航空公司微观机队规划方法研究

发布时间：2018-06-22 19:19

本文选题：机队规划 + 机型分配　；参考：《南京航空航天大学》2016年博士论文

【摘要】：机队是航空公司进行运输生产的载体,机队规划是航空公司重要战略性规划内容之一,是决定航空公司未来生死存亡的关键性战略决策,同时也是航空公司其它管理决策,例如航班计划、飞机排班、收益管理等决策工作的前提,机队规划的好坏将从根本上影响航空公司未来运输生产的效益,因此,在满足航空公司未来拟运营生产环境(航线网络结构、航班计划、需求与票价水平等)的基础上,研究飞机机队的规模与结构特征,对于航空公司优化运输产品结构,提高运输生产效益都具有重要的现实意义。针对宏观机队规划法无法准确反映出航班机型分配技术经济性能特点的缺陷,首先,本文针对单基地线形航线网络运营模式构建了基于“航班环”时序网络的机队规划数学模型,并设计了基于有向路径随机分解的模拟退火算法。算例结果表明,算法具有求解精度高、且速度快的特点。然后,将“航班环”时序网络推广至“航班串”时空网络,构建基于时空网络的机队规划随机优化数学模型,并设计了基于情景汇聚的两阶段算法。算例结果表明,在不确定需求下算法能够体现出更好的效益优势。最后,将旅客收益网络优化模型合并入基于航班时空网络的机队规划随机优化模型之中,利用上述两阶段算法进行了求解,算例结果表明该类收益增强型机队规划模型与算法能够更好的适用于枢纽轮辐式航线网络运营模式。针对基于航班机型分配微观机队规划法无法准确模拟航空公司未来航班计划的缺陷,本文研究了基于航线运力分配的微观机队规划法。首先,在综合考虑航线最大飞行频次限制、机队可用飞行时间限制,以及航线供需平衡限制等因素的基础上,构建基于随机需求下以机队运营利润最大化为目标的不确定机队规划数学模型,通过产生期望收益非线性函数的近似线性函数后,将旅客期望收益网络优化模型从不确定机队规划数学模型中分离,并设计了两个子模型迭代求解的算法。采用航空公司实际数据对模型与算法进行测试,通过与遗传算法计算结果进行比较,说明该分解算法的优势。在此基础上针对枢纽网络运营模式,将旅客收益网络优化模型并入该机队规划数学模型之中,利用上述分解算法再次进行求解,并采用算例验证该类机队规划模型在枢纽轮辐式航线网络运营模式下进行机队规划决策的优势。针对现有基于航线运力分配的微观机队规划法无法准确反映旅客需求波动动态性以及微观机队规划环境非垄断性的问题,本文提出了基于离散时间航线运力分配的机队规划模型,并设计了求解该模型的拉格朗日松弛解法。算例结果表明,该方法比基于航班机型分配机队规划法更加稳定、比基于航线运力分配机队规划法更能反映出旅客需求随时间动态波动过程中所需机队运力的差异性;另一方面,本文在现有基于航线运力分配的机队规划模型中考虑了竞争航空公司航线运力分配方案对于本航空公司的影响,构建了多航空公司竞争型机队规划数学模型,并基于均衡最优理论设计了启发式算法进行求解。算例结果表明,算法能够获取机队规划均衡解,且通过Monte Carlo模拟法验证了该竞争型微观机队规划法的优势。
[Abstract]:The fleet is the carrier for the airline to carry out the transportation, and the fleet planning is one of the important strategic planning contents of the airline. It is the key strategic decision to determine the future life and death of the airline. It is also the prerequisite for other decision making of the airline management, such as flight plans, aircraft scheduling, and revenue management. The good or bad will fundamentally affect the efficiency of the future transportation and production of the airline. Therefore, on the basis of meeting the future operating production environment of the airline (airline network structure, flight plan, demand and fare level, etc.), the scale and structure characteristics of aircraft fleet are studied, the structure of transport products is optimized for the airlines and the transport students are improved. The production benefit has important practical significance. In view of the defect that the macro fleet planning method can not accurately reflect the technical and economic performance characteristics of the flight model allocation, first, this paper constructs a fleet planning mathematical model based on the "flight loop" time series network for the single base line network operation mode, and designs a directed path based on the directed path. The simulated annealing algorithm of random decomposition shows that the algorithm has the characteristics of high accuracy and fast speed. Then, the "flight loop" time series network is popularized to the "flight string" space-time network, and the stochastic optimization mathematical model of fleet planning based on space-time network is constructed, and the two stage algorithm based on scene convergence is designed. The results show that the algorithm can show better benefit advantage under the uncertain demand. Finally, the passenger revenue network optimization model is incorporated into the stochastic optimization model of fleet planning based on flight space-time network, and the two stage algorithm is used to solve the problem. The results show that the model and algorithm of this type of revenue enhancement fleet planning and algorithm can be obtained. In this paper, the micro fleet planning method based on the distribution of airline force is studied in this paper. Firstly, the fleet can be used for the maximum flight frequency limitation, and the fleet is available. On the basis of the limitation of flight time and the balance of supply and demand, a mathematical model of uncertain fleet planning based on the maximum profit of fleet operation is built on the basis of random demand. By producing the approximate linear function of the expected return nonlinear function, the model of the passenger expected return network is never determined. The algorithm is separated from the mathematical model and the algorithm of two sub models is designed. The model and the algorithm are tested with the actual data of the airline. By comparing with the results of the genetic algorithm, the advantages of the decomposition algorithm are illustrated. The team planning mathematical model, using the above decomposition algorithm to solve the problem again, and using an example to verify the advantage of this kind of fleet planning model in the hub spoke airline network operation mode. In view of the existing micro machine team planning method based on the distribution of route capacity, it can not accurately reflect the fluctuation of passenger demand. In this paper, a fleet planning model based on the distribution of discrete time route is proposed, and a Lagrange relaxation method for solving the model is designed. The results show that the method is more stable than the flight model distribution unit planning method, which is based on the distribution of route capacity. The fleet planning method can reflect the difference of the fleet capacity required for the dynamic fluctuation of the passenger demand with time. On the other hand, this paper takes into account the influence of the competitive airline route allocation scheme to the airline in the existing fleet planning model based on the distribution of airline power, and constructs a multi airline competitive machine. The team plans the mathematical model and designs a heuristic algorithm based on the equilibrium optimal theory. The results show that the algorithm can obtain the equilibrium solution of the fleet planning, and the advantage of the competitive micro fleet planning method is verified by the Monte Carlo simulation method.
【学位授予单位】：南京航空航天大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：V352;F560.6

【相似文献】