Research On Semantic Conflict Detection And Diagnosis For Ontology Evolution In Open Environments

Ontology evolution is the timely adaptation of an ontology to the arisen changes and the consistent propagation of these changes to dependent artifacts. A survey of the state-of-art approaches to ontology evolution shows that most methods and tools were assumed to work in a centralized environment. However, some problems such as global knowledge, semantic conflicts and intractability arise from this assumption in an open dynamic environment like the Semantic Web. Evolving ontologies in an open environment neccesarily demands for non-standard reasoning technology of consistency checking, inconsistency localization and repair.Semantic conflicts detection and diagnosis are the major tasks for ontology evolution in open environment. On the basis of surveying the research on ontology evolution, this thesis focuses on the non-standard reasoning services for ontology consistency checking, semantic conflicts detection and diagnosis approaches to coping with the conflicting change sequences when multiple users independently make changes to an ontology in an open environment. The main contents are as followings:(1) We propose non-standard reasoning approaches for evolving ontology. In this method, the ontology document is analyzed by Jena, the results of which will be used as the inputs of the transformation algorithms, which are used to transferred the OWL ontology into Alloy model. Then Alloy Analyzer is used to judge the consistency. This method can remedy the deficiency of current reasoning methods of lacking ABox reasoning. Further, the counter-examples of the inconsistent ontologies are given to provide evidences to pinpoint the causes.(2) To improve the scale of the reasoning ontology, we present an improved SAT-based non-standard reasoning method for checking ontology consistency. In doing so, we define a logic based intermediate core-language, and present the transformation rules from OWL ontology language to the intermediate core-language. An experimental comparison shows that the proposed method can handle middle-sized ontologies. For the unsatisfiable ontology, we find out the cause by computing the minimal unsatisfiable sets.(3) We present a new method to detect conflicts for ontology evolution in open environment. The definitions of concepts and relationships of ontology evolution are the foundation of the research. We classifies conflicts as three groups:internal inconsistencies conflicts in change sequence, direct conflicts between the sequences and Inconsistent conflict between the sequences. For different conflicts, high effective detecting algorithms are presented with evaluation. This method can not only can find out the conflicts, but it also can get the maximum consistent changing sequence in a conflict situation to keep the evolution running.(4) We present a the semantic conflicts diagnosis approach, which focus on of the changing sequences of ontology evolution in open environment. This approach introduces conflict matrix to model the changing sequences, and uses statistical methods to get the minimal hitting sequence set of the matrix as the candidates for the diagnosis. By quantifying the Bayes reasoning mechanism with probability information, the diagnosis results can be measured quantitatively. It updates probability of each candidate, and the possibility rank result indicates the likelihood of the candidates to lead to the conflict.(5) To support the ontology evolution in an open environment, we designed and realized a prototype system. The system can detect and diagnosis the conflicts of the ontology evolution. Further, it provides consistency reasoning mechanism to prove the correctness of changing ontologies.