Application of Multi-Agent Theory in Social Networks

This paper discusses the dynamic multi-agent system is scattered by the configuration of many autonomous or semi-autonomous subsystem (agent) through the information network which is formed by the interconnection system. Individuals with relatively equal status in the system and each other for information exchange, coordinated movement. Group behavior is a common phenomenon in the nature, the whole group can in macro reflect a more complicated and more efficient than individual behavior. In engineering applications, some tasks cannot be achieved by a single system and needs to be done many subsystems together, such as uav reconnaissance, distributed collaborative filtering and observation, robot soccer game, sensor network data fusion, etc. Study of dynamic multi-agent system is the expansion of traditional control theory, on the traditional isolation system are already mature some basic concepts of system analysis and comprehensive, such as stability, balance, controllability, servo controller, etc., has a new meaning and new problems. In addition, in many industries and has a broad application prospect, such as economic system, the oil refining system, power system, transport system, process system and military system, etc. As a result, nearly five years related research has become the international leading one of the most active direction of the control theory.