| Since the Web founder Tim Berners-Lee proposed Semantic Web (Semantic Web) structure, ontology which is the intelligence basis of computer to understand the information has been a focus of research in recent years. However, the differences in ontology constructing of time, location, the purpose of structure and the knowleage of people, lead to prodigious diversities, even though ontology comes from the same area, resulting in the conflict between the ontology. In order to solve the above problem and achieve the existing ontology reuse, knowledge sharing and semantic interoperability, some researchers propose ontology integration technology. However, due to the distributed, dynamic and open characteristics of the semantic web make ontology integration very difficult. And the study of the current agent and multi-agent system (MAS) has been more mature. The dynamic, autonomous features of agent greatly satisfy the needs of ontology integration in semantic web, also lay the foundation of intelligent ontology integration.By comparison and analysis of existing ontology integration systems and based on the in-depth study of ontology, agent and multi-agent theory. This paper designs and realizes the MAS-based ontology integration system. Firstly, the architecture of this system is described, and the functions of agent in the system are defined detailedly. Secondly, as ontology mapping is the key to Ontology Integration, the process of MAS-based ontolgoy mapping and integration is discussed. Thirdly, the current widespread use of the Multi-Agent System Application Platform JADE(Java Agent Development Framework) is introduced. Finally, by using JAVA and the relevant class libraries in JADE, the communication and collaboration among agents and the prototype system are implemented intelligently and efficiently. |
