Abstract:

In recent years, researches on the complexity of air traffic become popular. How to select proper metrics to analyze the complexity is a key technology. In this paper, the importance of metrics selection to the complexity study is explained based on a brief review of previous researches. Combined with existing conclusions, a complexity metrics system is established by selecting metrics which are well-defined and can be quantitatively analyzed. The grey correlation cluster method is used to explore the pattern of the distribution of metrics and refine metrics for the characteristics of multi-dimensional and large amount of information. The refined metrics can both fully describe the problem and reduce the dimensions when analyze air traffic complexity. According to the radar data of 16th sector in Guangzhou region of China, a class of metrics is calculated and cluster analyzed, which verifies the accuracy of this method.

Key words: air transportation, air traffic complexity, refinement of metrics, metrics system, grey correlation cluster

摘要：

近年来，空中交通复杂性的研究蓬勃发展，如何选取指标分析复杂性成为一个关键技术.本文在对以往复杂性研究进行简单回顾的基础上，明确了指标选取对于复杂性研究的重要意义.结合已有成果，选取定义明确、可定量分析的指标构建了复杂性指标体系.针对指标体系维数多、信息量大等特点，选取灰色关联聚类方法挖掘指标数值的分布规律，实现对指标的提取精炼，从而在运用指标体系分析空中交通复杂性时，既能全面描述问题，又能降低指标维度.根据广州地区16扇区雷达数据，对一类指标进行了计算和聚类分析，验证了该方法的准确性.

关键词: 航空运输, 空中交通复杂性, 指标精炼, 指标体系, 灰色关联聚类