基于改进粒子群聚类的无线传感器网络能量均衡分簇策略  

Clustering strategy for energy balance of wireless sensor networks based on improved particle swarm optimization clustering algorithm

作　　者：李洪兵[1,2] 余成波[1] 闫俊辉[1] 李彦林[1] 

机构地区：[1]重庆理工大学远程测试与控制技术研究所,重庆400050 [2]重庆三峡学院,重庆404000

出　　处：《计算机应用研究》2011年第2期657-660,共4页Application Research of Computers

基　　金：重庆市自然科学基金重点项目(CSTC2007BA2023);重庆市教委资助项目(KJ101107);重庆市九龙坡科技计划资助项目(九龙坡科委发[2009]52号);重庆市科技创新资助项目(渝经信科技[2010]9号)

摘　　要：针对无线传感器网络能量约束特点,为实现节点能耗均衡、最大化网络寿命,提出了一种基于改进粒子群聚类的无线传感器网络能量均衡分簇算法。首先根据距离汇聚节点远近将网络进行区域划分和等级标定,以不同概率确定不同等级区域的分簇数量和规模。在活动等级区域内引入相同数量的粒子,根据K-均值聚类法形成多个初始粒子群,修改带惯性权重的粒子群算法,修改粒子飞行规则,并行智能搜索聚类。多个粒子群体的总结学习等优点加快了聚类收敛速度,克服了对初始聚类中心点选择较敏感的问题,形成了传感器节点位置的最优分簇,避免了网络热点问题,促进了网络能耗均衡,最大化网络寿命。理论分析和仿真实验结果表明了本算法对网络节点能耗均衡分簇的有效性。According to energy constraints of WSN,this paper presented a clustering strategy for energy balance based on the improved particle swarm optimization clustering algorithm,in order to balance the nodes＇ energy consumption and maximize the network＇s lifetime.First,divided the WSN into some hierarchical regions according to the distance from the sensor nodes to the sink node.Adopted different probabilities in different hierarchical regions to determine the clustering number and size.Then introduced particles with the same number of the nodes to the active hierarchical region.It formed a number of initial swarm particles by K-means clustering method.The inertia weight-based particle swarm optimization algorithm was amended as well as the flying rules of the particles to parallel intelligent searching and clustering.The advantage of summarizing and learning the particle swarms speeded up the convergence and overcame the issues that the initial clustering centers were sensitive to the clustering results.It also avoided the hot issues of the WSN,balanced the network energy consumption,and maximized the lifetime of the network.Theoretical analysis and the simulation results show it＇s effectiveness to the energy consumption balance.

关 键 词：无线传感器网络 能量均衡 分簇策略 粒子群 K-均值聚类 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP212[自动化与计算机技术—计算机科学与技术]