突发事件下大规模空中交通流量管理的组合优化模型

为了解决危险天气和军航活动对管制运行的影响，使用动态网络流方法对突发事件下短期空中交通流量调度问题展开研究。首先，结合中国民航管制的特点，分析了突发事件对流量管理的影响。同时，根据航路航线网络及其高度层的特点，给出了网络和高度层的数学描述，并根据机型将航空器分成大、中和小3种交通流，介绍了使用多品种流描述3种机型的必要性；其次，根据网络拥挤程度随时间和流量变化的特点、危险天气随机变化的特点、空中等待和地面等待费用差异的特点构建了3个优化目标，考虑机场容量、扇区容量、航班连续性和扇区连续性约束建立了多目标优化模型；再次，针对军航活动时需要协调空域的特点，多品种流模型求解时间复杂度高，不能适应短期流量管理的缺陷，改进了逐步宽容约束法，设计了一种阶段式求解的近似算法；最后，以西南空管局管制的空域为例，利用实际流量数据，设计了3个场景进行仿真。结果表明，所提模型和算法能有效求解突发事件下的短期流量调度问题，算法效率比起传统算法在大流量下更具优势。

Combined optimization method for large-scale air traffic flow management under emergencies

To address the disturbance like extreme weather conditions and air traffic control for military activities, a dynamic network flow method is proposed. Firstly, combined with the characteristics of civil aviation control in China, the impact of emergencies on traffic management is analyzed. Also, this paper uses the graph theory to establish a mathematical model for en-route network and flight level. Then the aircraft is divided into large, medium, and small traffic flows, introducing the necessity of using the multi-commodity flow method. Secondly, considering the condition of network congestion varies with time and traffic flow, the random changes of dangerous weather, and the characteristics of air holding and ground holding cost, three optimization objectives are defined, then the multi-objective mathematical model is constructed with the constraint of airport capacity, sector capacity, flight continuity, and sector continuity. Thirdly, to deal with the demand conflict with military activities and the defects of high time complexity in the multi-commodity flow model that can not meet the demand of short-term flow management, the progressive tolerance constraint method is improved, and a phased solution approximation algorithm is designed. Finally, a part of the airspace of the Southwest Air Traffic Control Administration is used for simulation in three scenarios. Simulation results show that the proposed model and algorithm can effectively solve the short-term traffic management problem under emergencies. The algorithm is more efficient than the traditional algorithm under large traffic flows.