基于元胞自动机的PI主动队列管理方法研究

针对传统的PI(Proportional Integral)算法收敛速度慢等问题,基于瞬时到达速率提出了一种新的PI主动队列管理算法PICA(Proportional Integral Based on Cellular Automa-ton).首先,该算法结合瞬时队长和瞬时到达速率建立了丢包策略,并利用元胞自动机刻画了数据包的动态特性.同时,通过仿真实验,将该算法与传统的PI算法以及RPI(Rate basedProportional and Integral)算法进行比较,结果发现PICA算法在有效传输数据包、时延和丢包率等方面的性能都较优.

Study on PI active queue management based on cellular automaton

In order to mitigate the slow of convergence rate in traditional Proportional Integral algorithm, a novel active queue management algorithm(Proportional Integral Based on Cellular Automaton, PICA) is proposed based on instantaneous arrival rate. At first, combined with instantaneous queue length and instantaneous arrival rate, the dropping strategy is presented in this algorithm, and the dynamic characteristic of packet is depicted by cellular automaton. Then, a simulation was conducted to study on the algorithm performance between PICA and PI, as well as RPI(Rate based Proportional and Integral). The result shows that it is better performance in efficient transmission packets, delay and dropping rate for PICA algorithm.