Self-adaptive Network Service Research Based on Multi-Agents in Hierarchy

To meet topology changes causing by the growing web application dynamically,and with Agent of the high degree of flexibility, collaboration, autonomy, and self-adaptive characteristics, this paper discusses a self-adaptive interaction strategy based on Multi-Agents in hierarchy and analyzes an abstract model in which Multi-Agents are divided into the main Agents and sub-Agents in different layers. Each main Agent is responsible for management and coordination of near a number of sub-Agents, which can balance the load, And collaborative interaction between main Agents or sub-Agents in same layer can keep service well. Multi-Agents stratified strategy can reduce data reported to high level, shorten the routing. and with its autonomy and coordination of each point, while avoiding obstacles due to single point of failure. The strategy can maintain the consistency of network topology between all Agents and ensure the effectiveness of routing in real-time. It can also improve network access routing shooting,and also improve the quality of service (QoS).