基于双向强化学习与动态码率调节无线Mesh网络协议.PDFVIP

优先出版 计 算 机 应 用 研 究 第33 卷 基于双向强化学习与动态码率调节的无线Mesh 网络协议* 1 1 2 陈永建 ，周 艳 ，崔海宁 (1．肇庆学院 电子信息与机电工程学院，广东 肇庆 526061 ；2 ．吉林大学 物理学院，长春 130012) 摘 要：针对现有无线mesh 网络协议的用户体验质量(QoE)较差的问题，提出一种基于双向强化学习与动态码率调节的 无线mesh 网络协议。首先，设计了兼容不同服务类型的无线Mesh 网络QoE 度量框架；然后，设计了基于双向强化学 习的无线mesh 网络路由协议；最终，结合QoE 感知的差异化报文调度策略与数据流源节点码率动态调节算法进一步优 化终端用户的QoE 质量。基于NS-2 仿真平台的对比实验结果显示，本协议可明显地提高无线mesh 网络的QoE 指标， 同时具有较低的控制开销。 关键词：无线Mesh 网络；移动互联网；强化学习；用户体验质量；路由协议；码率调节 中图分类号：TP393.02 Bidirectional reinforcement learning and dynamic data rate adjustment based wireless Mesh network protocol 1 1 2 Chen Yongjian , Zhou Yan , Cui Haining (1. Faculty of Electronic Information Mechatronic Engineering, Zhaoqing University, Zhaoqing Guangdong 526061, China; 2. College of Physics, Jilin University, Changchun 130012, China) Abstract: Concerned the problem that the existing wireless mesh network(WMN) protocols show pool performance in Quality of Experience(QoE), a bidirectional reinforcement learning and dynamic data rate adjustment based wireless WMN protocol is proposed. Firstly, different types of services compatible QoE measure framework of WMN is designed; then, bidirectional reinforcement learning based WMN routing protocol is designed ; lastly, QoE aware differential packets scheduling strategy and dynamic adjustment for data rate in source nodes of data flow are combined to optimize the QoE quality of terminal user further. NS-2 simulation platform based compared experimental results show th