User Behavior Analysis And Network Optimization Strategy In Measurement Report Based Mobile Networks

With the vigorous development of mobile Internet, the mobile application becomes inexhaustible, which leads to a high demand on the performance of the wireless access network, such as higher signal-to-interference-ratio (SIR). Because of the real-time feedback on the performance of mobile network, the measurement report technology makes it more realistic and effective response to the mobile network performance. At the same time, through large data mining technology, the relationship between the measurement report data and the performance of wireless access network it reflected can be mined through large data mining technology, which is helpful to enhance the performance of the mobile network optimization from the practical application aspect and user experience aspect. Meanwhile, the downlink ergodic rate can be improved combined with the self-organization network, which can ease the load pressure caused by tidal effect.In this thesis, the author gives a short introduction of measurement report technology, and designs an acquisition display system for the mobile network performance analysis through the research and improvement of the measurement report technology, which realizes the real-time acquisition, display and feedback analysis on the dynamic user behavior and the performance of the mobile network.In order to study the influence of the user distribution on the performance of mobile network, this thesis designs an improved K-means algorithm by mining the measurement report data. The particular region user distribution probability function of time is concluded and the user distribution model is established accompanied with simulation results and corresponding analysis. Meanwhile, the influence on the degree of SIR by collected data is analyzed by using the decision tree and the prediction model for SIR degree is established.To solve the problem found by the display system, two user access method are proposed and the downlink ergodic rates based on those use access method are derived. Though the simulation results, this thesis proves the correctness of formula and draw a conclusion that when the LPN density is low, it is better to use distance based user access method while it is better to use cluster based user access method when the LPN density is high. It is obtained that with the increase of user density, the ergodic rate goes down and this can be eased by the collaboration of LPNs.Based on the performance display system and self-organizing network, the mobile network can realize the self-check, self-configuration and self-optimization by using the two user access policy.