USE OF SPLINE-EXTRAPOLATION TO IMPROVE THE QUALITY INDICATORS OF TELECOMMUNICATION SYSTEMS

Abstract

The problem of predicting self-similar traffic with significant and frequent ripples and the property of long-term dependence is considered. For short-term prediction of self-similar traffic, which is performed outside the considered time interval, a method of spline extrapolation based on various spline functions (cubic and cubic B-splines) is proposed. We recommend a method for estimating traffic prediction errors for each traffic prediction option using cubic or cubic B-splines. Comparison of the results of predicting self-similar traffic using various spline functions (cubic and cubic B-splines) show that the accuracy of the prediction can be improved by using cubic B-splines. According to the results of the study, it is established that cubic B-splines are characterized by high constructively, adequacy to the tasks and simplicity of practical implementation. Their use makes it possible to greatly simplify computational processes using formalized mathematical constructs. The use of the proposed method of spline extrapolation based on cubic B-splines has several advantages over other known methods, namely, ease of practical implementation, high accuracy of forecasting, the ability to accurately extrapolate peak "bursts" of traffic, which is especially important when solving problems in real time. The results of predicting self-similar traffic will allow the operator, at the stage of design and further operation of the mobile communication network, to provide the required amount of buffer devices for the network hardware and software, thereby avoiding network congestion and exceeding standard values of traffic service quality indicators. Proposed is the prospect of further research through the use of wavelet-extrapolation in order to improve the accuracy of prediction. Practical significance is determined by the developed recommendations for using the spline-extrapolation method in solving various problems of designing networks of various technologies and structures.