Abstract: In order to mitigate the network congestion, a novel active queue management algorithm TDPQW is proposed by twice dropping packets method. In this algorithm, the mathematic formula of actual queue length and waiting time are deducted with M/G/1 queuing model, and the dropping strategy in queue front and in queue random location is presented. A simulation is conducted to research on the algorithm performance between TDPQW and RED, as well as DROP-TAIL. The result shows that it has better adaptability for TDPQW.

Key words: active queue management, twice dropping packets, M/G/1, queue length, waiting time

摘要： 针对网络拥塞现象，基于两次丢包方法建立了一种新的主动队列管理算法TDPQW。该算法利用M/G/1排队模型推导了实际队列长度和等待时间的数学表达式，以此提出在队列头部和队中随机位置进行丢包的策略。同时，通过仿真实验对比分析了该算法与RED、DROP-TAIL算法的性能，结果表明TDPQW具有较好的适应性。

关键词: 主动队列管理, 两次丢包, M/G/1, 队列长度, 等待时间