【作者】 姜涛；

【导师】 朱金福；

【作者基本信息】 南京航空航天大学 ， 管理科学与工程， 2007， 博士

【摘要】 航线网络是航空公司的生存之本,科学地构建航线网络是航空公司求得发展的重要手段。航空公司的其它工作如航班计划、运行控制、收益管理等都是在航线网络的基础之上,围绕着已经布局好的航线网络进行的,因此航线网络结构的合理与否对于航空公司的效益将产生深远的影响。随着航空运输业规模的不断扩大,能够充分体现规模经济的中枢辐射航线网络受到了越来越多地关注。以往对于中枢辐射航线网络的优化设计采用的都是确定型的方法,但是优化设计涉及的航空运输需求和成本往往具有不确定性,因此针对中枢辐射航线网络鲁棒优化的有关问题展开研究。通过对点对点航线网络结构、严格的以及非严格的中枢辐射航线网络结构的定量化比较研究,得到了航线网络结构与市场规模的关系。当航空运输发展到一定规模时,中枢辐射航线网络结构能够充分发挥出自身的优势,印证了中枢辐射航线网络体现规模经济的特点。当中转衔接的两条航线的夹角小于一特定的临界值时,将中转运输的方式改为直达运输能够进一步增加航空公司的利润。对不确定情形下枢纽机场选择问题采用偏差鲁棒优化方法进行了研究,提出了新的求解算法,将已有算法的复杂性由降低到.针对航空运输需求和成本具有不确定性的特点,在需求和成本各种可能取值的概率分布未知的情形下,建立了多种中枢辐射航线网络的鲁棒优化模型。首先基于枢纽机场的选择,建立了严格的和非严格的中枢辐射航线网络鲁棒优化枢纽机场选择模型——S-HS-R-C p -Hub和NS-HS-R-C p -Hub;然后将开辟航线的成本考虑在内,在枢纽机场选定的情况下,建立了严格的中枢辐射航线网络鲁棒优化航线选择模型——S-HS-R-C p -Airline;最后基于枢纽边的选择,建立了严格的中枢辐射航线网络鲁棒优化枢纽边选择模型——S-HS-R-C q -Hub-Arc。提出了上述中枢辐射航线网络鲁棒优化模型的求解算法。在枢纽个数较少的情况下,改进了基于最短路求解确定型模型S-HS-C p -Hub的算法用于S-HS-R-C p -Hub和NS-HS-R-C p -Hub的求解;在枢纽个数较多的情况下,基于禁忌算法,给出了S-HS-R-C p -Hub的求解算法。针对S-HS-R-C p -Airline,将Benders Decomposition算法进行改进,给出了模型具体求解的算法。将求解确定型模型S-HS-C q -Hub-Arc的枚举法加以改进,给出了求解S-HS-R-C q -Hub-Arc的算法。对于上述建立的中枢辐射航线网络鲁棒优化模型,在我国十五城市的基础上,分别进行了构建中枢辐射航线网络的实例分析,并验证了模型求解算法的有效性。更多还原

【Abstract】 The airline network is the foundation of the airline company, and the important means of the airline company which try for development is to design the airline netwok by scientific method. Other works of the airline company such as flight schedule, operation control, revenue management are based on the airline network. Whether the airline network is reasonable or not has far-reaching effect on the benefit of the airline company. The hub-and-spoke airline network can achieve economies of scale. This network has attracted more and more attention when the scale of air transportation is expansion. The previous research on the hub-and-spoke airline network design was for the certain scenario. But in practice the demand and cost of air transportation are uncertain. So this article studies the hub-and-spoke airline network robust optimization design for the uncertain scenario.Through the quantificational comparison on the point-to-point and the strict and the nonstrict hub-and-spoke airline networks, we obtain the connection of the airline network configuration and the scale of the market. The hub-and-spoke airline network can exert its predominance when the air transportation reaches some scale. The characteristic of the hub-and-spoke airline network is confirmed. The transfer transportation mode ought to be substituted by the nonstop transportation mode when the included angle of the two airlines in the transfer transportation mode is smaller than a specific value. The benefit of the airline company can increase.This article studies the problem of the hub airport selection by robust deviation optimization. We propose new algorithm to solve the problem. The complexity of the existing algorithm is reduced from min. In the case of the demand and cost with the uncertainty, this article establishes several robust optimization models of the hub-and-spoke airline network. The distribution of probability is unknown. First of all we establish the hub airport selection models of the strict and nonstrict hub-and-spoke airline networks. Then taking into account the cost of opening airline, we establish the airline selection model of the strict hub-and-spoke airline network when hub airports have been selected in advance. In the end based on selecting hub arc, the hub arc selection model of the strict hub-and-spoke airline network is established.This article proposes the algorithms to solve the models which have been established before. When the number of hub airports is small, the algorithm which can solve the certain model is improved to solve S-HS-R-C p -Hub and NS-HS-R-C p -Hub. The improved algorithm is based on the shortest path algorithm. When the number of hub airports is large, the algorithm is proposed to solve S-HS-R-C p -Hub and the algorithm is based on the tabu search algorithm. We improve the Benders decomposition algorithm to solve S-HS-R-C p -Airline. We improve the enumeration algorithm to solve S-HS-R-C q -Hub-Arc. The enumeration algorithm can solve the certain model which is based on selecting hub arc. We separately use the models which have been established before to construct the hub-and-spoke airline network on fifteen cities of China. These instances validate the algorithms which are proposed to solve the robust optimization models.更多还原