摘要
对航班地面保障事件执行日志展开分析以及对国内外航班延误报告进行研究,针对航班地面保障事件执行效率低下的影响因素发现问题,明确关联与因果的统一定义,给出量化影响因素的时间差参考量,提出一种面向实际保障过程的基于增剪边规则的强化因果发现算法,通过最大化奖励分数得到最优的影响因素与目标事件因果图。实验结果表明,提出方法能有效得到真实因果图,在FDR、TPR、SHD与生成的因果图的合理性上超过其它对比方法。
Through the analysis of the execution log of flight ground support events and the study of flight delay reports at home and abroad,aiming at the problem of finding the influencing factors of the low execution efficiency of flight ground support events,the unified definition of causality was defined.The time difference consideration of quantifying the influencing factors was given.A reinforcement causal discovery algorithm based on the trimming edge rule was proposed,which was oriented to the actual support process.By maximizing the reward score,the optimal causality diagram of the influencing factors and the target events was obtained.Experimental results show that the proposed method can effectively obtain the real cause and effect diagram,and the rationality of FDR,TPR,SHD and the generated causal graph is better than that of other compared methods.
作者
邢志伟
李正锋
罗谦
夏欢
张涛
XING Zhi-wei;LI Zheng-feng;LUO Qian;XIA Huan;ZHANG Tao(Electronic Information and Automation Institute,Civil Aviation University of China,Tianjin 300300,China;Engineering Technology Research Center,The Second Institute of Civil Aviation Administration of China,Chengdu 610041,China)
出处
《计算机工程与设计》
北大核心
2024年第4期1126-1133,共8页
Computer Engineering and Design
基金
国家自然科学基金项目(62106269)
国家重点研发计划基金项目(2018YFB1601200)。
关键词
航班地面保障
航班离港延误
因果关系
因果发现
强化学习
因果图构建
航班保障效率
flight ground support
flight departure delay
causality
causal discovery
reinforcement learning
construction of causal graph
flight ground support efficiency
