电子学报 ›› 2018, Vol. 46 ›› Issue (10): 2325-2332.DOI: 10.3969/j.issn.0372-2112.2018.10.003

• 学术论文 • 上一篇    下一篇

一种基于Q-Learning策略的自适应移动物联网路由新算法

张德干1,2, 葛辉1,2, 刘晓欢1,2, 张晓丹3, 李文斌1,2   

  1. 1. 天津理工大学天津市智能计算及软件新技术重点实验室, 天津 300384;
    2. 天津理工大学计算机视觉与系统省部共建教育部重点实验室, 天津 300384;
    3. 中国科学技术信息研究所, 北京 100038
  • 收稿日期:2017-08-16 修回日期:2017-12-09 出版日期:2018-10-25
    • 通讯作者:
    • 刘晓欢
    • 作者简介:
    • 张德干,男,1970年1月出生,湖北黄冈英山县人.博士,教授,博导.主要研究方向为物联网、移动计算、网络通信、智能控制等.E-mail:zhangdegan@tsinghua.org.cn;葛辉,男,1993年3月出生,山东济南人.天津理工大学硕士研究生.主要研究方向为物联网、网络通信等.E-mail:1464736574@qq.com;张晓丹,女,1972年5月出生,吉林通化人.中国科学技术信息研究所研究员.研究方向为机器学习等.E-mail:183650594@qq.com;李文斌,男,1991年6月出生,天津市人.天津理工大学硕士研究生.主要研究方向为网络通信.E-mail:549411835@qq.com
    • 基金资助:
    • 国家自然科学基金 (No.61571328); 天津市重大科技专项 (No.15ZXDSGX00050,No.16ZXFWGX00010); 天津市科技支撑重点项目 (No.17YFZCGX00360); 天津市自然科学基金 (No.15JCYBJC46500); 天津市科技创新和131人才团队 (No.TD12-5016,2015-23,No.TD13-5025)

A Kind of New Routing Algorithm with Adaptivity for Mobile IOT Based on Q-Learning

ZHANG De-gan1,2, GE Hui1,2, LIU Xiao-huan1,2, ZHANG Xiao-dan3, LI Wen-bin1,2   

  1. 1.Tianjin Key Lab of Intelligent Computing & Novel Software Technology, Tianjin University of Technology, Tianjin 300384, China;
    2.Key Laboratory of Computer Vision and System (Tianjin University of Technology), Ministry of Education, Tianjin University of Technology, Tianjin 300384, China;
    3.Institute of Institute of Scientific and Technical Information of China, Beijing 100038, China
  • Received:2017-08-16 Revised:2017-12-09 Online:2018-10-25 Published:2018-10-25
    • Corresponding author:
    • LIU Xiao-huan
    • Supported by:
    • National Natural Science Foundation of China (No.61571328); Key Science and Technology Project of Tianjin Municipal (No.15ZXDSGX00050, No.16ZXFWGX00010); Key Program of Tianjin science and technology support project (No.17YFZCGX00360); Natural Science Foundation of Tianjin Municipality,  China (No.15JCYBJC46500); Tianjin Science and Technology Innovation and 131 Talent Team (No.TD12-5016, 2015-23, No.TD13-5025)

摘要: 针对移动物(车)联网的路由问题,通过对车辆的运动特点及造成链路断裂的原因进行的详细分析,我们建立了链路维持时间模型,并将维持时间作为设计路由算法的重要参数.Q-Learning作为一种启发式机器学习策略,能够通过与周围环境交互来动态地调整路由路径.基于此,我们设计了一种自适应的路由新算法.它将学习任务分散在每一个车辆节点中,通过周期性的与周围节点交换信标信息来维护可靠的路由路径.利用NS-2模拟器对该算法的性能进行了评估,结果表明,在不同的网络场景中,该算法在递交率、端到端的延时以及平均跳数等方面均表现出很好的效果.

关键词: 机器学习, 移动物联网, 拓扑, 动态, 路由

Abstract: In order to solve the routing problem of mobile IOT (IOV), based on our analyzing the details about motion characteristics of the vehicle and the reasons that cause links down, we set up link model of the duration time and using the duration time as key parameter to design the new routing method. Q-Learning as a kind of heuristic machine learning strategy is able to dynamically adjust the routing path through interaction with the surrounding environment. So a kind of new routing algorithm with adaptivity for mobile IOT based on Q-learning has been presented in this paper. It distributes the learning task into each vehicle node and maintains the reliable routing path by continuously exchanging the beacon information with the neighbor nodes. With the NS-2 simulator, the performance of the algorithm is tested. The results show that it has better performances on delivery, end-to-end delay and average hops in many mobile applications.

Key words: machine learning, mobile IOT, Topology, dynamic, routing

中图分类号: