A novel measurement-based neural fuzzy method is proposed for traffic modeling of an output buffer at a single N×N node in communication networks in this paper. The inputs of the system model are four of the firs...
详细信息
A novel measurement-based neural fuzzy method is proposed for traffic modeling of an output buffer at a single N×N node in communication networks in this paper. The inputs of the system model are four of the first-order and second-order statistics of the measured traf-fic parameters and the output is packet loss rate. The firstorder Sugeno fuzzy model is adopted. Simulation stud-ies show that the absolute RMSE is 0.0021 and the relative RMSE is 1.45%between the output of the proposed model system and the actual statistics respectively. This methodis suitable for real-time processing because of its simplicity and quickness in calculation.
暂无评论