Research On Dynamic Description Logics Based On DL-Lite

With the rapid development of internet and its popularization in everyone’s life, description logic has been paid more and more attention to and researched by both at home and abroad. Description logic is a kind of knowledge formalism method based on logic and it is suitable for depicting the static domain knowledge, which has a strong ability of description, decidability features, efficient reasoning algorithms and reasoning tools as support. However, description logic can not depict problems with action characteristics.Researchers have proposed all kinds of dynamic expansion forms of diverse description logics and thus various dynamic description logics are built. But at present, in all dynamic description logics’ decidability algorithms, TBox of description logic’s part is limited not to include general concept inclusion axioms(GCIs). The reason is that as soon as GCIs are brought in, the algorithm becomes undecidable. DL-Lite is a light-weight description logic, which is cut for capturing concept modeling constructs and easy to be handled. Its reasoning complexity is low. What’s more, it is an easy-handled child part of OWL2 recommended as in recent years the Web ontology language. These characteristics make DL-Lite especially suitable as a kind of conceptual layer for data-intensive applications under large data.We start from the above two motives. Firstly, we choose DL-Lite to be extended dynamically to build a dynamic description logic whose algorithm supports GCIs. Then, it provides an effective reasoning tool for the above algorithm. Finally, from the perspective of application, we extend the logical tool and study its typical application examples. In this paper, based on the above ideas we carry out research work and specific work is mainly as follows:(1) Due to the current situation that present dynamic description logic decidability algorithms don’t support general concept inclusion axioms, first of all, description logic DL-Lite is combined with dynamic logic and action theory, and a new dynamic description logic pr RDDL- Lite where disjointness that involves roles in TBox is forbidden is built. Then, we give its syntax and semantics as well as the logic’s tableau algorithm design and introduce the whole process of the algorithm design. Finally its termination, reliability and completeness are analyzed and proved.(2) In order to be better applied to practical problems, there is a need in a quick and convenient way to decide the satisfiability problem of pr RDDL- Lite formuals. Therefore, on the basis of the above algorithm, the corresponding pr RDDL- Lite(Tableau) reasoning machine software is developed, and the software functions, design as well as tool demonstration are introduced to let users understand and use more easily.(3) Knowledge base in Semantic Web is always in evolution. In constrast to data base, knowledge base based on description logic DL-Lite is also faced with the same problem of transaction rollback and transaction concurrency. Thus based on pr RDDL- Lite, the converse action and concurrent action are brought in and a new dynamic description logic supporting the converse action and concurrent action is built, which is studied to be applied to the dynamic evolution of DL-Lite knowledge base.