基于改进果蝇算法的无线传感器网络覆盖优化研究

发布时间：2018-01-20 06:33

本文关键词： 无线传感器网络覆盖优化可变步长果蝇算法传感器节点　出处：《安徽大学》2017年硕士论文　论文类型：学位论文

【摘要】：无线传感器网络是一种分布式传感网络,是由大量固定或移动的无线传感器节点以自组织和多跳传输的方式组成。传感器节点采集的监测数据,可以通过逐跳的方式在多个节点之间进行传输。无线传感器网络具有网络设置灵活、网络服务质量高等优点,因此广泛应用于军事、智能交通、环境监控、医疗卫生等多个领域。在传统的无线传感器网络中,网络覆盖和节点部署等技术已经获得很多的研究成果,但随着网络通信技术的快速发展,人们对于无线传感器网络的需求变得更大。传统的节点部署策略就会出现部署速度慢,覆盖范围小,服务质量差等问题。无线传感器网络节点部署主要分为可移动传感器节点的网络覆盖和固定位置传感器节点的网络覆盖,这两种节点部署方式都存在一些相同的问题。例如:有些区域的节点过于密集,造成网络信号覆盖的亢余,而有的区域节点过于稀疏,造成该区域信号强度不够,成为网络盲区。于是,为了提高网络覆盖率和网络服务质量,通常就会通过增加节点数量的方式来实现,结果造成一些节点冗余,资源的利用率降低,网络结构变复杂,系统能耗变大等问题。本论文针对这两种节点部署方式,运用一种改进的果蝇算法,实现对无线传感器网络覆盖的优化。目前已有多种智能算法运用在无线传感器网络的覆盖优化问题上,例如粒子群算法、鱼群算法、遗传算法等。但是这些算法在无线传感器网络问题上,或算法复杂度高,导致计算速度太慢,或算法性能差,导致计算结果精度太低,或算法参数太多,导致网络模型复杂。针对这些问题,本文将改进的果蝇算法与无线传感器网络的两种覆盖模型结合,通过对比试验,验证在无线传感器网络覆盖优化问题上,本文的解决方案优于以往的解决方案,实现对网络覆盖的进一步优化。本文主要的工作集中于以下几点:1、提出一种改进的果蝇算法:可变步长果蝇算法。算法将整个搜索过程分为若干个周期,这样做可以增加搜索过程的多样性,大大减小局部收敛的可能性。其次算法在每个周期内采用Sin(x)函数,使步长在单位周期T内可以跌宕变化。这样既能保证算法有很强的全局搜索能力,可以实现快速收敛,又能使算法可以在小范围内完成高精度的搜索,结果具有更好的收敛效果。2、使用多个经典测试函数对可变步长果蝇算法的性能进行检测,体现算法在寻优问题上的有效性和优越性。通过实验结果的展示与分析,验证了相对于其它几种智能算法,可变步长果蝇算法具有更好的搜索性能和更高的稳定性。3、针对可移动传感器节点的网-络覆盖,首先建立网络模型,然后结合可变步长果蝇算法提出优化流程,在仿真环境下进行模拟实验,体现优化方法的有效性和优越性。通过一系列的对比试验和数据展示,验证了相对于其它智能算法,可变步长果蝇算法能更有效的结合可移动节点网络覆盖模型,进一步提高网络的覆盖率,实现对网络覆盖的优化。4、针对固定位置传感器节点的网络覆盖,首先建立网络模型,然后结合可变步长果蝇算法提出优化流程,在仿真环境下进行模拟实验,体现优化方法的有效性和优越性。通过一系列的对比试验和数据展示,验证了相对于其它智能算法,可变步长果蝇算法能更有效的结合固定位置节点网络覆盖模型,进一步提高网络覆盖率并降低网络能耗,实现对网络覆盖的优化。
[Abstract]:Wireless sensor network is a distributed sensor network is composed of a large number of fixed or mobile wireless sensor nodes in a self-organized and multi hop transmission mode. The monitoring data collected by sensor nodes, can be transmitted between a plurality of nodes through hop by hop. Wireless sensor network has set up a flexible network, higher quality of network service the advantages, it is widely applied to the military, intelligent transportation, environmental monitoring, medical and health fields. In traditional wireless sensor networks, network coverage and node deployment technology has obtained research results very much, but with the rapid development of network communication technology, the wireless sensor network needs to get bigger. The traditional node deployment strategy will be deployed to slow, the coverage is small and the problem of poor quality of service. The wireless sensor network node deployment is divided into Can the network coverage of mobile sensor nodes and the fixed position of the sensor node network coverage, the two nodes are some of the same questions. For example: nodes in some areas is too dense, resulting in more network coverage, and some regional nodes is too sparse, the signal intensity is not enough, the network become blind. So, in order to improve the efficiency and quality of network service overlay network, usually achieved by increasing the number of nodes, resulting in some redundant nodes, the utilization rate of resources is reduced, the network structure is complicated, the energy consumption of the system change and other issues. This thesis focuses on the two kinds of node deployment, using an improved algorithm of Drosophila and optimize coverage of wireless sensor networks. There are many intelligent algorithms used in the coverage problem of wireless sensor networks, such as particle swarm optimization Method, fish swarm algorithm, genetic algorithm and so on. But these algorithms in wireless sensor networks, or the complexity of the algorithm is high, so the calculation speed is too slow, or the algorithm performance is poor, which results in low precision, too much or cause the algorithm parameters, the models of complex networks. According to these problems, this paper will cover two the improved algorithm and the Drosophila model combined with the wireless sensor network, through the contrast test, verify the coverage problem in wireless sensor networks, the solution is better than the previous solutions, to further optimize the network coverage. This paper mainly focuses on the following points: 1, this paper proposes an improved algorithm: Drosophila a variable step algorithm. The algorithm will search the entire Drosophila process is divided into several periods, this can increase the diversity of the searching process, greatly reduce the possibility of local convergence. The second algorithm in each The Sin cycle (x) function, can make step ups and changes during the period T units. This can ensure the algorithm has strong global search ability, can achieve fast convergence, and can make the algorithm can achieve precision search in a small range,.2 has better convergence effect, the use of multiple the classic test functions on the performance of variable step algorithm for detection of Drosophila, reflect the efficiency and superiority of the algorithms in the optimization problem. And through the analysis of experimental results, verified compared with other several intelligent algorithms, a variable step search algorithm has better performance in Drosophila and higher stability of.3 for mobile sensor nodes the network coverage, first establish the network model, and then combined with the variable step algorithm is proposed to optimize the process of Drosophila melanogaster, a simulation experiment was carried out in the simulation environment, reflect the effectiveness and superiority of the optimization method. Through a series of experiments and data display, verified compared with other intelligent algorithms, a variable step algorithm with Drosophila more effective mobile node network coverage model, further improve the network coverage, to realize the optimization of.4 network coverage, according to the fixed position of the sensor node network coverage, network model is firstly established, then combined with the variable step algorithm is proposed to optimize the process of Drosophila melanogaster, a simulation experiment was carried out in the simulation environment, reflect the effectiveness and superiority of the optimization method. Through a series of experiments and data display, verified compared with other intelligent algorithms, a variable step algorithm with fixed position of Drosophila node network coverage model more effectively, further improve the network the coverage rate and reduce the energy consumption of the network, to realize the optimization of network coverage.

【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5

【参考文献】