提出一种基于ARIMA-GARCH和抛物线模型相结合的网络流量动态预测方法,既发挥ARIMA-GARCH模型处理和预测网络流量内部短期变化特征的能力,又融合了抛物线模型捕捉网络长期动态变化特征的性能,即构成AGP组合模型.仿真实例表明,AGP组合模型能够很好地预测网络流量的动态变化规律,在网络服务控制方面有广泛应用前景.
A dynamic prediction method of network traffic based on the combination of ARIMA-GARCH and parabolic models was proposed. This method not only gave full play to ARIMA-GARCH model's capability of dealing with and predicting network traffic of short internal changes, but also integrated the advantage of parabolic model in capturing long-term dynamic changes of network. It was short for AGP model. Simulation experiments show that AGP model is of great application prospect in the field of network service control based on its well-founded prediction of dynamic change rules of network traffic.