A Learning Based Adaptive Network Selection Strategy In Dynamic Hetnet Environments

With the rapid growth and improvement of network technologies,the network structure is going towards increasingly heterogeneous.Meanwhile,the intelligent programming of the core network makes the available radio resource being more changeable rather than static.In such a dynamic and heterogeneous network environment,how to help terminal users select optimal networks to access is challenging.Prior implementations of network selection are usually applicable for the environment with static radio resource,while cannot handle the unpredictable dynamics in next generation network environments.To this end,this paper proposes a learning based adaptive network selection strategy in dynamic HetNet environments.It considers both the fluctuation of radio resource and variation of user demand.The access network selection scenario is modeled as a multiagent coordination problem,in which a bunch of rationally terminal users compete to maximize their benefits with incomplete information about the environment(no prior knowledge of network resource and other users' choices).Then a learning based strategy is proposed,which enables users to adaptively adjust their selections in response to the gradually or abruptly changing environment.The system is experimentally shown to converge to Nash equilibrium,which also turns out to be both Pareto optimal and socially optimal.Extensive simulation results show that the approach achieves significantly better performance compared with two learning and non-learning based approaches in terms of load balancing,user payoff and the overall bandwidth utilization efficiency.In addition,the system has a good robustness performance under the condition with non-compliant terminal users.To sum up,the learning based network selection strategy can well adapt to the dynamic network environment and ensure the load balancing among heterogeneous networks,which provides a reference for the dynamic heterogeneous network selection study.