Neural Network Algorithms for Multi Step Ahead Prediction

Multimedia services became a major part of the internet network traffic. The bursty characteristics of the video traffic, produced by applications like video on demand, video broadcasting or videoconferencing, make it difficult to fulfill the Quality of Service (QoS) of the multimedia applications. Therefore it is important to utilize congestion control procedures. One of the procedures used to fulfill the QoS are traffic prediction and dynamic bandwidth allocation. Neural networks belong to vastly used tools for traffic prediction. In this paper, we propose three algorithms for multistep ahead prediction with the use of Nonlinear AutoRegressive model with eXogeneous inputs (NARX) neural network and the Multilayer Perceptron (MLP) based on separation of different frames, together with the prediction of difference values. At first we briefly describe the characteristics of the video traffic. Then we introduce theoretical fundamentals of the NARX neural network and multilayer perceptron. Then we describe the proposed algorithms and in the last section we present the results of video traffic prediction using the proposed algorithms for multi step ahead video traffic prediction. KeywordsNeural Networks; Nonlinear Autoregressive Model With Exogeneous Inputs; Multi Step Ahead Prediction; Video Traffic Prediction