Research On QoE-oriented Wireless Resource Allocation Scheme Over Hyper Dense Wireless Network

With the development of the wireless network,voice service,web browsing,file download and video services become more and more popular in people's daily life.Meanwhile,the rapid growth of intelligent terminals,and network will be hyper dense deployed.It is hoped that people have almost the same experience in the hyper dense scenario.In the hyper dense deployment scenarios,the interference of wireless resource is more serious;meanwhile,wireless resource's scarcity,the dynamics and complex of the wireless environment pose challenges to the wireless resource allocation.In order to make it easier for hyper dense network's wireless resource management,this paper introduces the centralized control-based hyper dense wireless network architecture.For hyper dense deployment scenarios,several typical services' related issues are studied,this paper has done the following researches:1.With the QoE(quality of experience)sensing function of the controller in software defined hyper dense wireless network architecture,this paper has analyzed the classification of different services QoE assessment methods and finished the research in modeling and simulation of VOIP(Voice Over IP and VOIP Protocols),web browsing,file download and video services QoE assessment algorithms.2.In the hyper dense wireless network,some users in overlap region of adjacent cells can be served by the adjacent cells.In order to optimize all the users' wireless resource allocation,this paper has proposed a QoE-oriented wireless resource allocation scheme over hyper dense wireless network.And we use the cross-layer architecture which considers the parameters in application layer,media access control layer and physical layer jointly.Then,we build the mathematical model of the resource allocation problem.Based on the Binary Constrained Particle Swarm Optimization(B_CPSO)algorithm,the optimization results can be found.In addition,Simulation results show that the proposed scheme have better performance than the other schemes in terms of maximizing average users' MOS value for the whole system as well as maintaining fairness among users.