The Design Of Management Information System Based On Agent And The Study On Human Computer Interaction Scheduling

With the development of the science and technology and the improvement of the management, it is an important problem for an enterprise to consider how to obtain the huge economic benefit under the new management style - Enterprise Information Management. This dissertation designs a Human Computer Interaction System based on Agent, which is the project of Qinshang Factory Management Information System. The design of the whole system adopts the three-layer system that integrates the C/S and the B/S structures. The Enterprise Management Information System based on multi-agent is introduced. The communication between client and server and the communication among sub information units are all achieved by "communication agent". This dissertation also explicates the realization of such three-layer information system with Powerbuilder and other development tools. Meanwhile, according to the four steps in database application system design: system demand analysis, general structure design, logic structure design and physics structure design, the analysis and the design of database in the Enterprise Management Information System are elucidated, and the technique of Powerbuilder for the implementation of the human interface based on the Web is also explained in the dissertation. In order to strengthen the coordination among systems, the dissertation constructs the Agent in the Information System, the enterprise information integration platform based on Agent - an interaction system with three abstract layers: the interface layer, the coordination layer and the model layer, which is able to mine the useful data from the huge dataset to improve the query completeness, accelerate the query speed and facilitate the query process. The dissertation elucidates the three key techniques for realize this structure: information guidance, correlation feedback and initiative service. The key of information guidance is the information classification and the construction of user model. This dissertation utilizes the fuzzy neural network for the data mining, and then constructs the information classification knowledge and user model. The steps are: 1. construct the fuzzy neural network and training dataset; 2. prune the structure fuzzy neural network; 3. cluster the value of hidden neuron; 4. extract the fuzzy rule. The fuzzy neural network for data mining is also relatively fully explained in the dissertation.