Argumentative dialogue revision in multi-agent system

Most of the traditional belief revisions were built on a completely reliable set of static beliefs between the believable intelligent agents, which are unrealistic due to their inability to identify the credibility of faith. Argumentative dialogue is a basic method for intelligent agents in multi-agent system to eliminate differences between them and reach a consensus. In this paper we will establish an argumentative dialogue framework for intelligent agents´ dialogue process by using the argumentative theory and transforming dialogue beliefs of agents into arguments, and building an original argumentative dialogue tree based on the relationship of arguments. Then we calculate the credible belief set of each agent after turning the original argumentative dialogue tree into standard argumentative dialogue tree. At last we calculate the acceptable belief set to make each agent´s mind reach an agreement. This method is characterized by high efficiency, good adaptability, easy realization, and removing forged or incredible beliefs to a certain extent.