Research Of AQM Strategy Based On Single Neuron Adaptive PID In Congestion Control Of Networks

Internet has experienced an explosively growth since 1990's. Widespread use of Internet, as well as the appearance of varied network applications has made network congestion a significant problem. So how to avoid and control congestion is one of the most active fields in the computer networks.In this paper, the innovation lies in proposing the method of applying the single neuron PID adaptive control strategy to active queue management, by reason of three following respects: 1. Active queue management is an important congestion control mechanism, and its quality concerns the quality of the performance of whole network. 2. The famous algorithm RED and mutation algorithms are designed and improved according to project experience, and lack the effective theoretical direction. 3. The new strategy that this paper designs, regarding control theory as guidelines, combines PID's advantage with the single neuron's. It not only gains the PID's advantages, the real-time character, intuition and simple realization etc., but also overcomes PID's disadvantages in adjusting parameters and difficult disposal facing non-linear system.First of all, this paper gives a general introduction to the background, cause of IP network congestion, and the development of congestion control mechanism. With that, congestion control strategies in terminal and routers are introduced separately. AQM as the key congestion control strategies in the routers is analyzed especially.Then, this paper expounds the cause, the principle and designs of the single neuron PID adaptive control strategy, and carries on research and proves from two respects. On one hand, proceeding from mathematics model that TCP flow controls, analyze RED algorithm, PI algorithm, increment type PID algorithm and the improved single neuron PID algorithm, according to control theory, and carry on emulation under the environment of Matlab software and compare them in the dynamic performance of the whole system; On the other hand, adopt network simulation software NS2 to prove and compare four kinds of algorithms through the way of the simulated test. Thus, the algorithm that this paper puts forward is superior to RED, PI and increment type PID algorithm on the stability, robustness andadaptivity.Finally, Summarize exploration of the congestion control mechanism of Internet network and look into the distance. Hope to instruct us how to analyze and solve the problem with the ripe control theory, and can play a certain role to the development of the congestion control strategy.