空中交通管制安全是确保飞行安全的重要因素,为提高空中交通管制系统的安全性,需要对空中交通管制员人为差错风险进行评估.首先,基于SHEL模型构建了管制员人为差错评估指标体系,之后针对BP神经网络算法容易陷入局部极小值的缺点,通过引入遗传算法优化神经网络的权值和阈值,建立了一种混合算法;经实例评估表明,GA-BP算法与传统的BP神经网络模型相比具有更高的预测精度,具有较好的适用性和可行性,为加强航空业安全管理体系建设提供了一定的参考依据.
The safety of air traffic control is an important factor to ensure flight safety,in order to improve the safety of air traffic control system,it is necessary to assess the risk of human error of air traffic control system.A SHEL model was constructed based on the controller's error evaluation index system.Because the BP neural network algorithm is easy to fall into local minimum,the paper introduces genetic algorithm to optimize neural network weights and threshold to establish a hybrid algorithm.Evaluation showed that comparing with the traditional prediction BP neural network model with GA-BP algorithm is more accurate,feasible and applicable.It provides certain reference for strengthening the construction of aviation safety management system.