Abstract:

Efficient network traffic prediction method can improve the efficiency of network management. On account of problems of network traffic if as burst, time-varying, nonlinear problems happen that caused by various coefficients, a distributed network traffic prediction method was proposed obeyed by intelligent optimization. The fruit fly optimization algorithm was adopted in this method to optimize the smoothing coefficients of traditional triple exponential smoothing forecasting model. By predicting network traffic that is collected within time windows, this method effectively improves the efficiency of network traffic prediction. Simulation indicates that, compared with traditional triple exponential smoothing forecasting model, the proposed prediction model can solve the problem of prediction error caused by smoothing coefficient. The optimal smoothing coefficient can be selected adaptively, thus improves the prediction accuracy.

Key words: traffic prediction, fruit fly optimization algorithm, exponential smoothing