基于Fisher聚类算法的航班飞行高度偏差分析研究  

A Study on Flight Altitude Discrepancy Base on the Fisher Ordinal Samples Cluster Method

作　　者：俞萍[1,2] 池宏[2] 谭显春[2] 许保光[2] 刘磊[2] 

机构地区：[1]中国科学技术大学管理学院,安徽合肥230026 [2]中国科学院科技政策与管理科学研究所,北京100190

出　　处：《中国管理科学》2010年第5期130-136,共7页Chinese Journal of Management Science

摘　　要：目前计算机飞行计划(Computerized Flight Plan,CFP)中的计划飞行高度与实际飞行高度之间存在偏差。为了能识别出航路中易发生显著偏差的子航段,便于对其进行重点监控,本文提出一种基于Fisher有序样本聚类算法的航班飞行高度偏差分析方法,采用修正的Fisher有序样本聚类算法对航路进行分段,在分段的基础上,对每个子航段上的高度偏差平均水平及波动情况进行统计分析,作为高度偏差的评价指标。分析结果可帮助识别航路上易发生显著实际与计划飞行高度偏差的子航段,为签派人员制作飞行计划以及在机组人员进行航前准备提供决策依据,具有较好的实践意义。There now exists a discrepancy between the practical flight altitude and the one calculated by CFP（Computerized Flight Plan）,which is unfavorable to security as well as cost-saving of flights.Along air route,flight altitude discrepancy is variable between segments.According to this character,this paper studies the flight altitude discrepancy based on air route segmentation.The study applies the weighted Fisher ordinal samples cluster method to cluster around those air route points with similar discrepancy level,meanwhile keeping their original order in air route,consequently carrying out air route segmentation.The outcome can tell on which air route segments the discrepancy is striking and needs to be given special attention,which is of benefit to improve the rationality of flight plan（takeoff fuel） and helps pilots support takeoff preparation.

关 键 词：航班运控管理 飞行高度偏差 飞行计划 Fisher有序样本聚类 

分 类 号：F562.6[经济管理—产业经济] V355[航空宇航科学与技术—人机与环境工程]