Research Of Combined Reasoning Mechanism In Ontology Reasoning

With the rapid development of Internet, the data on the network is also exploding. There are many conveniences brought to people. At the same time it becomes more difficult to obtain effective knowledge. As a kind of effective management and application method for knowledge, ontology is widely applied in various fields. However, there are a lot of implicit knowledge included in the ontology. And reasoning machine can be very convenient for acquiring the knowledge hidden in ontology. At present, the reasoner is usually used to implement the reasoning of the whole ontology by means of description logic or rule based reasoning algorithm. The efficiency of the Abox reasoned by description logic is low, since the whole instance level are needed to considered. And the knowledge hidden in ontology is hardly acquired by the description logic reasoning. At the same time, the completeness of the Tbox reasoning part can not be guarantee by rule based reasoner.Considering the problems above, the inference algorithms and features of the inference engine are studied in this thesis. And the performance of analytical reasoning machine in ontology reasoning is discussed. A combined reasoning model is proposed in this dissertation to extend the function of the reasoning machine. And the efficiency of reasoning are improved through this reasoning model. In this thesis the main works are completed as follows:The research results of ontology and knowledge reasoning are discussed. The mechanism and realization technology of ontology knowledge reasoning are deeply studied. The characteristics of common and special purpose reasoning machine are compared and analyzed, and the combined reasoning model is built according to the characteristics of the inference engine. The Abox and Tbox are separated by OWL API, during the reasoning process. The Tbox part of ontology are reasoned by Pellet. And the Abox part are reasoned by Jena with the rules which are rewritten by ontology axioms. In the process of reasoning, the efficiency is improved by merging operation of ontology reasoning rules. The ontology construction methods are analyzed and compared. And an ontology about person are created through the Protege. After the creating of the ontology, a part of reasoning rules are defined.Finally, the performance of the combined reasoning machine is verified by experiments. And a query reasoning system based on ontology is constructed. The concept and instances can be easily inquired through the system. And the implicit knowledge can be obtained conveniently through the general rules.