基于蚁群聚类算法的扇区复杂性分析

Analysis and Verification of Sector Complexity Based on Ant Colony Clustering Algorithm

为了准确评价扇区的复杂程度,提高空域精细化管理能力,针对传统K-means算法易产生局部最优解的缺陷,研究了基于蚁群聚类算法的扇区复杂性分类方法。首先选择进近管制扇区作为分析对象,构建可定量分析的复杂性指标体系,利用主成分分析法精简指标;然后通过蚁群聚类算法对多个扇区进行聚类分析,选择Silhouette指标评估聚类质量,最终得到扇区的复杂程度分类。以西安和杭州进近扇区为例,综合复杂程度将扇区分为3类,通过仿真软件验证了聚类结果和蚁群聚类算法的有效性和可靠性。该结果及方法可以为空域规划和管理起到辅助决策支撑作用。

In order to accurately evaluate the complexity of sectors and improve the ability of refined airspace management,a sector complexity classification method based on ant colony clustering algorithm was studied in view of the defect that the traditional K-means algorithm is prone to produce local optimal solutions.First,approach control sectors were selected as the analysis object,a quantitatively analyzed complexity index system was constructed,and principal component analysis was used to streamline the indicators.Then,ant colony clustering algorithm was used to perform cluster analysis on multiple sectors at different periods of time,and Silhouette index was chosen to evaluate the quality of clustering,and finally the classification of the complexity of the sector was obtained.Took the Xi’an and Hangzhou approach sectors as examples,the sectors were divided into three categories based on comprehensive complexity.The clustering results and the effectiveness and reliability of the ant colony clustering algorithm were verified by simulation software.The results and methods can play an auxiliary decision support role for airspace planning and management.