An active queue management algorithm for network congestion based on intelligent prediction control

Queue management is an important means to improve the performance of a network and avoid network congestion.At present,passive queue management dominates the queue management field.Recently active queue management(AQM) has been getting more and more attention,along with the random early detection(RED) algorithm,the first proposed AQM algorithm.In order to analyze a system made up of clients and a gateway,a strict mathematical model was developed to describe the system.A kind of RED controller with fast generalized prediction control was included to analyze congestion control.The system's architecture and identification are discussed in detail.A series of simulations were made,showing that fuzzy generalized predictive control(FGPC) is feasible in queue management and effective in controlling queue length,avoiding network congestion,and reducing round trip delay.