基于Q学习的无线传感网分簇拓扑控制算法  

A Clustering Topology Algorithm Based on Q-learning in WSN

作　　者：阎新芳[1] 王晓晓[1] 冯岩[1] 严晶晶[1] 

机构地区：[1]郑州大学信息工程学院,河南郑州450001

出　　处：《郑州大学学报（工学版）》2015年第2期85-88,共4页Journal of Zhengzhou University（Engineering Science）

基　　金：郑州市科技攻关基金资助项目(20120555)

摘　　要：为了延长大规模无线传感器网络的生命周期,在ETBG算法的基础上提出基于Q学习的分簇拓扑控制算法.该算法利用有序加权平均(OWA)算子多属性决策的方法确定节点的权值,利用Q学习算法对节点进行周期性的学习训练,按照每条路径的Q值进行最优路径的选择,然后就可以实现网络的拓扑控制.仿真分析表明,基于Q学习算法形成的簇树机制解决了ETBG算法在生成簇树过程中未能寻找到最佳路径而造成数据传输时能量损耗过多的问题,从而达到延长网络生命周期的目的.To prolong the lifetime of wireless sensor network, a Clustering Topology Algorithm Based on Q- learning in WSN （CTQL） is proposed on the basis of classical clustering algorithms such as ETBG. The Or- dered Weighted Average （OWA） operator multi-attribute decision making method is used to determine the weight of the nodes, and Q learning algorithm is used to periodically train the cluster heads. So the Q value of the optimal path is selected of this algorithm and the topology control is realized. Through simulation study shows that the use of Q-learning algorithm to resolve the problem that much energy consumption of ETBG algo- rithm fails to find the best path and CTQL effectively extend the network lifetime.

关 键 词：无线传感器网络 Q学习 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]