| The forecast to traffic,data flow is a very practical problem for the telecommunication service provider, it can produce a direct influence on the temporary expansion or reduction that the telecommunication service provider deal with the Voice Channels and the Data Service Channels during the holidays. Some or all of the GSM Voice Channels TCH are set as dynamic channels of GPRS / EDGE data services, so we had to consider the influence that the data flow carried in the dynamic channels to the configuration of the Voice Channels.This paper first briefly introduces the major research contents and innovations; Then, this paper introduces the detailed research to the forecasting arithmetic models which are commonly used in the current industry such as Index Regression Model, k quadratic polynomial regression model (Parabolic model), k third power polynomial regression model, k fourth power polynomial regression model, k fifth power polynomial regression model, LMS algorithm model (least mean square model), and integrates these algorithms into a united large program that gives comparison to the forecasting arithmetics, then take some massive historical data to validate forecast accuracy, so the best forecasting arithmetic is selected.Then again, this paper introduces the method that plot the data flow value (bytes) as equivalent to the value of traffic (Erl) through the model of Campbell proposed by China Mobile Group Design Institute Co., Ltd.; After that, aim at the contents that the project researches, we write a GSM / GPRS / EDGE system that be used to forecast the traffic and allocate carrier frequencies, and this paper introduces a part of functions of the system database connection public class-DBConn.cs, and the process improving on and perfecting the data management structure of traffic and the data import module of traffic of the system platform.Finally, this paper carries out a certain sum up, talks freely about the harvest that write paper and research topic. |
PDF Full Text Request