中文摘要：

主动队列管理（ActiveQueueManagement，简称AQM）是网络拥塞控制领域的一个热点．针对主动队列管理中PID算法超调量大，抖动大的问题，提出一种GPID（GearshiftPID）的拥塞控制算法．该算法特点是：积分作用大小跟随瞬时队列长度的偏差大小变动而变动，队列长度的偏差越大，则积分系数值越小，收敛速度越慢，超调量越大；反之，偏差越小，积分系数值越大，收敛速度越快，超调量越大．该算法可以有效的控制队列长度的偏差变化，调节瞬时队列长度趋于队列期望值收敛．仿真实验结果表明，该算法收敛速度快，链路利用率高，平均队列长度更趋于期望值．

英文摘要：

Active queue management （ AQM ） is one of the hot research field in network congestion. In this paper, a congestion con- trol algorithm Gearshift integral PID （ GPID ） is proposed, which tends to address the problem of overshoot and big jitter of PID in AQM. The algorithm is characteristic of dynamic changes of the integral role on PID in accordance with the deviation of stimutaneous queue length, that is, the larger the deviation in queue length is, the smaller the integral coefficients are, the faster the convergence rate is and the bigger the overshoot is and vice versa. This algorithm can effectively control the deviation of the queue length so that it can attain the expected length. The simulation result shows that GPID has a faster convergence speed, a higher rate of link utiliza- tion and that the average queue length is closer to the expected length.