Abstract: This paper studies the relationship between the changing rate of drop probability and the queue stability, and specifically researches computing function of dropping probability of Adaptive Random Early Detection(ARED) algorithm and Random Exponent Marking(REM) algorithm respectively. As a result, a Non-Linear Adaptive Congestion Control Algorithm(NLACCA) is proposed. Based on the Active Queue Management(AQM) scheme, which provides a non-linear adaptation to the dropping probability function of the ARED, NLACCA enables the dropping probability gradient to vary along with the deviation that is between the instantaneous queue length and the target queue length. NLACCA can not only reduce the jitter of the target queue length, but also improve the stability of the algorithm. Simulation results demonstrate that the NLACCA algorithm outperforms in most scenarios, such as the jitter of queue length, delay, and packets dropped.

Key words: congestion control, Random Early Detection(RED), Random Exponent Marking(REM), non-linear adaptive control

摘要： 研究丢弃概率的变化率与队列长度稳定性间的关系，分析ARED算法及REM算法的丢弃概率计算函数，采用非线性化函数计算丢弃概率，提出一种非线性自适应拥塞控制算法(NLACCA)，根据队列长度与目标队列长度中值的偏离程度动态地调整丢弃概率的变化率，从而减小队列长度波动，提高算法稳定性。在NS-2上进行的大量实验结果表明，该算法具有队列长度抖动性小、平均时延低、丢包数少等特点。

关键词: 拥塞控制, 随机早期检测, 随机指数标记, 非线性自适应控制

CLC Number: 

• TP301