Research On The Application Based On DFR Of Multi- Agent's Social Behaviors Mechanism

Agent is a very important concept widely used in AI/DAI, the research of whose behavior is arousing an increasing attention in different environment, especially in social environment. Denett put it forward that it can be explored into from the physical perspective, designing perspective and consciousness perspective. The present research is more inclined to be focused on the consciousness perspective, in an effort to predict its behavior by studying its faith, wishes and other consciousness with Agent as its rational subject. Structurally, Agent and society featuring an large number of agents are systems of nonlinear nature, the whose changes belong to the dynamics development process. Because of the limit of single agent's power and the limited resources possessed, it, when assigned a mission or have a job, will turn to other Agents in the society for help, asking a number of them to make up an organization to accomplish the task. The interaction among the Agents, taking the form of collaboration, consultation and negotiation, calls for the formulation of social behavior rules, so as to regulate and guide the social behavior of Agent and guarantee the accomplishment of the common goal, and should work out the evaluation standards of its social behavior, which is conducive to discriminating between its social and nonsocial behavior.As to the logic structure of Agent and the dynamic fuzzy feature of social behavior, this essay put forward a formalized descriptive method to construct relative social behavior model, and also analyzed and researched how Agent behavior associated with its internal state property. This essay elaborated the formation mechanization, exchange, behavior credibility of Agent social behavior, and at last supplied the construction of simulative human resource market to exemplify the success of the research of this essay.