摘要
影响航班到达延误的因素众多,涉及节假日、时段、天气、航空器故障等。本文提出一种基于神经网络多输入–单输出的航班到达延误预测方法,并利用遗传算法优化神经网络的结构与参数,从而进一步提升预测精度。最后,以2015年美国亚特兰大机场为例,给出了预测结果,并与传统神经网络进行性能对比,从而验证了本研究的有效性。研究表明:研究表明,经过遗传算法优化后的神经网络的误差比仅为单独使用神经网络误差的58%。
There are many factors affecting flight arrival delay, including holidays, time periods, weather, aircraft failures, etc. In this paper, a multi input single output prediction method of flight arrival delay based on neural network is proposed, and genetic algorithm is used to optimize the structure and parameters of neural network, so as to further improve the prediction accuracy. Finally, taking Atlanta Airport in the United States in 2015 as an example, the prediction results are given and compared with the performance of traditional neural networks to verify the effectiveness of this study. The research shows that the error ratio of neural network optimized by genetic algorithm is only 58% of the error of neural network alone.
出处
《应用数学进展》
2024年第7期3481-3487,共7页
Advances in Applied Mathematics
