摘要： 基于常规控制理论的主动队列管理(AQM)算法在复杂动态网络环境下对参数变化比较敏感，难以保证队列稳定性且缺乏鲁棒性。针对上述问题提出基于队列长度和链路速率相对变化率的模糊AQM算法，以队列长度与期望队列长度以及链路速率与链路容量的相对误差量作为网络拥塞指示，采用模糊推理得出中间节点的丢包概率。仿真实验表明，该算法具有良好的队列稳定性和较小的队列延时，对网络的非线性和负载波动等不确定因素具有鲁棒性。

关键词: 主动队列管理, 拥塞控制, 模糊推理, 鲁棒性

Abstract: Most Active Queue Management(AQM) algorithms based on control theory have low robustness and queue stability under the complex network environment with uncertainties since those algorithms are more sensitive to the variations of network parameters. Aiming at the above problem, this paper presents a novel fuzzy AQM algorithm based on the relative changes of queue length and link rate, which introduces two relative errors as congestion notifications and also as the inputs of fuzzy inference system, one is the error between current queue length and the desired values, the other is the error between the link rate and link capacity, an appropriate dropping probability at router is determined by a set of fuzzy rules. Simulation results show that the algorithm has better performance on queue stability and less delay, at the same time it has good robustness for nonlinearity and load variation.

Key words: Active Queue Management(AQM), congestion control, fuzzy inference, robustness

中图分类号: 

• TP312