Information Retrieval System Based On Mobile Agent And Optimal Search Theory

With the rapid development of network technology, Internet plays an important role in people's living, and the tremendous volume of information and heterogeneity of the network give information retrieval a great challenge. Traditional technology and methods can solve the problem of information retrieval to some extent, but it's still unable to meet the growth and heterogeneity of the network.Mobile agent technology combines Distributed Computing Technology and Artificial Intelligence. Mobile agent is a kind of compute body which has several features such as initiative, interoperability, reactivity and autonomy. It can migrate to search information from node to node and send back the result after finishing the task. The emergence of mobile agent technology has provided a revolutionary method for information retrieval on Internet. The thesis proposes an information retrieval system model based on mobile agent whose participators are: clients, a search server, a dock server, a communication server and a lot of servers which provide information resource.To find an efficient strategy for mobile agent's retrieval task, we refer to the Optimal Search Theory which is evolved from the Statistical Decision Theory in Operations Research and that is a theory of finding how those resources can be allocated to maximize the probability of detection, or to minimize the cost. We can use the Optimal Search Theory to optimize our system to gain higher detection probability. The thesis also studies non-adptive optimal search strategy and adaptive optimal search strategy.At last, we use IBM's aglet as the mobile agent platform, and realize the information retrieval system with Java language that we needn't concerned the platform. The results of the experiment indicate that the system has attained the desired objective and the optimal search strategy improves the search efficiency of the system. And father more, we compare non-adptive optimal search strategy and adaptive optimal search strategy, ananysis the results and pave a way to the future study and research.