Inconsistent Reasoning For Ontology

When people get new information, they are apt to change cognitive state. This may cause inconsistency in Knowledge Base sometimes. Ontology is a way to express knowledge. So many manipulations could lead to conflicts in some degree, such as polysemy and ontology migration and so on.Ontology reasoned is the reason tool of ontology system, which is mainly based on two mechanisms. One is rule-based reasoning, and the other one is description logic reasoning. No matter which one it is, the reasoner would check Knowledge Base to ensure consistency before reasoning. When inconsistency happens, the reasoner will immediately stop reasoning operation. Even though the inconsistency can be detected, there is no solution in reasoner to solve this issue, just leaving it to the system engineer. So it is becoming more and more important to improve the performance of reasoners. We hope the improved reasoner could allow reasoning with inconsistent ontology.Now solutions to this problem have been proposed. There are two thoughts in this field. One is to revise the inconsistent ontology to get a good one for reasoning, such as the application of belief revision in contradictory ontology. The other one is to adopt non-standard reasoning based on the incoherent ontology, so the reason could continue with inconsistent premise, such as paraconsistent reasoning. As we know, ontology revision could lead to lose useful information sometime. According to disadvantage, a new proposal has been proposed to solve inconsistent ontology reasoning.In this thesis, the thought of belief non-revision is been used to manage the inconsistent ontology reasoning. The basic idea of non-revision is that we put new belief into the belief set regardless of conflict and take the reasoned conclusion as the final result, which there is no information against with. Then two methods has been come up. One is to translate ontology description language to the expression in first order logic, so the reason can be performed in FOL. The other one is to improve tableau algorithm in view of belief non-revision to achieve inconsistent reason. And we propose the basic frame of improved reasoners and the translation process.