基于灵活性产品的航班网络舱位优化与控制
|
摘要
灵活性产品已在服务行业中应用多年，公司采用销售灵活性产品的方式，保留在合适的时候再确定产品细节的权利，以应对需求的不确定性。公司会在产品售出以后和提供服务以前确定产品的具体形式并告知顾客具体细节。实践表明，在资源容量一定且需求预测存在误差时，灵活性产品有助于提高公司的收益。然而人们对灵活性产品的相关问题进行系统的学术性研究的时间还不长，成果也非常有限。
本文的研究目的是将灵活性产品的概念引入到航空客运收益管理中，提出航空公司销售灵活性产品的流程和方法，重点研究基于灵活性产品的网络舱位优化控制。首先从收益管理系统的角度分析了航空公司进行网络舱位优化控制所需要的数据源，以数据源的数据结构为基础，提出了构建灵活性产品的方法和算法流程。以反应航班间时空衔接关系的航班网络结构为基础，构建了不考虑灵活性产品的航班网络舱位优化模型、基于灵活性产品的航班网络舱位优化模型、将灵活性产品分配到确定航班中的模型、分配灵活性产品前的模型和分配灵活性产品后的模型。
模拟十二种需求场景，采用遗传算法对算例进行求解，对比分析不销售灵活性产品和销售灵活性产品产生的期望收益的大小。通过数据研究表明，在预订期第一阶段，每种需求场景下销售灵活性产品的期望收益均高于不销售灵活性产品的期望收益；同时发现，在同一预测精度下，销售灵活性产品时期望收益的增加幅度表现出随着需求均值下降而增加的趋势。在销售灵活性产品和不销售灵活性产品的情况下，分别模拟15次订票情形，分析预订期第二阶段分配灵活性产品对期望收益的影响。数据研究表明，分配灵活性产品后的平均期望收益仍然高于不销售灵活性产品的平均期望收益，而且期望收益的增加幅度大于第一阶段。
因此得出结论，本文基于灵活性产品的航班网络舱位优化模型有助于航空公司提高整体航班网络的期望收益。
Flexible products have been popular for many years in practice in service industry.When selling a flexible product, a firm retains the right to specify some of its details later.The relevant point in time is after the sale, but often before the provision of the product orservice, depending on the customers’need to know the exact specification in advance.Practice shows that the resulting flexibility can help to increase revenues if capacity is fixedand the demand to come is difficult to forecast. But flexible products have only recentlygained attention in the academic literature on revenue management.
The purpose of this thesis is to introduce flexible products into airlines revenuemanagement, proposing the process and method of selling flexible products in airlines,mainly studying flight network inventory control incorporating flexible products. First, thedata source needed by network inventory control is analyzed in the viewpoint of revenuemanagement system. Based on the data structure, the way to structure flexible products isproposed. On the basis of the flight network which can reflect the relationship in time andspace between flights, five flight network inventory control models are set, including modelswith and without flexible products，model of allocating flexible products to fixed flights, andmodels used before and after allocating flexible products.
Through simulating twelve demand scenes and solving models with genetic algorithm,the difference of expected revenue produced by providing and not providing flexible productsis compared. The result shows that, during the first booking period, the expected revenuesproduced by providing flexible products are always more, and the increased amplitudes aregreater as demand reducing. Simulating fifteen booking situations, computing the expectedrevenues at the time of allocating flexible products, the result shows that the expectedrevenues produced by providing flexible products are still more, and the increased amplitudesare greater than the first period.
The conclusion indicates that the models building in this thesis are helpful of increasingthe expected revenue for airlines.
本文的研究目的是将灵活性产品的概念引入到航空客运收益管理中，提出航空公司销售灵活性产品的流程和方法，重点研究基于灵活性产品的网络舱位优化控制。首先从收益管理系统的角度分析了航空公司进行网络舱位优化控制所需要的数据源，以数据源的数据结构为基础，提出了构建灵活性产品的方法和算法流程。以反应航班间时空衔接关系的航班网络结构为基础，构建了不考虑灵活性产品的航班网络舱位优化模型、基于灵活性产品的航班网络舱位优化模型、将灵活性产品分配到确定航班中的模型、分配灵活性产品前的模型和分配灵活性产品后的模型。
模拟十二种需求场景，采用遗传算法对算例进行求解，对比分析不销售灵活性产品和销售灵活性产品产生的期望收益的大小。通过数据研究表明，在预订期第一阶段，每种需求场景下销售灵活性产品的期望收益均高于不销售灵活性产品的期望收益；同时发现，在同一预测精度下，销售灵活性产品时期望收益的增加幅度表现出随着需求均值下降而增加的趋势。在销售灵活性产品和不销售灵活性产品的情况下，分别模拟15次订票情形，分析预订期第二阶段分配灵活性产品对期望收益的影响。数据研究表明，分配灵活性产品后的平均期望收益仍然高于不销售灵活性产品的平均期望收益，而且期望收益的增加幅度大于第一阶段。
因此得出结论，本文基于灵活性产品的航班网络舱位优化模型有助于航空公司提高整体航班网络的期望收益。
Flexible products have been popular for many years in practice in service industry.When selling a flexible product, a firm retains the right to specify some of its details later.The relevant point in time is after the sale, but often before the provision of the product orservice, depending on the customers’need to know the exact specification in advance.Practice shows that the resulting flexibility can help to increase revenues if capacity is fixedand the demand to come is difficult to forecast. But flexible products have only recentlygained attention in the academic literature on revenue management.
The purpose of this thesis is to introduce flexible products into airlines revenuemanagement, proposing the process and method of selling flexible products in airlines,mainly studying flight network inventory control incorporating flexible products. First, thedata source needed by network inventory control is analyzed in the viewpoint of revenuemanagement system. Based on the data structure, the way to structure flexible products isproposed. On the basis of the flight network which can reflect the relationship in time andspace between flights, five flight network inventory control models are set, including modelswith and without flexible products，model of allocating flexible products to fixed flights, andmodels used before and after allocating flexible products.
Through simulating twelve demand scenes and solving models with genetic algorithm,the difference of expected revenue produced by providing and not providing flexible productsis compared. The result shows that, during the first booking period, the expected revenuesproduced by providing flexible products are always more, and the increased amplitudes aregreater as demand reducing. Simulating fifteen booking situations, computing the expectedrevenues at the time of allocating flexible products, the result shows that the expectedrevenues produced by providing flexible products are still more, and the increased amplitudesare greater than the first period.
The conclusion indicates that the models building in this thesis are helpful of increasingthe expected revenue for airlines.
引文
[1]杨威.机票产品迎合市场需求呈现多元化发展[J].中国民用航空, 2011,124(4):80-80
[2] Gallego G, Phillips R. Revenue Management of Flexible Products[J]. Manufacturing and ServiceOperations Management, 2004,6(4):321-337
[3]叶纪律,朱金福.不定期机票销售数量的优化控制方法研究[J].系统工程学报, 2009,5(24):626-630
[4]彭峥.我国航线网络的发展与展望[J].中国民用航空, 2009,105(9):51-53
[5]菲利普斯（Phillips, R. L.）著;陈旭,慕银平译.定价与收益优化[M].北京:中国财政经济出版社,2008:3-4
[6] Littlewood K. Forecasting and control of passenger bookings[J]. AGIFORS Symposium Proceedings,1972,V4(2):95–117
[7] Richter H. The differential revenue method to determine optimal seat allotments by fare type[J].AGIFORS Symposium Proceedings, 1982,22:339-362
[8] Belobaba P. P. Air Travel Demand and Airline Seat Inventory Management[D]. MIT: Cambridge, MA:1987
[9] Belobaba P. P. Application of a probabilistic decision model to airline seat inventory control[J].Operations Research, 1989,V37(2):183–197
[10] Wollmer R.D., An Airline Seat Management Model for a Single Leg Route when Lower Fare ClassesBook First [J].Operations Research, 1992,40:26-37
[11] Brumelle S.L., McGill J.I. Airline seat allocation with multiple nested fare classes[J]. OperationsReaserch, 1993,41:127-137
[12] Lee T.C., Hersh M. A model for dynamic airline seat inventory control with multiple seat bookings[J].Transportation Science, 1993,27:252-265
[13] Glover F, Glover R, Lorenzo J, et al. The Passenger Mix Problem in the ScheduledAirlines[J] .Interface, 1982,12:73-79
[14 ]Wollmer R.D. A hub-spoke seat management model[R]. Long Beach, CA: Mc Donnell DouglasCorporation,1986
[15] Smith B C, Penn C.W. Analysis of alternative origin-destination control strategies[J]. OperationsResearch, 1988,40:26-37
[16] Curry R.E. Optimal airlines seat allocation with fare classes nested by origins and destinations[J].Transportation Science, 1990,24(3):193-204
[17] Wong J.T., Koppelman F.S., Daskin M.S. Flexible assignment approach to itinerary seat allocation[J].Transportation Reasearch B, 1993,27:33-48
[18] Moller A, WEBER K. A new approach to O&D revenue management based on scenario trees[J].Journal of Revenue and Pricing Management, 2004,3(3):265-276
[19] COOPER W.L. Asymptotic behavior of an allocation policy for revenue management[J]. OperationsResearch, 2002,50(4):720-727
[20] Chen Lijian. Multistage stochastic programming models for airline revenue management[R].Evanston: Northwestern University, 2004
[21] Secomandi N. An analysis of the control-algorithm resolving issue in finite-horizon dynamic resourceallocation problems[R]. Pittsburgh: Carnegie Mellon University, 2004
[22] Williamson E.L. Airline network seat control:Methodologies and Revenue Impacts[D].MIT:Cambridge, 1992
[23] De Boer S.V., Sanne. The impact of dynamic capacity management on airline seat inventory control[J].Journal of Revenue & Pricing Management, 2004V2(4):315–330
[24]董裴君.航空公司网络舱位控制方法研究[D].南京:南京航空航天大学研究生院, 2007:24-37
[25] De Boer S.V., Freling R, Piersma N. Mathematical programming for network revenue managementrevisited[J]. European Journal of Operational Research, 2002,137(1):72–92
[26] Talluri KT, van Ryzin G.J. A randomized linear programming method for computing network bidprices. Transport Science, 1999,33(2):207–216
[27] Gallego G, Iyengar G, Phillips R, et al. Managing flexible products on a network.[R]. New York:University of Coumbia, 2004
[28] Petrick A, Klein R. Revenue Management of Flexible Product Presentation[A]. The 2005 Meeting ofthe GOR Group“Revenue management & Dynamic Pricing”[C], Frankfurt, Germany, 2005
[29] Muller B, Revenue Management with Flexible Products, Models and Methods for BroadcastingIndustry [M]. Berlin: Springer-Verlag, 2007
[30]周蓉,肖文莉.基于收益管理的高灵活产品两阶段销售模型研究[J].管理学报, 2009,l:57-61
[31] Petrick A, Claudius Steinhardt, Jochen G nsch , et al. Using flexible products to cope with demanduncertainty in revenue management[J]. OR Spectrum, 2009,34(1):251-242
[32]潘海莹.国内航空公司收益管理的应用和研究[D].厦门:厦门大学, 2007
[33]杨思梁,刘军.关于航空客运收益管理的一些基本概念[J].民航经济与技术, 1998,4:37-41
[34]衡红军,王红,王治宝.航空公司收益管理决策支持系统[J].计算机工程与应用,2002,28(9):213-214
[35]张永莉.航空公司收益管理理论、应用与创新研究[D].天津：天津大学, 2005
[36]梅虎,朱金福,汪侠.我国航空旅客选择行为结构特征及趋势[J].商业时代, 2007,13:95-97
[37]高强,朱金福,陈可嘉.航空收益管理中多航段舱位控制模型[J].交通运输工程学报,2005,4(5):82-85
[38]雷英杰. MATLAb遗传算法工具箱及应用[M].西安:西安电子科技大学出版社, 2005:147-148
[2] Gallego G, Phillips R. Revenue Management of Flexible Products[J]. Manufacturing and ServiceOperations Management, 2004,6(4):321-337
[3]叶纪律,朱金福.不定期机票销售数量的优化控制方法研究[J].系统工程学报, 2009,5(24):626-630
[4]彭峥.我国航线网络的发展与展望[J].中国民用航空, 2009,105(9):51-53
[5]菲利普斯（Phillips, R. L.）著;陈旭,慕银平译.定价与收益优化[M].北京:中国财政经济出版社,2008:3-4
[6] Littlewood K. Forecasting and control of passenger bookings[J]. AGIFORS Symposium Proceedings,1972,V4(2):95–117
[7] Richter H. The differential revenue method to determine optimal seat allotments by fare type[J].AGIFORS Symposium Proceedings, 1982,22:339-362
[8] Belobaba P. P. Air Travel Demand and Airline Seat Inventory Management[D]. MIT: Cambridge, MA:1987
[9] Belobaba P. P. Application of a probabilistic decision model to airline seat inventory control[J].Operations Research, 1989,V37(2):183–197
[10] Wollmer R.D., An Airline Seat Management Model for a Single Leg Route when Lower Fare ClassesBook First [J].Operations Research, 1992,40:26-37
[11] Brumelle S.L., McGill J.I. Airline seat allocation with multiple nested fare classes[J]. OperationsReaserch, 1993,41:127-137
[12] Lee T.C., Hersh M. A model for dynamic airline seat inventory control with multiple seat bookings[J].Transportation Science, 1993,27:252-265
[13] Glover F, Glover R, Lorenzo J, et al. The Passenger Mix Problem in the ScheduledAirlines[J] .Interface, 1982,12:73-79
[14 ]Wollmer R.D. A hub-spoke seat management model[R]. Long Beach, CA: Mc Donnell DouglasCorporation,1986
[15] Smith B C, Penn C.W. Analysis of alternative origin-destination control strategies[J]. OperationsResearch, 1988,40:26-37
[16] Curry R.E. Optimal airlines seat allocation with fare classes nested by origins and destinations[J].Transportation Science, 1990,24(3):193-204
[17] Wong J.T., Koppelman F.S., Daskin M.S. Flexible assignment approach to itinerary seat allocation[J].Transportation Reasearch B, 1993,27:33-48
[18] Moller A, WEBER K. A new approach to O&D revenue management based on scenario trees[J].Journal of Revenue and Pricing Management, 2004,3(3):265-276
[19] COOPER W.L. Asymptotic behavior of an allocation policy for revenue management[J]. OperationsResearch, 2002,50(4):720-727
[20] Chen Lijian. Multistage stochastic programming models for airline revenue management[R].Evanston: Northwestern University, 2004
[21] Secomandi N. An analysis of the control-algorithm resolving issue in finite-horizon dynamic resourceallocation problems[R]. Pittsburgh: Carnegie Mellon University, 2004
[22] Williamson E.L. Airline network seat control:Methodologies and Revenue Impacts[D].MIT:Cambridge, 1992
[23] De Boer S.V., Sanne. The impact of dynamic capacity management on airline seat inventory control[J].Journal of Revenue & Pricing Management, 2004V2(4):315–330
[24]董裴君.航空公司网络舱位控制方法研究[D].南京:南京航空航天大学研究生院, 2007:24-37
[25] De Boer S.V., Freling R, Piersma N. Mathematical programming for network revenue managementrevisited[J]. European Journal of Operational Research, 2002,137(1):72–92
[26] Talluri KT, van Ryzin G.J. A randomized linear programming method for computing network bidprices. Transport Science, 1999,33(2):207–216
[27] Gallego G, Iyengar G, Phillips R, et al. Managing flexible products on a network.[R]. New York:University of Coumbia, 2004
[28] Petrick A, Klein R. Revenue Management of Flexible Product Presentation[A]. The 2005 Meeting ofthe GOR Group“Revenue management & Dynamic Pricing”[C], Frankfurt, Germany, 2005
[29] Muller B, Revenue Management with Flexible Products, Models and Methods for BroadcastingIndustry [M]. Berlin: Springer-Verlag, 2007
[30]周蓉,肖文莉.基于收益管理的高灵活产品两阶段销售模型研究[J].管理学报, 2009,l:57-61
[31] Petrick A, Claudius Steinhardt, Jochen G nsch , et al. Using flexible products to cope with demanduncertainty in revenue management[J]. OR Spectrum, 2009,34(1):251-242
[32]潘海莹.国内航空公司收益管理的应用和研究[D].厦门:厦门大学, 2007
[33]杨思梁,刘军.关于航空客运收益管理的一些基本概念[J].民航经济与技术, 1998,4:37-41
[34]衡红军,王红,王治宝.航空公司收益管理决策支持系统[J].计算机工程与应用,2002,28(9):213-214
[35]张永莉.航空公司收益管理理论、应用与创新研究[D].天津：天津大学, 2005
[36]梅虎,朱金福,汪侠.我国航空旅客选择行为结构特征及趋势[J].商业时代, 2007,13:95-97
[37]高强,朱金福,陈可嘉.航空收益管理中多航段舱位控制模型[J].交通运输工程学报,2005,4(5):82-85
[38]雷英杰. MATLAb遗传算法工具箱及应用[M].西安:西安电子科技大学出版社, 2005:147-148