A Fuzzy Predictive Controller Base on Neural Network Model for Active Queue Management

摘要：

Active queue management (AQM) has been an important means of congestion control for Internet. Designing an effective AQM algorithm is very difficult because of dynamic delay of network. After analyzing the weaknesses of RED and PI controller from control theory theoretic standpoint, we present a fuzzy predictive controller based on neural network model, namely FPCNN, for AQM supporting TCP flows in this paper. FPCNN is composed of a neural network predictive model and a fuzzy logic controller. In comparison to the PI AQM, FPCNN exhibits robust response to the changes of network parameters like number of TCP sessions and round trip time.

关键词： AQM congestion control neural network fuzzy predictive control

作者: L. Wang X.Z. Zhou

作者单位: School of Engineering and Management,Nanjing University,Nanjing 210093,China;College of Automation,N School of Engineering and Management,Nanjing University,Nanjing 210093,China

会议类型: 国际会议

会议名称: 2008高等智能国际会议(2008 International Conference on Advanced Intelligence)

会议地点: 北京

会议语种:英文

在线出版日期: 2008-10-18（万方平台首次上网日期，不代表论文的发表时间）