| 摘要: |
| 针对海上无线网状网通信环境复杂多变、船舶节点具有特殊移动模型等特点，提出一种基于Q-Learning的自适应路由（Q-Learning Based Adaptive Routing，QLAR）算法。综合考虑海上无线电波传播特性、船舶航程信息以及相应海区气象信息等因素的影响，提出链路可靠性、链路稳定性和节点航程相似度等概念，并对链路状态进行评估；然后，根据链路状态评估结果，利用Q-Learning算法寻找源、目的节点间最稳定的路径以传输数据分组；最后，利用OPNET搭建仿真平台对算法进行测试。仿真结果表明，与4种对比算法中性能最优的算法相比，QLAR算法最高可提升分组投递率4.89%，降低平均分组时延17.42%，减少归一化路由开销21.99%。 |
| 关键词: 海上无线网状网 自适应路由 Q-Learning 链路可靠性 链路稳定性 航程相似度 |
| DOI： |
|
| 基金项目:国家自然科学基金资助项目（61371091,61801074）；中央高校基本科研业务费专项资金资助（3132017078）；大连海事大学“十三五”重点科研项目（3132016318） |
|
| A Q-Learning Based Adaptive Routing Algorithm for Maritime Wireless Mesh Networks |
| ZHANG Qiang,CHEN Xiaojing,HE Rongxi,WANG Yuqing |
| (1.Information Science and Technology College,Dalian Maritime University,Dalian 116026,China;
2.School of Electrical Engineering,Dalian University of Science and Technology,Dalian 116052,China) |
| Abstract: |
| A Q-Learning based adaptive routing(QLAR) algorithm is proposed for maritime wireless mesh networks(MWMN) with the complex and variable communication environment and the special mobility model of ship nodes.With an integrated consideration of the characteristics of marine radio propagation,ship voyage information and meteorological information of corresponding sea area,the concepts of link reliability,link stability and ship voyage similarity are proposed,and the link status is evaluated.Moreover,according to the link status,the Q-Learning algorithm is used to look for the most stable path between the source and destination nodes for the transmission of data packets.Finally,the performance of the proposed algorithm is evaluated by using OPNET to build a simulation platform.The simulation results show that compared with the algorithm with the optimal performance in the four compared algorithms,the QLAR algorithm can mostly increase the packet delivery rate by 4.89%,decrease the average packet delay by 17.42% and reduce the normalized routing overhead by 21.99%. |
| Key words: maritime wireless mesh networks adaptive routing Q-Learning link reliability link stability voyage similarity |
