基于猴群算法的传感器优化布置方法研究

发布时间：2018-04-09 18:30

本文选题：猴群算法　切入点：传感器优化布置　出处：《兰州交通大学》2016年硕士论文

【摘要】：传感器的优化布置是一类典型的组合优化问题。目前,传感器优化布置的方法有多种,但都存在各自的缺点。猴群算法是近年提出的一种智能仿生算法,适用于求解多变量、多峰值的函数优化问题。利用猴群算法进行传感器的优化布置,是目前国内外学者广泛关注和研究的热点问题之一。本文在总结猴群算法国内外研究现状及成果的前提下,对猴群算法进行了相应的改进,使其适应传感器优化布置的需要。本文研究的内容如下:(1)介绍了传感器优化布置的意义,对猴群算法的国内外研究现状进行了综述,总结了猴群算法的研究成果,给出了猴群算法改进和提高的方向,确立了传感器优化布置的数学模型。(2)针对猴群算法初始化种群随机性大、固定爬步长不利于搜索局部最优解的问题,提出了一种改进的猴群算法。该算法以MAC矩阵(Modal Assurance Criterion,模态置信矩阵)作为目标函数,通过正态分布的方法构造初始种群来增强猴群的多样性;采用自适应的变动爬步长,提高算法运行的速度和求解精度。(3)针对猴群算法跳区间固定、优秀猴子特征信息不能传承等缺陷,提出了野草猴群算法。该算法在改进的猴群算法基础上,采用自适应的跳过程,并引入以适应度为基准的野草繁殖进化和竞争生存机制,解决了优秀猴子特征信息的传承问题,进一步提高算法的求解精度。(4)猴群算法在附近区域进行最优解的搜寻时,难免存在搜索盲区,易导致某些最优解隐藏在步长覆盖的区域错失“良机”,降低了算法搜寻全局最优解的能力。针对该问题提出了基于蜂群采蜜行为的猴群算法。该算法在改进的猴群算法基础上,引入蜂群算法的采蜜行为,利用蜂群搜寻机制对所有区域进行搜索后,再将初步遴选出来的猴子进行猴群算法的基本搜索,改善了算法的搜索性能。(5)用8个测试函数及常用算法分别对上述3种改进后的算法进行测试分析,结果表明,改进后的猴群算法求解精度和收敛速度都得到了提高,算法性能改善明显。(6)建立了糊底机涂胶机构算例的有限元模型,通过上述3种改进后的算法对其进行传感器的优化布置方案选择,并对它们的特点进行了横向对比。
[Abstract]:The optimal arrangement of sensors is a typical combinatorial optimization problem.At present, there are many methods for optimizing sensor layout, but each has its own shortcomings.Monkey swarm algorithm is an intelligent bionic algorithm proposed in recent years, which is suitable for solving multivariable and multi-peak function optimization problems.The optimal arrangement of sensors using monkey swarm algorithm is one of the hot issues that scholars at home and abroad pay close attention to.On the premise of summarizing the research status and achievements of monkey swarm algorithm at home and abroad, this paper improves the algorithm to meet the needs of optimal sensor layout.The contents of this paper are as follows: (1) the significance of sensor optimization is introduced, the research status of monkey swarm algorithm at home and abroad is summarized, the research results of monkey swarm algorithm are summarized, and the direction of improvement and improvement of monkey swarm algorithm is given.The mathematical model of optimal sensor placement is established. (2) aiming at the problem that the initialization of the population is random and the fixed crawling step is not conducive to searching for the local optimal solution, an improved monkey swarm algorithm is proposed.In this algorithm, the MAC matrix Modal Assurance criteria (modal confidence matrix) is taken as the objective function, the initial population is constructed by normal distribution method to enhance the diversity of the monkey group, and the adaptive variable crawling step is used to enhance the diversity of the monkey group.To improve the speed and accuracy of the algorithm, a wild grass monkey swarm algorithm is proposed to solve the problems of fixed jump interval of monkey swarm algorithm, and the excellent monkey characteristic information can not be passed on.Based on the improved monkey swarm algorithm, the adaptive jump process is adopted, and the mechanism of weed propagation evolution and competitive survival based on fitness is introduced to solve the problem of the transmission of excellent monkey characteristic information.Further improving the accuracy of the algorithm. (4) when the monkey swarm algorithm searches for the optimal solution in the nearby area, it is inevitable that there are blind areas, which can easily lead to the missing "opportunity" of some optimal solutions hidden in the region covered by the step size.The ability of the algorithm to search for the global optimal solution is reduced.To solve this problem, a monkey colony algorithm based on honeybee behavior is proposed.On the basis of the improved monkey swarm algorithm, the honey collecting behavior of the bee colony algorithm is introduced. After all the regions are searched by the bee colony search mechanism, the monkeys selected from the preliminary algorithm are searched for the basic search of the monkey swarm algorithm.The search performance of the improved algorithm is improved. (5) eight test functions and common algorithms are used to test and analyze the above three improved algorithms. The results show that the precision and convergence speed of the improved monkey swarm algorithm are improved.The finite element model of the glue coating mechanism of the paste machine is established. Through the above three improved algorithms, the optimal arrangement scheme of the sensor is selected, and their characteristics are compared horizontally.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;TP212

【参考文献】