基于谱聚类的空域扇区分割  

Airspace sector partitioning based on spectral clustering algorithm

作　　者：王超[1] 王晓英[1] 王飞[1] 

机构地区：[1]中国民航大学空中交通管理学院,天津300300

出　　处：《计算机应用研究》2014年第6期1889-1893,共5页Application Research of Computers

基　　金：国家科技支撑计划资助项目(2011BAH24B08);中央高校基本科研业务费专项基金资助项目(ZXH2011A002;ZXH2012C005)

摘　　要：为解决当前空域扇区结构对交通流适应性差的问题,进行了空域扇区分割与边界优化算法的研究。根据管制空域的基本航线网络结构和空中交通流量,提出了顶点间联系程度的计算模型;结合顶点融合与边航线离散化方法,建立了空域运行图模型;应用谱聚类算法实现了空域图顶点的有效分割,解决了近距平行航路分割误差问题,实现了扇区凸壳的构造与划分;提出了基于MAKLINK图的扇区间边界优化路径选择算法。实例分析表明,该算法能根据空中交通流量需求进行扇区优化设计,运行时间较短,约为10 s。To solve the problem that the current airspace sector structure could not accommodate the current air traffic flow,this paper studied airspace sector partition and boundary optimization methods.According to the basic air route network structure of controlled airspace and air traffic flow therein,it proposed a connection model between vertices.Combined with vertices fusion and route discretization method,it established an airspace operation graph model.By spectral clustering algorithm it clustered airspace graph vertices,and solved the problem of bad partition for closely parallel routes,and constructed and partitioned sector convex hull.Based on MAKLINK graph theory,it proposed an optimal path selection algorithm of sector boundaries.Experients result shows that it can use the proposed algorithms to design and optimize airspace sectors according to the requirements of air traffic flow,and the calcultion time is about 10 s.

关 键 词：空域规划 扇区划分 图论 谱聚类 MAKLINK图 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]