Knowledge Reasoning Method Based On Inconsistent Ontology Non-revision

With the widely application of ontology, the validity of ontology is more and more critical. But the existence of an inconsistent ontology is inevitable, such as reuse of other author’s ontology, polysemy of words, integration of heterogeneous ontology and no comprehensive inspection in the process of ontology evolution, would lead to an inconsistent ontology. Knowledge reasoning method based on ontology is divided into two categories, the description logic-based knowledge reasoning method and the rule-based knowledge reasoning method. In pratical reasoning the two would be forced to stop in case of meeting an inconsistent ontology. In order to solve this problem, it is necessary to deal with the inconsistent ontology.To solve the problem that current ontology inference engine can’t reason based on an inconsistent ontology, this thesis applies the belief non-revision method to knowledge reasoning based on ontology, designs a knowledge reasoning method based on an inconsistent ontology by using the belief non-revision method, puts forward the framework of the reasoning method.Firstly, the problem currently in the ontology inference engine is discussed. Through understanding of the related theory of ontology and method of reasoning, pinpoints the reasoning task of ontology inference engine and the essence that ontology inference engine can not reason based on an inconsistent ontology.Secondly, the belief non-revision method is introduced. Through understanding of the relevant theory about belief non-revision, pinpoints that this method can undertake reasoning based an inconsistent belief set, and is able to get a reliable conclusion. A knowledge reasoning algorithm based on inconsistent ontology non-revision is designed by using this reasoning method to make ontology inference engine reasoning based on an inconsistent ontology.Thirdly, a refined algorithm about knowledge reasoning based on description logic and based on rule is designed. Aiming at the existing problem of knowledge reasoning based on ontology, this thesis refers to the key to the implementation of belief non-revision method to improve the original reasoning algorithm, making the reasoning operation not affected by the quality of ontology, and serving the knowledge without opposite information as the final conclusion of the system.Finally, a knowledge reasoning prototype system based on inconsistent ontology non-revision is designed and implemented. Adopting some concrete instances, the rationality and the availability of the reasoning prototype system are tested. This system being able to reason based on an inconsistent ontology availably is verified at last.