Research On Optimization Algorithms Of Handover Management In Dense Networks

The Densification of network is an important trend of the next generation mobile communication system,and is also one of the key technologies.With the intensive deployment of a variety of low-power nodes,the network becomes increasingly complex,which improves capacity of the network,but also brings great challenges on handover management.Therefore,how to achieve the users’ smooth movement is a hot spot of the next generation mobile communication system.To this end,the handover management algorithms in dense network is focused researched in the thesis,and the main contents are as follows:1.An adaptive handover algorithm based on RSRP prediction and loadTo solve the problem that the users are more prone to handoff frequently due to the small radius and intensive deployment of the small base station in dense network,an adaptive handoff algorithm based on the RSRP prediction and load is proposed in this thesis.The algorithm not only considers the load of networks but also considers the RSRP from the users’ side and concerns the influence of the changes in RSRP on the handoff deicision.Firstly,the adaptive least squares method is utilized to predict the RSRP value of each candidate base station(including the Macro base station and the small base station)in the algorithm.Then,the corresponding SINR is estimated according to the predicted RSRP value and the candidate base stations are filtered on the basis of the load utility value of the base station and the estimated SINR as well as the predicted RSRP.Finally,the user selects the base station with highest throughput as the target base station for switching.The simulation results show that the proposed algorithm can obtain lower interrupt probability and ping-pong rate and the higher throughput in comparison with the current switching algorithm.2.The handoff management solution based on learningAiming at the problems that the handover performance and throughput decrease due to the different coverage of various base stations,a handoff management algorithm solution based on learning is proposed for dense heterogeneous networks.The program includes a new context-aware scheduling mechanism and two learning-based handoff management algorithms: one is MAB(Multi Armed Bandit)-based learning algorithm and the other is satisfaction-based learning algorithm.In the scheme,the optimization model of the total transmission rate of the base station(including the Macro base station and the Pico base station)is firstly established.Then,the proposed two kinds of learning-based mobility management algorithms and the proposed new scheduling mechanism are combined for autonomous optimization.The simulation results show that,regardless of the user’s handoff performance and throughput,the scheme has better performance gain compared with the traditional handover management scheme.