论文部分内容阅读
在AOS高级在轨系统链路协议的基础上,分析了具有自相似特性的业务流量,提出了一种基于FARIMA模型的自相似预测的链路优化模型。该预测模型基于分数阶统计理论,在估计赫斯特参数的基础上,根据不同的时间粒度,提前预测突发业务量的到来,从而降低了网络丢包率。仿真表明模型在20点预测内具有较好拟合性,在一定置信区间下具有较好的预测成功概率和较低的虚警概率,同时使网络丢包率大幅下降。
Based on AOS advanced on-orbit system link protocol, this paper analyzes the traffic with self-similar features and proposes a link optimization model based on FARIMA model for self-similar prediction. Based on the fractional statistics theory, this prediction model predicts the arrival of burst traffic based on different Hurst parameters and reduces the packet loss rate. The simulation shows that the model has a good fit within the 20-point prediction, has a better probability of predicting success and a lower false alarm probability under a certain confidence interval, and greatly reduces the network packet loss rate.