Abstract
Aiming at industries which sell the time limited products like airlines, hotels and car renters, if they don't sell their products by a specific time, these products would become worthless. In the contrast, the demand may increase due to seasonal or peak time. Under the fixed flight seats, hotel rooms, and cars, sellers must deny some customers when the capacities are all reserved. For consumers, they might not meet reservations on time because of some reasons, and they should have punishments. To increase the total revenue, sellers can use overbook customers to avoid loss from no-show customers. In this thesis, we construct an airline revenue management model concerning seat allocation. There are a series flights with different departure time at one day and sellers have to sell tickets by a specific time. To increase revenue, we accept the overbooked passengers and let them be transferred to next two flights. Through separating whole system into subsystems, we construct an overbooking and transference model and calculate terminal function, and then we decide the transference numbers. As the number of flights increase, the dimension of the proposed model increase sharply, so we proposed another simplified model for solving the problem. Furthermore, numerical examples are provided to show the solution procedure and the results are summarized as follows: (1) airlines can rise up revenue by transferring passengers to next flights; however, if the arrival rate is low, then the increasing amount of revenue is limited; (2) under the same arrival rate, let passengers be transferred to next two flights can have more revenue than passengers only be transferred to next flight; and (3) the overbooking limit can be estimated through marginal cost and price.