一种基于深度Q学习的移动自组网稳定路由  

A deep Q learning-based stable routing for mobile ad hoc networks

作　　者：高翔 马少斌 张成文 GAO Xiang;MA Shaobin;ZHANG Chengwen(School of Digital Media,Lanzhou University of Arts and Science,Lanzhou 730000,China)

机构地区：[1]兰州文理学院数字媒体学院,甘肃兰州730000

出　　处：《南昌大学学报（工科版）》2022年第1期91-96,共6页Journal of Nanchang University(Engineering & Technology)

基　　金：甘肃省高等学校产业支撑引导项目(2019C-09)。

摘　　要：节点的移动降低了认知无线电-移动自组网(CR-MANETs)路由的稳定性。为此,针对CR-MANETs网络,提出基于深度Q学习的稳定路由(DQLR)。设计DQLR路由的目的在于最小化端到端队列时延和提高链路稳定性。DQLR路由先通过队列可用空间和链路的连通时间计算链路成本。然后,建立深度Q学习模型,并选择具有最低Q值作为下一跳转发节点,进而构建稳定路由。最后,利用SimPy仿真框架建立仿真平台,分析了DQLR路由的性能。仿真结果表明,相比于基于链路稳定的MANETs路由(LSR),提出的DQLR路由降低了队列时延和路由时延,并提高了数据包传递率。The movement of nodes reduces the stability of routing in cognitive radio-mobile ad hoc networks(CR-MANETs).Therefore,deep Q learning-based stable routing(DQLR)was proposed for CR-MANETs.The goal of DQLR was to minimize end-to-end queue delay and improve stability of link.DQLR routing first calculates the link cost by the available space in the queue and the connection time of the link.Then,the deep Q learning model was established,and node with lowest Q value was selected as the next hop forwarding node in order to construct stable routing.Finally,the simulation platform was established by using SimPy simulation framework,and the performance of DQLR routing was analyzed.Simulation results showed that,compared with stable link MANETs routing(LSR),the proposed DQLRS routing reduced the queue delay and routing delay,and improved the packet transfer rate.

关 键 词：认知无线电-移动自组网 稳定路由 深度Q学习 链路的稳定值 Q值 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]