Network Congestion Algorithm based on Smith Predictor Fuzzy AQM

Jianjun WU

Hunan City University, Yiyang Hunan 413000, CHINA

Abstract: Since there are many problems of the traditional network congestion algorithm like the lack of ro- bustness and low steady-state, this paper proposes network congestion algorithm based on Smith predictor fuzzy AQM. This algorithm first analyzes the TCP congestion control mathematical model and the Smith predictive control theory, and then by using the fuzzy control in control theory and Smith Predictor compen- sation designs a new AQM algorithm to eliminate the negative effects of jitter and delay on the network rou- ter and queue-to-end delay jitter, thus this thesis achieves the purpose of achieving the desired stability and transient performance in the delay network. Finally, simulation experiment of network congestion algorithm based on Smith predictor fuzzy AQM are conducted. The results show that: the proposed algorithm improves the performance of the system, and it has good robustness and steady-state.