A Prediction and Learning Based Approach to Network Selection in Dynamic Environments

Abstract

The heterogeneous property in the next generation wireless network arises challenges of network selection problem. Existing approaches are mainly implemented in static network environments while cannot handle unpredictable dynamics in practice. In this paper, we propose a prediction and learning based approach, which considers both the fluctuation of radio resource and the variation of user demand. The network selection scenario is modeled as a multiagent coordination problem, in which a population of rational agents compete to maximize their benefits with incomplete information (no prior knowledge of network bandwidth and other users’ demands). Terminal users adaptively adjust their selections in response to the gradually or abruptly changing environment. The system is shown to converge to Nash equilibrium, which also turns out to be both Pareto optimal and socially optimal. Extensive simulation results show that our approach achieves significantly better performance compared with two existing approaches in terms of load balancing, user payoff and the overall bandwidth utilization efficiency.