改进粒子群算法及其在传感器网络定位中的应用

发布时间：2018-05-26 12:19

本文选题：群集智能 + 粒子群优化算法　；参考：《辽宁工程技术大学》2014年硕士论文

【摘要】：粒子群优化算法(PSO)是一种群集智能搜索算法,来源于对鸟类捕食的行为模拟和其模型构建。因其定义简单,为解决复杂优化问题另辟蹊径且十分有效,因此,许多学者对其十分关注,且经过研究,该算法已在众多领域得以广泛应用。但由于其理论还很不完善,还存在过早收敛的问题。到目前为止,为改善这些不足,许许多多的改进算法被提出。本文在此基础上,以提高算法性能为最终目的,也深入研究了关于PSO的改进方法,并将此方法成功应用于传感器网络的节点定位中。本文先是通过对粒子轨迹和算法收敛性的分析,分别对简化PSO系统和一般化PSO系统进行研究分析。另外对无约束的轨迹实例分析,更直观的描述了粒子的收敛性、周期性和离散性。基于此前分析找出致使过早收敛的成分和对其的解决方式。经过研究分析,得出算法的全局收敛条件和局部收敛条件。本文通过对基于惯性权重、学习因子、收缩因子、混合算法的改进PSO算法的研究,以传感器的节点定位为研究背景,充分利用混沌映射的优势,并结合PSO,提出了新的改进算法,仿真研究证实,该算法确实可以优化传感器网络的节点定位问题,并使其定位精度得以提高,定位速度得以加快,在传感器节点定位问题上是一种确实可行的解决途径。
[Abstract]:Particle Swarm Optimization (PSO) is a cluster intelligent search algorithm derived from the behavior simulation and modeling of bird prey. Because its definition is simple and it is very effective to solve complex optimization problems, many scholars pay close attention to it, and through research, the algorithm has been widely used in many fields. But because its theory is still very imperfect, there is still the problem of premature convergence. So far, in order to improve these deficiencies, many improved algorithms have been proposed. On the basis of this, and with the aim of improving the performance of the algorithm, this paper also deeply studies the improved method of PSO, and successfully applies this method to node localization in sensor networks. Firstly, by analyzing the particle trajectories and the convergence of the algorithm, the simplified PSO system and the generalized PSO system are studied and analyzed respectively in this paper. In addition, the convergence, periodicity and dispersion of particles are described more intuitively by the analysis of unconstrained trajectory examples. Based on the previous analysis, find out the components that lead to premature convergence and the solution to it. By studying and analyzing, the global and local convergence conditions of the algorithm are obtained. In this paper, an improved PSO algorithm based on inertial weight, learning factor, contraction factor and hybrid algorithm is studied. Based on the sensor node location, the advantage of chaotic mapping is fully utilized, and a new improved algorithm is proposed. The simulation results show that the proposed algorithm can effectively optimize the node location problem of sensor networks and improve the accuracy and speed of localization. It is a feasible solution to the problem of sensor node location.
【学位授予单位】：辽宁工程技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5;TP18

【参考文献】