WSN中基于改进粒子群优化算法的分簇拓扑算法研究

发布时间：2018-04-12 22:07

本文选题：无线传感器网络 + 分簇拓扑　；参考：《郑州大学》2017年硕士论文

【摘要】：无线传感器网络(Wireless Sensor Network,WSN)是一项多种学科技术高度交叉的综合性前沿研究领域,受到社会各界的高度重视。鉴于WSN能量受限及多跳等特点,使得WSN路由协议成为WSN研究领域的难点。分层路由由于其低能耗且易扩展的特性已成为WSN路由协议的研究重点,而分簇拓扑技术作为分层路由的重要部分,亦备受重视。本文首先介绍一种基于梯度的有网关的分簇拓扑算法(a Hierarchical Clustering Algorithm based on Gradient with Gateway,HCAGG)。该算法综合考虑节点的剩余能量和平均距离两项因素选取簇头;在建立簇树阶段,网关节点的引入有效降低了较远簇头间直接通信的高能耗。但是该算法在簇头选举时未考虑簇内节点能量的分布,而网关节点的选取过分依赖于梯度。针对HCAGG算法中存在的问题,本文提出一种基于改进粒子群优化算法的分簇拓扑算法(a Clustering Topology based on Modified Particle Swarm Optimization,CTMPSO)。针对PSO算法易早熟收敛的缺点,首先提出一种改进粒子群优化算法MPSO,该算法引入人工鱼群算法中的视野及随机行为,增强粒子搜寻的多样性,避免粒子过早陷入局部最优,并通过Sphere函数和Ratrigrin函数验证了该MPSO算法的有效性。同时,结合WSN及MPSO算法特点,构建多局部最优觅食场景及适合离散WSN环境的连续评价函数,使MPSO算法更好的应用于WSN。此外,针对待优化目标特点,采用比较法逐步确定搜索区域。针对HCAGG中簇头的问题,综合考虑邻居节点剩余能量、到节点的平均距离以及簇内能量分布重新构建适应度函数,利用MPSO算法搜寻最佳节点担任新簇头。针对HCAGG中网关的问题,提出一种两跳内最优网关并结合MPSO算法对网关进行优化。实验结果表明,该CTMPSO算法能显著延长全网的存活期。最后给出CTMPSO算法节点上、下线的维护与更新策略,以提高算法的自适应性。同时,针对WSN新兴领域中要求节点移动的场景,给出CTMPSO算法移动情况下的维护更新策略,扩大算法的应用领域。
[Abstract]:Wireless Sensor Network (WSNs) is a comprehensive frontier research field, which is highly intersected by many disciplines and technologies, and is highly valued by all walks of life.Due to the characteristics of WSN energy limitation and multi-hop, WSN routing protocol becomes a difficult problem in the field of WSN research.Because of its low energy consumption and easy to extend, hierarchical routing has become the focus of WSN routing protocols. As an important part of hierarchical routing, clustering topology has been paid more attention.In this paper, we first introduce a Hierarchical Clustering Algorithm based on Gradient with Gateway-HCAGG algorithm based on gradients with gateways.The algorithm considers both residual energy and average distance of nodes to select cluster heads. In the stage of building cluster tree, the introduction of gateway node can effectively reduce the high energy consumption of direct communication between remote cluster heads.However, the algorithm does not consider the energy distribution of the nodes in the cluster when the cluster head is elected, and the selection of the gateway nodes is too dependent on the gradient.Aiming at the problems in HCAGG algorithm, this paper presents a Clustering Topology based on Modified Particle Swarm optimization algorithm based on improved particle swarm optimization (PSO) algorithm.In order to overcome the disadvantage of premature convergence of PSO algorithm, an improved particle swarm optimization algorithm (MPSOs) is proposed. The algorithm introduces the field of vision and random behavior of artificial fish swarm algorithm to enhance the diversity of particle search, and to avoid the particle falling into local optimum prematurely.The validity of the MPSO algorithm is verified by Sphere function and Ratrigrin function.At the same time, combined with the characteristics of WSN and MPSO algorithm, the multi-local optimal foraging scene and the continuous evaluation function suitable for discrete WSN environment are constructed, so that the MPSO algorithm can be applied to WSN better.In addition, according to the characteristics of the target to be optimized, the search area is determined step by step by comparison method.Considering the residual energy of neighbor nodes, the average distance to the nodes and the energy distribution within the cluster, the fitness function is reconstructed to solve the cluster head problem in HCAGG, and the MPSO algorithm is used to search for the best node as the new cluster head.Aiming at the problem of gateway in HCAGG, this paper presents a two-hop optimal gateway and optimizes the gateway with MPSO algorithm.Experimental results show that the CTMPSO algorithm can significantly prolong the lifetime of the whole network.Finally, the maintenance and update strategy of the CTMPSO algorithm node is given to improve the adaptability of the algorithm.At the same time, aiming at the scene of node moving in the emerging field of WSN, the maintenance and update strategy of CTMPSO algorithm is given to expand the application field of the algorithm.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5;TP212.9

【相似文献】