Nowadays, to make additional profits, revenue management system plays an important role for airline. This study presents a framework of airline revenue management decision support system. The system includes 3 sub-systems, interface, model management and database management which integrate OLTP and OLAP techniques. Several algorithms for flight demand forecasting, overbooking and seat inventory control are designed. Cluster algorithm is employed to decrease online run-time for flight demand forecasting without considering season, week and event attributes of flights. Overbooking algorithm is built in terms of binary distribution of passenger arrival. Multi-segments seat inventory control algorithm based-on expected marginal seat revenue (EMSR) provides a new way to control 3 legs flight. The system has been applied successfully in Xiamen Airline of China. The results have shown a great efficiency of this system.
[1]
Jeffrey I. McGill,et al.
Revenue Management: Research Overview and Prospects
,
1999,
Transp. Sci..
[2]
Richard E. Chatwin,et al.
Multiperiod Airline Overbooking with a Single Fare Class
,
1998,
Oper. Res..
[3]
Jeffrey S. Zickus.
Forecasting for airline network revenue management : revenue and competitive impacts
,
1998
.
[4]
Richard E. Chatwin,et al.
Continuous-Time Airline Overbooking with Time-Dependent Fares and Refunds
,
1999,
Transp. Sci..
[5]
Richard Robert Wickham.
Evaluation of forecasting techniques for short-term demand of air transportation
,
1995
.
[6]
Peter P. Belobaba,et al.
Survey Paper - Airline Yield Management An Overview of Seat Inventory Control
,
1987,
Transp. Sci..