基于排队网络模型的机场群航班时刻资源优化配置关键技术研究

The strategy of air-ground integration and collaborative optimization will greatly alleviate the contradiction between the rapid development of civil aviation transport industry and limited airport and airspace resources. In addition to the.airspace optimization measures, how to optimize the allocation of the.multi-airport slot, effectively alleviate the airport congestion and flight delay.is also one of the key technology of the collaborative optimization flow management. This project makes clustering analysis of air traffic flow pattern with self-organizing mapping neural network. The air traffic queuing network model based on Coxian distribution is put forward. The quantitative analysis of the correlation between the delay and the air traffic flow control measures of multi-airport is carried out by using the theory of random matrix. With the improved clustering and quantile regression algorithm, multi-airport "coupling" capacity is evaluated. The capacity evaluation method of the balance between the delay and the airport revenue is put forward. Based on the above factors, the multi-objective optimization of multi-airport slot model is established. Based on cooperative co-evolutionary genetic algorithm and gradient projection method for solving constrained optimization problems is proposed to get the best target weights of every airport delay weight value of multi-objective optimization problems. The project archives the optimal allocation of multi-airport slot resources under premise of the delay target level. The implementation of the project will provide a more advanced technology for the flight operation simulation and flight schedule optimization of multi-airport.

空地一体化协同优化战略的提出将极大地缓解目前快速发展的民航运输业与有限的机场、空域资源之间的矛盾。除空域优化措施之外，如何对机场群航班时刻资源优化配置，有效缓解机场拥堵及航班延误也是协同优化流量管理的关键技术之一。本项目利用自组织特征映射神经网络对交通流模式进行聚类分析，提出基于Coxian分布的机场群地空排队网络模型；并利用随机矩阵理论对机场群延误与流控措施间的相关性进行量化分析；利用改进聚类、分位数回归算法对机场群“耦合”容量进行评估分析；提出延误损失与机场收益平衡的容量评估方法。在对上述多因素分析的基础上，构建机场群航班时刻多目标优化模型，提出利用合作协同遗传算法和梯度投影法求解约束优化问题以获得机场延误权重值未知的多目标优化问题的最佳目标权重，并在满足延误目标水平前提下实现机场群航班时刻资源优化配置。项目的实施将为我国机场群航班运行仿真及航班时刻编排提供更加完善的技术手段。

空地一体化协同优化战略的提出将极大地缓解目前快速发展的民航运输业与有限的机场、空域资源之间的矛盾。除空域优化措施之外，如何对机场群航班时刻资源优化配置，有效缓解机场拥堵及航班延误也是协同优化流量管理的关键技术之一。本项目开展了机场群地空排队网络模型研究；利用改进聚类、分位数回归算法对机场群“耦合”容量进行评估分析；提出延误损失与机场收益平衡的容量评估方法。在对上述多因素分析的基础上，构建机场群航班时刻多目标优化模型，提出利用合作协同遗传算法和梯度投影法求解约束优化问题以获得机场延误权重值未知的多目标优化问题的最佳目标权重，并在满足延误目标水平前提下实现机场群航班时刻资源优化配置。项目组共发表标注该基金课题号的论文10篇文章已发表。其中SCI论文发表2篇，EI论文1篇，第二标注2篇，第三标注1篇。北大中文核心论文5篇，均为第一标注；其他中文论文1篇，第一标注一篇，超额完成论文要求。本课题结合数据挖掘理论、空中交通管理理论、运筹学理论等，综合应用时空聚类技术、排队网络模型分析技术、多目标规划方法、智能优化算法、计算机仿真等理论方法实现机场群地空排队网络模型和航班时刻优化算法。这些理论方法在机场群航班时刻优化领域的到了验证与应用，为今后学者的研究工作提供了典型案例与经验，这些成果会对未来的研究工作起到抛砖引玉的作用。项目的实施将为我国机场群航班运行仿真及航班时刻编排提供更加完善的技术手段。