基于灰色预测可变裕度PID网络自适应算法

摘要

为了综合控制拥塞链路的队列长度,提高AQM系统对动态网络环境的自适应能力,提出了一种基于灰色预测和考虑可变裕度PID控制的自适应TCP网络主动队列管理.首先,建立相角和幅值裕度与网络参数相关的PID自适应主动队列(TCP/AQM)控制论模型,该模型可以根据网络参数的变化而动态改变控制参数,以提高AQM网络动态自适应能力,及系统的鲁棒性；其次,将灰色预测引入该模型,实现路由器队列长度的超前预测,补偿带有PID反馈模块的AQM算法给队列造成的时滞影响.与其他算法的仿真结果相比较,该设计算法能够使信息流在较短的时间内稳定在期望队列长度阈值附近.%In order to overcome the weakness of existing AQM algorithms based on the control theory and enhance AQM algorithm ' s self-adapting, this paper designed a new adaptive AQM algorithm of PID with variable gain and phase margins based on grey prediction. First, established the model of PID control for adaptive AQM. In this model, gain and phase margins of PID were changed with the variation of network parameters(round-trip time,link capacity,load factor) ,in order to enhance algorithm ' s self-adapting. Second, introduced grey prediction algorithm into model to achieve feedback data' s advanced prediction , which could compensate PID feedback control for AQM' s delay. Contrasted with traditional PID algorithm,adaptive AQM algorithm of PID with variable gain and phase margins, adaptive AQM algorithm of PID with variable gain and phase margins based on grey prediction can adjust itself to queue' s threshold rapidly and stably.