Research Of Multi-agent Active Network Congestion Control Strategy

In recent years,with the rapid growth of the network scale,there are more and more traffic. But traditional routing algorithms cannot make full resource utilization, which always lead to network congestion and it has become a bottleneck which restrictst he growth and applieation of the network. So how to realize network congestion control is a brand-new research subjeet.This thesis mainly focuses on the research of networke on gestion control method based on multi-Agent and Active network Algorithm. The PSO algorithm is improved, the path option method based on improved PSO is proposed and a network congestion control system based on intelligent multi-agent is designed and implemented. The main contents are as follows:The PSO is improved with multi-agents. Method adjusting velocity threshold of PSO based on multi-agents is put forward, which guarantees local and global searching ability. The algorithm can get the best value rapidly and mean while it can avoid converging at local best value. Also,it imProvesin computing speed and precision. Simulation results demonstrate the validity and feasibility of the proposed algorithm. A network congestion control method is put forward based on improved PSO path optimization to more QoS data stream. The method makes network load balancing and decrease resource consumption to avoid congestion in the condition of meeting bandwidth, delay and cost constrains.A network congestion control system based on intelligent multi-agent is designed and implemented and the system is applied to solve the problems in congestion control. The cause is analyzed and a congestion control method based on multi-Agent is proposed to overcome the shortcomings of the traditional method. The congestion can be controlled fast and locally through the cooperation of local Agent and neighbors, including configuration for TCP Protoeol to avoid the deterioration.