FCM聚类算法的改进及在GIS电缆管理系统中的应用

发布时间：2018-01-28 09:52

本文关键词： 电缆管理地理信息系统聚类分析 Hilbert R-树空间索引　出处：《江西理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：电力工业是国家经济发展的基础型资源,关系到经济的稳定上升和人民的日常生活。电缆作为电力系统的传输介质,它的安全高效运行是国民经济和居民生活质量的重要保障。传统的电缆管理存在着数字化程度不高,原有的资料信息不完整,人工普查信息位置不准确,电缆规划、建设和抢修维护存在盲目性等一系列弊端。地理信息系统(Geographic Information System,GIS)综合地图学以及遥感和计算机科学,在不同的场景有着重要的应用,特别是在电力线路管理当中。因此科学有效的建立基于GIS的电缆管理系统对供电公司电缆管理的意义十分重大。本文主要针对在基于GIS的电缆管理系统开发与应用研究过程中提出的新算法和新技术进行探讨研究。主要内容如下:(1)提出了一种基于灰狼优化的模糊C-均值改进聚类算法对电缆空间数据聚类为了解决模糊C-均值聚类(Fuzzy C-Means,FCM)存在易受初始随机产生的聚类中心影响和容易出现早熟收敛的问题,本文提出一种将灰狼优化算法(Grey Wolf Optimizer,GWO)和FCM相结合的新聚类算法(GWO-FCM)。使用GWO优秀的全局寻优性能对FCM的聚类中心进行优化,模拟灰狼优秀的搜寻猎物行为找到一组最佳的聚类中心来提高FCM的聚类效果。(2)提出了一种混合聚类的Hilbert R-树索引算法建立电缆空间数据索引机制针对Hilbert R-树索引算法在空间对象分布不均的时候,结点容易造成死空间和大量的重叠,特别是在叶子节点和其他结点的重叠面积增大时,导致出现多路查询,降低了检索的效率。提出了一种基于混合聚类的Hilbert R-树索引机制,将灰狼优化结合模糊聚类的混合聚类算法融合到Hilbert R-树索引中。具体方法是先利用Hilbert R-树对空间对象编码,然后用混合聚类方法对数据进行分类,来减少各结点的重叠和访问外存的OI/次数,从而提高检索效率。(3)开发了一套基于GIS的电缆管理系统为了满足电力电缆的科学管理,结合本文提出的理论创新,开发了基于地理信息系统技术的电缆管理系统,并且成功应用在了国家电网义乌供电公司。将结合混合聚类的Hilbert R-树索引应用到系统开发中。同时介绍了电缆管理系统的建设目标、系统结构设计、功能设计、开发环境及工具。分析了系统实现的关键技术、相应功能和应用价值。
[Abstract]:Electric power industry is the basic resource of national economic development, which is related to the steady rise of economy and the daily life of people. The cable is used as the transmission medium of power system. Its safe and efficient operation is an important guarantee of the national economy and the quality of life of residents. The traditional cable management is not digitalized, the original information is incomplete, and the location of artificial census information is not accurate. There are a series of disadvantages in cable planning, construction and repair maintenance. Geographic Information system (GIS) Information System. Integrated cartography, remote sensing and computer science have important applications in different scenarios. Especially in the power line management. So it is very important to establish the cable management system based on GIS scientifically and effectively. This paper mainly focuses on the cable management system based on GIS. The new algorithms and techniques proposed in the process of development and application are discussed and studied. The main contents are as follows:. (. 1) an improved fuzzy C-means clustering algorithm based on gray wolf optimization is proposed for cable spatial data clustering in order to solve fuzzy C-means clustering. In this paper, we propose a grey wolf optimization algorithm, grey Wolf Optimizer, which is easy to be affected by the initial random cluster center and prone to premature convergence. GWO) is a new clustering algorithm combined with FCM. The excellent global optimization performance of GWO is used to optimize the clustering center of FCM. Simulating the excellent hunting behavior of the gray wolf to find a set of best cluster centers to improve the clustering effect of FCM. In this paper, a hybrid clustering Hilbert R- tree index algorithm is proposed to build a cable spatial data indexing mechanism. When the spatial objects are unevenly distributed, the Hilbert R- tree indexing algorithm is proposed. Nodes are easy to cause dead space and a large number of overlaps, especially when the overlapping area of leaf nodes and other nodes increases, resulting in the emergence of multiple queries. The efficiency of retrieval is reduced. A Hilbert R- tree indexing mechanism based on hybrid clustering is proposed. A hybrid clustering algorithm based on gray wolf optimization and fuzzy clustering is integrated into Hilbert R- tree index. Firstly, Hilbert R- tree is used to encode spatial objects. Then the data are classified by hybrid clustering method to reduce the overlap of each node and the number of OI / access to external memory. In order to satisfy the scientific management of power cable, a set of cable management system based on GIS is developed. A cable management system based on GIS technology is developed. And successfully applied in Yiwu Power supply Company of State Grid. The index of Hilbert R- tree combined with hybrid clustering is applied to the system development. At the same time, the construction goal of cable management system is introduced. System structure design, function design, development environment and tools. The key technology, corresponding function and application value of system realization are analyzed.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13;TM75;P208

【参考文献】