Network traffic forecasting provides key information for network management, resource allocation, traffic attack detection. However, traditional linear and non-linear network traffic forecasting models cannot achieve enough prediction accuracy for future traffic prediction. In order to resolve this problem, a network traffic prediction method based on SA (Simulated Annealing) optimized ARIMA (Autoregressive Integrated Moving Average model)-BPNN (Back Propagation Neural Network) is proposed in this paper, which makes comprehensive use of linear model ARIMA, non-linear model BPNN and optimization algorithm SA. With enhancement of the BPNN global optimization ability, it can fully realize the potential of mining linear and non-linear laws of historical network traffic data, hence improving the prediction accuracy. This paper selects the historical network traffic data of two different sampling points in the WIDE project to predict, and utilizes the MAE(Mean Absolute Error), RMSE(Root Mean Square Error), and the MAPE(Mean Absolute Percentage Error) as the evaluation index of the prediction effect. Experimental results show that our proposed method outperformed traditional network traffic prediction model, with several improvements in network traffic prediction accuracy.

网络流量预测为网络管理，资源分配，流量攻击检测提供关键信息。但是，传统的线性和非线性网络流量预测模型无法为未来的流量预测获得足够的预测精度。为了解决这个问题，本文提出了一种基于SA（模拟退火）优化的ARIMA（自回归综合移动平均模型）-BPNN（反向传播神经网络）的网络流量预测方法，该方法综合利用了线性模型ARIMA ，非线性模型BPNN和优化算法SA。通过增强BPNN全局优化能力，可以充分挖掘历史网络流量数据的线性和非线性规律的潜力，从而提高了预测的准确性。本文选择WIDE项目中两个不同采样点的历史网络流量数据进行预测，并利用MAE（均值绝对误差），RMSE（均方根误差）和MAPE（均值绝对百分比误差）进行评估。预测效果的指标。实验结果表明，本文提出的方法优于传统的网络流量预测模型，在网络流量预测精度上有一些改进。