Journal of Computer Applications ›› 2017, Vol. 37 ›› Issue (1): 299-304.DOI: 10.11772/j.issn.1001-9081.2017.01.0299

Previous Articles    

Dynamic estimation about service time of flight support based on Bayesian network

XING Zhiwei1, TANG Yunxiao1, LUO Qian2   

  1. 1. College of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300, China;
    2. Information Filiale, The Second Research Institute of Civil Aviation Administration of China, Chengdu Sichuan 610041, China
  • Received:2016-07-08 Revised:2016-09-05 Online:2017-01-10 Published:2017-01-09
  • Supported by:
    This work is partially supported by the Joint Funds of the National Natural Science Foundation of China and Civil Aviation Administration of China (U1533203), the Fundamental Research Funds for the Central Universities (3122014P003).

基于贝叶斯网络的航班保障服务时间动态估计

邢志伟1, 唐云霄1, 罗谦2   

  1. 1. 中国民航大学 电子信息与自动化学院, 天津 300300;
    2. 中国民航局第二研究所 信息技术分公司, 成都 610041
  • 通讯作者: 邢志伟
  • 作者简介:邢志伟(1970-),男,辽宁沈阳人,教授,博士,主要研究方向:民航装备与系统、机场交通信息与控制;唐云霄(1989-),男,安徽阜阳人,硕士研究生,主要研究方向:民航装备与系统、机场交通信息与控制;罗谦(1975-),男,四川绵阳人,高级工程师,博士,主要研究方向:机场运营管理、数据挖掘。
  • 基金资助:
    国家自然科学基金委员会-中国民用航空局联合研究基金资助项目(U1533203);中央高校基本科研业务费基金资助项目(3122014P003)。

Abstract: Concerning the problems of estimating the service time of airport flight support, and the particularity, complexity, and influence factors' uncertainty of flight support service process, an estimation model of flight support service time based on Bayesian Network (BN) was proposed. The knowledge of aviation experts and the machine learning of historical data were combined by the proposed model, and the incremental learning characteristic of BN was used to adjust the BN model dynamically, so as to make itself adapt to new conditions and constantly update the service time estimates of flight support. By using the data selected from a large domestic hub airport information system, the proposed BN model was trained via the Expectation Maximization (EM) algorithm to obtain the test results. The analysis of experimental results and model evaluation show that the proposed method can effectively estimate the service time of flight support and has higher accuracy. In addition, the sensitivity analysis demonstrates that the flight density during flight arrival time has the strongest influence on flight support service time.

Key words: flight support service, machine learning, Bayesian Network (BN), incremental learning, Expectation Maximization (EM), sensitivity analysis

摘要: 针对航班保障服务时间估计的问题,考虑到航班保障服务流程的特殊性、复杂性以及影响因素的不确定性,提出了一种基于贝叶斯网络(BN)的航班保障服务时间估计模型。该模型把航空领域的专家知识与历史数据的机器学习相结合,使用贝叶斯网络的增量学习特性动态地调整BN模型,使其适应新的变化,进而不断更新航班保障服务时间的估计值。使用国内某大型枢纽机场信息系统内提取的数据,通过期望最大化(EM)方法对模型进行训练,得到了测试结果。实验结果分析与模型评价表明,所提方法能有效估计航班保障服务时间且具有较高的准确度。敏感性分析表明,航班到达时段的航班密度对航班保障服务时间影响最强。

关键词: 航班保障服务, 机器学习, 贝叶斯网络, 增量学习, 期望最大化, 敏感性分析

CLC Number: