基于强化学习的M2M通信上行链路节能优化算法  

Optimization algorithm of M2M communication uplink energy saving based on reinforcement learning

作　　者：李奇越[1] 周娜娜 柳传嘉 王建平[1] 孙伟[1] LI Qiyue;ZHOU Nana;LIU Chuanjia;WANG Jianping;SUN Wei(School of Electrical Engineering and Automation, Hefei University of Technology, Hefei 230009, China)

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2020年第7期913-918,991,共7页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(51877060)。

摘　　要：针对机器对机器(machine-to-machine,M2M)通信在加强的长期演进(long term evolution-advanced,LTE-A)网络上行链路能量效率问题,文章提出了一种基于强化学习的M2M通信上行链路节能优化算法。首先建立M2M通信能量效率模型,并将其重构为二维背包问题;然后使用强化学习的方法,引进并训练指针网络模型;最后通过主动搜索的策略解决该背包问题。仿真结果表明,相比于经典算法,当设备规模很大时,该算法性能更优,保证设备服务质量(quality of service,QoS)需求和公平性的同时,优化系统能效并降低数据的丢包率。For the uplink energy efficiency problem of machine-to-machine(M2M)communication in long term evolution-advanced(LTE-A)network,this paper proposes an uplink energy-saving optimization algorithm for M2M communication based on reinforcement learning.Firstly,the model with the highest uplink energy efficiency for M2M communication in LTE-A network is established and converted into a two-dimensional knapsack problem.Secondly,the pointer network model is introduced and then trained with the method of reinforcement learning.Finally,the active search strategy is used to solve the knapsack problem.The simulation results show that the proposed algorithm outperforms the classical algorithm when the device scale is large.The algorithm can cut down loss rate of data packet and optimize energy efficiency of system while ensuring the fairness and quality of service(QoS)requirements between M2M devices.

关 键 词：机器对机器(M2M) 加强的长期演进(LTE-A)网络 服务质量(QoS) 能量效率 强化学习 

分 类 号：TP393.01[自动化与计算机技术—计算机应用技术] TN929.531[自动化与计算机技术—计算机科学与技术]