基于强化学习的WLAN节点自适应调整技术

Adaptive adjustment technology of WLAN node based on reinforcement learning

为提高WLAN无线网络节点的吞吐量,针对IEEE802.11无线网络中的速率自适应技术进行研究。基于一定程度的强化学习策略,提出多臂老虎机问题的速率自适应算法。与已提出的一些算法在信道稳定的情况下、信道逐渐变化的情况下、信道随机变化的情况下依次进行对比,观察它们的吞吐量变化情况,分析所提出算法性能的优劣性。仿真结果表明,当信道环境稳定或比较稳定时,该算法要优于其它自适应算法。

To improve the throughput of WLAN wireless network nodes,the rate adaptation technology in IEEE802.11 wireless networks was studied.Based on the certain extent reinforcement learning,one rate adaption algorithm of multi-armed bandit was proposed.Some algorithms were compared with the algorithm proposed in the environment of static channel,gradual channel variation and random channel variation.The performance of the proposed algorithm was analyzed by observing their throughput changes.Simulation results show that the algorithm proposed is superior to other adaptive algorithms when the channel environment is stable.