An Business Intelligence System Based On Multi-Agent

Currently the Business Intelligent System includes four parts, which are Data Collection, Data Warehouse, Data Mining and OLAP. For the Business Intelligent System's data source is simple, response agent, expanded intelligent meta and data mining technology are introduced into the Business Intelligent system which based on Multi-agent System, so that to construct an efficient and intelligent platform.For overcome the insufficiency of present Business Intelligent, the intelligent agent engine is applied, and regards the multidimensional query as a kind of data source, which cause the OLAP multidimensional query not to be the passive query any more. Agent client is established on the framework of response agent, which proposes the tendency analysis on its own initiative. By the client's real-time query, historical query records and other changed information, Multi-agent environment is designed to solve the problem of communications among agent clients which improves the concurrency. The main component of agent server is the Intelligent meta container, which collects Intelligent meta and analyses associate-rule of Intelligent-meta through data mining algorithm which based on the Intelligent-meta tree. Some result-set will be stored in the local-view, which will be post to user immediately without computing of Data Warehouse. The purpose of local-view is to enhance the efficiency of system response and reduce the computing burden.The key to all system's data can carry on effectively is the realization of intelligent meta. On the basis of CWM, the count of subjects and dimensions are extended by the data"meta-intelligent meta Switch", which packs the query information using XML language. Packed Intelligent meta exist in the Multi-agent environment and realize request and exchange in this environment.