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 overshoot is a major category of air transport security risks, the paper comprehensive use of Copula function and Logistic regression to establish risk analysis model of flight operations. Combining with the characteristics of the model, an evolutionary algorithm is designed to search for the flight operation state space, and a high-risk QAR overrun operation can be found through the model optimization process. Empirical research on a QAR overrun, which is one of the most important airline companies in the country, has obtained the high-risk operation that easily leads to such overrun and provides quantitative basis for pilots’ targeted training.