基于指标体系的扇区复杂性评估方法

为了全面分析扇区复杂性，将其分解为结构复杂性和运行复杂性.借鉴已有的研究成果，围绕结构特征和运行特征，分别建立了多维指标体系.利用主成分分析方法提炼指标信息，评估扇区的结构复杂性和运行复杂性.最后采用k-means 聚类算法对多个扇区进行聚类分析，选取Dunn 指标评价聚类质量，实现了对扇区复杂程度的最佳等级划分，同时对复杂性指标分析结果进行了验证.实例表明，复杂性计算结果能够较好地体现多个指标的综合影响，区分不同扇区的复杂程度，聚类结果与实际情况相符.该结论可以为空域规划和管理提供参考意见.

An Evaluation Method of Sector Complexity Based on Metrics System

In order to analyze sector complexity comprehensively, this paper decomposes it into structure complexity and operation complexity. Multi-dimensional metrics focusing on airspace structure characteristics and traffic operation characteristics are constructed respectively according to present research. Primary component analysis is used to refine metrics information, from which structure complexity and operation complexity are evaluated. Multiple sectors are divided into different clusters by k-means clustering algorithm. The Dunn indicator is used to evaluate clustering results, which can help us decide the optimal clustering number. The clustering results are treated as the best classification for sector complexity, which could verify the metrics evaluation results. In the case of sector samples, sector structure complexity and operation complexity can well reflect the comprehensive effect of multiple metrics, distinguish complex degree of different sectors. The clustering results are consistent with the actual situations. These conclusions could provide recommendations for airspace planning and management.