Multi-Agent Belief Merger and Interaction Based On BDI

In the field of intelligent information retrieval, it´s meaningful to speculate a user´s true intentions through the interaction and belief merger. How to improve the rate of the Precision and the Recall has become a focus. We first discussed the interactive process in multi-agent system in which there exists an agent with deficient knowledge. Using it, we can deduce a user´s intention accessing to his true intention. But the problem of belief conflict can not be slide over when the process is described. We used a measure of Majority operator to solve the problem. The method can improve the efficiency of IR partially resulted from incomplete and indefinite intentions of a user´s queries.