基于copula函数的飞行操作风险分析  

Risk Analysis of Operations in Flight Based on Copula

作　　者：邵雪焱[1] 池宏[1] 高敏刚[1] 

机构地区：[1]中国科学院科技政策与管理科学研究所,北京100190

出　　处：《数理统计与管理》2012年第5期771-779,共9页Journal of Applied Statistics and Management

摘　　要：QAR超限是航空运输的一大类安全隐患,本文综合运用Copula函数和Logistic回归建立飞行操作的风险分析模型。结合模型特点设计进化算法,实现对飞行操作状态空间的搜索,通过模型寻优过程找到QAR超限的高风险操作。对航空公司一类重点监控QAR超限的实证研究,得到易导致该类超限的高风险操作,为飞行员针对性训练提供了量化依据。QAR exceedances are hidden danger to air transportation. This paper combines copula with logistic regression to establish risk analysis model of operations in flight. According to optimal objectives of the model, we design evolutionary algorithm to explore operations state space, in order to find the sub-space with higher exceedance frequency. Through empirical study of a specific QAR exceedance, we find some high risk operations associated with it. These results give a basis for pilots＇ special training.

关 键 词：QAR超限 飞行操作 风险 COPULA 进化计算 

分 类 号：O212[理学—概率论与数理统计] O221[理学—数学]