Based Agent Network Desing And Implementation

With the rapid development of the network techniques and the spreading of network teaching system, the "answering" in network teaching system has been an urgent problem. At present, the construction of "answering" in our country is still at the elementary step, and the answer system is even less. Whereas the answer system existed now has the problem as the following: monotonous method, low intelligence and independence, and the knowledge base without the ability of self-study.Agent technology is now the field to be studied in rapid development, and is widely used in business, industry, military and education. The characteristics of Agent technology such as being intelligent, automatic and moveable make it possible to solve the network problems existed in our country now. The purpose of this thesis is to make the advantage of Agent designing thought and Multi-Agent technology to solve the problems.According to this purpose, the thesis firstly illustrates the correlative conception and nomenclature of Agent theory and machine study theory; then has a detailed study and analysis on the key technologies concerned; and finally, puts forward the system model based on Automatic Answer Agent and Knowledge Study Agent .The Automatic Answer Agent model accomplishes picking up the participles, key characters which describing natural language, and calculating the corresponding key weight; realizes the passing of data among Agents by communication Agent; achieves the index of knowledge base by using the Intelligence Matching Agent; and brings about an integrated evaluation an the system automatic answer result by using Answer Explaining Agent. In Knowledge Base Study Agent model, it adapts machine study technology and demonstrating study model with the support of the Expert Auditing Agent, thus realizes the extension of knowledge base.This thesis takes the course "Visual Foxpro6.0 Program&Design" as the experiment case, adopts .net platform, and combines with Agent thought, accomplishes the Web Answer System Based on Agent (WASBA). The result indicates the design has basically reach the purpose, realizes automatic answer and self-study of the system knowledge base. Compared with other answer system, it has improved obviously in initiative and intelligence.