Evaluation of statistical models in 3G and 4G networks

Abstract: This report evaluates the performance of the statistical models based on queuing theory which are used to enhance simulations of downlink packet data traffic in mobile networks. The models are used to estimate user throughput based on system load in the cells of the simulation and the aim of the report is to conclude whether or not these estimations are representative for the traffic occurring in existing mobile networks. By collecting data from measurements in live networks and comparing the measured cell loads and number of users with the values estimated by the models, the performance of the models has been evaluated. The analysis shows that the user estimations made are representative for existing networks with low loads, but become increasingly inaccurate with increased load in the networks’ cells. This is caused in part by the model making unrealistic assumptions and in part by the model simply being unfit to estimate the number of users at very high loads.