空中交通复杂性指标体系精炼方法研究

为了准确评估空中交通复杂性,需要构建多维复杂性指标体系并对指标进行精炼.借鉴已有的研究成果,构建指标体系,使用K-means聚类方法对指标进行聚类分析,与初始指标分类相互验证,最后使用主成分分析方法提炼指标内涵.以广州地区16扇区的航班数据为例,发现聚类结果不仅与经验结果一致,还可以对指标进行更精细的划分；指标经提炼后保留了98%的信息,表达维度大大降低,从而验证了指标精炼方法的实用性和有效性.

Research on Refinement Method of Air Traffic Complexity Metrics System

In order to evaluate air traffic complexity accurately,a multi-dimensional metrics system should be constructed and refined.Firstly,according to present research,build a metrics system.Then the K-means clustering algorithm is used to analyze metrics which can validate its results to initial classifications of metrics.Finally meanings of metrics are refined with the method of primary component analysis.In the case of the flight data of 16 th sector in Guangzhou region,clustering results are not only in accord with experience,but also divided more accurately and meticulously.Refined metrics hold 98 percent of information and reduce original metrics＇ dimensions.Results verify the practicality and effectiveness of this method.