Designning & Researching A Distributed Data Mining System Based On Multi-Agent System

Computer network system has become a large-scale nondeterministic and dynamic distributed system day and day.But various kinds of existing network database management solutions can not adapt to decision-making and manage it efficiently and network database management is facing the new challenges.To apply multi-agent system to network and distributed data management is a brand-new thought of constructing the new generation network and distributed data mining system.This paper introduces the theories about data mining,distributed computing and agent technology.On the base of these theories , the basic theories of multi-agent system (MAS) and the communication manner between various Agents of MAS are researched emphasically in this paper.Furthermore,a theory model about distributed data mining system based on MAS is brought forward for the system distributed geographically and logicaly,in which the problem how to mining a large amount of distributed data can be soloved.The model adopts to a new technology named multi-agent in the field of artificial intelligence.In this paper,the complemention of the model and the communication between various agents as well mining arithmetic are discussed to a certain degree.At the same time, the feasibility and capability of the theory model are analysed in detail,including the puzzle and the key difficult technology that is anticipated to be resolved in the future.This model ,which incorporates the merits of moving agents and intellectual agents,can both provide the more intelligent surpport to implement the system's autonomous management for distributed data mining,and avoid the consumption of a large number of network broadband in order to solve the bottleneck problem of data transmission.