虚占时刻航班异常延误行为研究

【作者】 梁丽琴；

【导师】 邵培基；

【作者基本信息】 电子科技大学 ， 信息管理与电子商务， 2010， 硕士

【摘要】 随着空中交通流量的日渐增长,开通航线的不断增多,航班延误问题成为了制约着民航业的发展的一大难题。而航空公司虚占时刻的航班异常延误行为,在一定程度上不仅加剧了航班的拥堵状况,加大了机场的流量压力,而且造成了航空时刻资源的巨大的浪费。因此,如何科学、有效地检测和规避航班的这种虚占时刻行为,成为民航重点关注又亟待解决的重要问题。目前在我国民航信息系统中储存着海量的航班运行历史数据,但其应用只是实现数据的简单查询、统计功能,尚未能有效挖掘数据中隐藏的关联和规则。而针对虚占时刻的航班异常延误行为的分析,也主要凭经验来判断,缺乏科学的理论依据和有效的决策指导方法。本文借鉴相关领域的研究成果,针对航班延误的特点,结合数据挖掘的理论和思想,构建了虚占时刻航班异常延误行为研究综合模型。该综合模型融合了基于决策树的异常航班检测模型和和基于关联规则的异常航班机场关联影响模型。通过基于决策树算法的异常行为检测模型,可有效识别潜在具有虚占时刻异常延误行为倾向的航班;通过构建异常航班对机场的关联模型,以深入分析虚占时刻的异常延误航班对机场造成的影响。本文借助智能的数据挖掘工具,采用民航航班延误的真实数据,对提出的虚占时刻航班异常延误行为研究理论模型进行了实证研究,并总结了虚占时刻的航班异常延误行为的三点表象特征和两点数据特征,提出了四方面的行为管理策略建议。本文以商务智能、数据挖掘技术的相关理论和思想为指导,以大量文献资料的搜集和企业的调研访谈为基础,综合采用定性与定量分析,比较评价分析,实证与理论、管理与技术相结合等多种科学研究方法来展开研究。其研究平台和工具主要包括Microsoft SQL Server 2005、SPSS 16.0和Clementine 11.0。本文的研究思路可为航空领域构建相关的异常问题的检测模型提供借鉴,其研究成果可为促进民航利用科学的方法检测航班异常延误行为,进而采取有效的管理措施预防虚占时刻行为的发生提供决策的参考。更多还原

【Abstract】 With the growing air traffic, increasing airlines, flight delay become a major problem that restricts the development of aviation industry. As an abnormal flight delay phenomenon, the virtual occupation moment behavior not only exacerbated the congestion status of flights, but also increasing the airport’s traffic pressure and the waste of flight resources.Therefore, how to detect and avoid virtual occupation moment behavior scientifically and effectively, has becomes an important issue need to be solved immediately.Though there are amounts of flight history data in the information systems of civil aviation of China, but its application is only a simple query and statistics for data. And the analysis of occupation moment behavior was based on the industry experience, lacking of scientific and effective method for decision-making guidance.This paper presents an integrated model to research the characteristics of occupation moment behavior. This integrated model consists of two basic models, which are decision-tree-based anomaly detection model to the name of potential abnormal flight, and association rules-based model to explore the effect of correlation between the abnormal flight and airport.After the empirical study of the integrated theory model with the intelligent data mining tools, this paper summarizes the abnomal flight characteristics of virtual occupation moment behavior, and suggests some management strategies for this problem.This paper adopts many methods and tools for the research, such as comprehensive qualitative and quantitative analysis, empirical and theoretical study, combining management and technology.The research platform and tools include Microsoft SQL Server 2005, SPSS 16.0 and Clementine 11.0.The research thinking and methodology can be built for the related anomaly detection issues in aviation industry. And the results can be used to detect the abnormal behavior of virtual occupation moment, and then take effective management measures to prevent this behavior and regulate the behavior of air-company, to some extent to improve the situation of delayed flights and reduce the waste of flight moment resourses.更多还原