A Congestion Control Algorithm of Fuzzy Control in Routers

Evaluating on network state and forecasting the occurrence of congestion, active queue management (AQM) can abandon selectively those groups, so as to make source know the network state in time and adjust the dispatch speed. But mechanism now has defect in respondent speed, stability and environment sensitivity. As to these problems mentioned by the above, fuzzy control can bring relative advantages into play, actualizing the intellectualizing of control system. Based on the existed AQM algorithm - REM, this paper explores one kind of refinement algorithm - fuzzy REM (FUZREM). The test shows that this algorithm can stable queue length rapidly and achieve perfect control properties in the event of network dynamic change. Comparing to other available AQM algorithms, this algorithm improves on the efficiency of network congestion control.