Study On The Agent-Based Information Integration Framework

With the improvement of enterprise information level,enterprise can get more and more information. At the same time, the kinds and structures of information are more and more abundant (from traditional DBMS to XML document, and to so much semi-structured information in the Web). The traditional application is based on the information store layer. All of these make that application can't efficiently use and manage these kinds of heterogeneous data in enterprises. We bring forward an agent-based information integration framework model (ABII) that is adaptable and easily maintained. The ABII model is consisted of many autonomous agents. Agents cooperate by the model of centralization and collaboration, and wrap the heterogeneous data source by the resource agent. On the other hand, we express this integrated information in a standard data format by the use of XML. So this model can easily extend and integrate new data source.According to the character of agents, we classify them into three catalogs: User Agent, Registration Agent and Function Agent. By organizing the agent according the character, and maintaining a run-time object set in the Registration Agent to log the information of agents in the system, we can easily add or delete their agents in the system. Agents cooperate by the model of centralization and collaboration and contact the others by their own router module. Router module is in charge of sending out and receiving message. So the model is flexible and can support the dynamically changing of structure. In the data store layer, we wrap the heterogeneous data source by the use of resource agent. We express this integrated information in a standard data format by the use of XML and use Document Type Definition as the data model of the data structure definition of various data resource. So we can easily describe various unformulated data. In the model, we use XPath as a kind of public query language. In the resource agent, the public query based on XPath is translated to local query according data resource. So we can transparently query data and this operation is not in dependent on backward data resource.Aimed at the heterogeneousness problem of remote data source in the long-distance education software, we make ABII as an assistant supporting platform that is applied in the long-distance education.