【摘要】 在异构Macro-femto蜂窝网络中,随着日益增长的用户数量使得基站能耗问题变得更加严峻,提升整个移动系统能效的有效方式就是进行基站休眠。根据无模型理论提出一种基于Sarsa学习的动态基站休眠算法,算法通过基站学习环境中的用户流量,制定合理的休眠机制。仿真结果表明,提出的基于Sarsa学习的基站休眠算法能够有效提升系统能效。更多还原

【Abstract】 With the increasing mobile users, the problem of base station energy consumption becomes more serious in macro-femto heterogeneous cellular network. The technology of base station sleep is an effective way to improve the energy efficiency of the mobile system. In the paper, a dynamic base station sleep algorithm based on Sarsa learning is proposed based on model-free theory. In this algorithm, the base station learns the user traffic in the environment and interacts with other base stations to formulate a reasonable sleep mechanism. Simulation results show that the base station sleep algorithm based on Sarsa learning proposed in this paper can effectively improve the system energy efficiency.更多还原