基于超限快速决策树的无线多媒体传感器网络路由优化方法  

Routing Optimization Method of Wireless Multimedia Sensor Network Based on Super Limit Fast Decision Tree

作　　者：刘松[1] LIU Song(Anhui Vocational and Technical College of Commerce and Trade E-commerce College,Wuhu241000,China)

机构地区：[1]安徽商贸职业技术学院电子商务学院,芜湖241000

出　　处：《青岛大学学报（自然科学版）》2022年第3期33-38,共6页Journal of Qingdao University(Natural Science Edition)

基　　金：安徽省高等学校自然科学研究重点项目(批准号:KJ2019A255)资助;安徽省质量工程“创新发展行动计划项目”(批准号:2016cx49)资助。

摘　　要：为优化无线多媒体传感器网络路由,降低路由优化过程中能量消耗,提出基于超限快速决策树的无线多媒体传感器网络路由优化方法。首先,构建无线多媒体传感器网络路由模型,确定变量参数;然后,采用随机森林递归特征消除算法从变量参数中选取关键变量,构建路由多目标优化模型。最后,以关键变量为超限快速学习机输入值,通过学习机训练输出路由优化结果,完成网络路由优化。研究结果表明,该方法数据传输速率最优,且有效提升节点能耗均衡度。In order to optimize the routing of wireless multimedia sensor networks and reduce the energy consumption in the process of routing optimization,a routing optimization method for wireless multimedia sensor networks was proposed which was based on overrun fast decision tree.Firstly,the routing model of wireless multimedia sensor networks was constructed to determine the routing variable parameters of wireless multimedia sensor networks;Then,the random forest recursive feature elimination algorithm in the decision tree algorithm was used to select the key variables from the variable parameters to construct the routing multi-objective optimization model.Finally,taking the key variable as the input value of the overrun fast learning machine,the routing optimization results were output through the learning machine training to complete the network routing optimization.The results show that this method ensures the optimal data transmission rate in the experimental object and effectively improves the energy balance of nodes.

关 键 词：无线多媒体传感器 能量消耗 超限快速决策树 路由模型 递归特征消除 关键变量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]