The Communication Of Multiple Agents Based On BDI Model

Multi-Agent System(MAS)is a system that consists with a number of autonomous Agent which can communicate with each other. The computational element in MAS is Agent, which has two main properties or abilities. First, each Agent in system is some extend autonomous, that means it can decide to choose the actions by itself to achieve the goal. Second, each Agent can communicate with other Agents such as trust and cooperation, which is similar to human social behavior.The purpose of this thesis is to introduce and discuss the theory problems of Agent and MAS. Since there is few BDI model which contains the trust and cooperation among Agents,this thesis try to define several important concepts, which are defined through the original definition in BDI models, making the concepts intuitive, and to discuss the properties of these concepts.The first chapter is about the AI and MAS research situation. This part generally discuss AI's developing situation nowadays and some background information about MAS.The second chapter focuses on the individual Agent model, which contains the introduction of Agent and its properties, the comparison between deductive reasoning Agent and practical reasoning Agent. Then we lead to the BDI model.The third chapter is about Agent group based on BDI model. We use the original concepts for defining some new concepts such as trust, and then put out some theorems and proof to make the properties clear.