Research And Implementation Of Spliting Ontology Methods Based On Semantic Grid

With the fast development of the information technology, the demand that users obtain the information resources is being strengthened constantly. The traditional information search tool based on key word, can't realize more complicated semantic search and the inquiry on the basis of deducing. The Ontology as one kind modeling tool of the information system on semantic and knowledge level can describe the conceptual model. With the good concept hierarchical structure and support to reasoning from logic, the Ontology is suitable for realization of semantic search especially. But huge Ontology exercise was restricted in the performance of searching .When the RAM and compute ability of system can't be infinite expanded, the unique and viable method is to cuts up into some sub-Ontology and apply it in distribute environment.Aimed to the application of semantic web in the distribute system, the thesis researchs the technology of splitting Ontology in the surrounding of semantic grid.Although the current technique of grid doesn't support semantic reasons, the technique of grid supports various relations database. For this reason, the thesis designs a method of storing Ontology relations through relation database. The system scans the OWL Ontology by means of Semantic tools -Jena, and the same times, the affiliations between super class and sub class and the triple relations between different concepts were save in the database respectively. As a result, the integration of semantic web and grid was realized, and the complicated process of semantic reasoning was converted the familiarly searching operation in the relation database.The thesis still realizes one methods of splitting Ontology. On the basis of first code and splitting threshold, this method divides Ontology to different sets .Because every set have the mapping relation to itself and others, so we can unite the two sets to one by means of computing the contrasting relation and comparing the threshold, as a result, the splitting granularity was confirm. This method can make the lowest coupling degree with each other, minimize data redundancies that were produce in the process of splitting Ontology, and prove the feasibility of ontology applying in the distribute surrounding.