选择子网络的飞机恢复问题模型改进研究

Research on Model Improvement for Aircraft Recovery Based on Sub-Network Selection

摘要: 本文分析选择子网络的飞机恢复方法的缺陷，即非目标不正常航班可能导致额外的航班取消。针对该缺陷，本文对经典飞机恢复模型进行改进，增加虚拟航班的概念以减小非目标不正常航班的影响。最后，本文使用机器学习选择子网络并滚动求解，对改进模型的效果进行测试。实验结果对经典模型和改进模型的结果进行对比，显示出改进模型在减少航班取消数量上的显著优势。

Abstract: This paper analyzes the deficiency of aircraft recovery methods that employ sub-network selection, namely, the potential for non-target irregular flights to result in additional flight cancellations. To address this deficiency, this paper proposes an improvement to the classic aircraft recovery model by introducing the concept of virtual flights to mitigate the impact of non-target irregular flights. Finally, this paper utilizes machine learning to select sub-networks and employs a rolling approach to recover the whole network, thereby testing the effectiveness of the improved model. Experimental results comparing the outcomes of the classic model and the improved model demonstrate the significant advantage of the latter in reducing the number of flight cancellations.