individual belongs to the concept. Threshold concepts C~t for ~ then collect all the individuals that belong to C with degree ~ t. We generalize a well-known characterization of membership in EL concepts to construct a specific graded membership function deg, and investigate the complexity of reasoning in the Description Logic τEL(deg), which extends EL by threshold concepts defined using deg. We also compare the instance problem for threshold concepts of the form C>t in τEL(deg) with the relaxed instance queries of Ecke et al.

Fuzzy description logics (FDLs) are knowledge representation formalisms capable of dealing with imprecise knowledge by allowing intermediate membership degrees in the interpretation of concepts and roles. One option for dealing with these intermediate degrees is to use the so-called Gödel semantics. Despite its apparent simplicity, developing reasoning techniques for expressive FDLs under this semantics is a hard task. We present a tableau algorithm for deciding consistency of a SROIQ ontology under Gödel semantics. This is the first algorithm that can handle the full expressivity of SROIQ as well as the full Gödel semantics.

In a previous paper, we have introduced an extension of the lightweight Description Logic EL that allows us to define concepts in an approximate way. For this purpose, we have defined a graded membership function deg, which for each individual and concept yields a number in the interval [0; 1] expressing the degree to which the individual belongs to the concept. Threshold concepts C~t for ~ 2 ∈ {<, ≤, >, ≥} then collect all the individuals that belong to C with degree ~ t. We have then investigated the complexity of reasoning in the Description Logic τEL(deg), which is obtained from EL by adding such threshold concepts. In the present paper, we extend these results, which were obtained for reasoning without TBoxes, to the case of reasoning w.r.t. acyclic TBoxes. Surprisingly, this is not as easy as might have been expected. On the one hand, one must be quite careful to define acyclic TBoxes such that they still just introduce abbreviations for complex concepts, and thus can be unfolded. On the other hand, it turns out that, in contrast to the case of EL, adding acyclic TBoxes to τEL(deg) increases the complexity of reasoning by at least on level of the polynomial hierarchy.

In a recent research paper, we have proposed an extension of the lightweight Description Logic (DL) EL in which concepts can be defined in an approximate way. To this purpose, the notion of a graded membership function m, which instead of a Boolean membership value 0 or 1 yields a membership degree from the interval [0; 1], was introduced. Threshold concepts can then, for example, require that an individual belongs to a concept C with degree at least 0:8. Reasoning in the threshold DL T EL(m) obtained this way of course depends on the employed graded membership function m. The paper defines a specific such function, called deg, and determines the exact complexity of reasoning in T EL(deg). In addition, it shows how concept similarity measures (CSMs) ~ satisfying certain properties can be used to define graded membership functions m~, but it does not investigate the complexity of reasoning in the induced threshold DLs T EL(m~). In the present paper, we start filling this gap. In particular, we show that computability of ~ implies decidability of T EL(m~), and we introduce a class of CSMs for which reasoning in the induced threshold DLs has the same complexity as in T EL(deg).

In the last few years there has been a large effort for analysing the computational properties of reasoning in fuzzy Description Logics. This has led to a number of papers studying the complexity of these logics, depending on their chosen semantics. Surprisingly, despite being arguably the simplest form of fuzzy semantics, not much is known about the complexity of reasoning in fuzzy DLs w.r.t. witnessed models over the Gödel t-norm. We show that in the logic G-IALC, reasoning cannot be restricted to finitely valued models in general. Despite this negative result, we also show that all the standard reasoning problems can be solved in this logic in exponential time, matching the complexity of reasoning in classical ALC.

The notion of concept similarity is central to several ontology tasks and can be employed to realize relaxed versions of classical reasoning services. In this paper we investigate the reasoning service of answering instance queries in a relaxed fashion, where the query concept is relaxed by means of a concept similarity measure (CSM). To this end we investigate CSMs that assess the similarity of EL-concepts defined w.r.t. a general EL-TBox. We derive such a family of CSMs from a family of similarity measures for finite interpretations and show in both cases that the resulting measures enjoy a collection of formal properties. These properties allow us to devise an algorithm for computing relaxed instances w.r.t. general EL-TBoxes, where users can specify the „appropriate“ notion of similarity by instanciating our CSM appropriately.

Rough Description Logics have recently been studied as a means for representing and reasoning with imprecise knowledge. Real-world applications need to exploit reasoning over such knowledge in an efficient way. We describe how the combined approach to query answering can be extended to the rough setting. In particular, we extend both the canonical model and the rewriting procedure such that rough queries over rough EL ontologies can be answered by considering this information alone.

We investigate the computational complexity of axiom pinpointing in Description Logics, which is the task of finding minimal subsets of a knowledge base that have a given consequence. We consider the problems of enumerating such subsets with and without order, and show hardness results that already hold for the propositional Horn fragment, or for the Description Logic EL. We show complexity results for several other related decision and enumeration problems for these fragments that extend to more expressive logics. In particular we show that hardness of these problems depends not only on expressivity of the fragment but also on the shape of the axioms used.

The least common subsumer (lcs) w.r.t general EL-TBoxes does not need to exists in general due to cyclic axioms. In this report we present an algorithm for computing role-depth bounded EL-lcs based on the completion algorithm for EL. We extend this computation algorithm to a recently introduced probabilistic variant of EL: Prob-EL⁰¹.