バイアップがある多クラス静的レベニューマネジメントモデル (不確実環境下における意思決定数理の新展開)

Abstract

In the classical Littlewood's two-period model of static revenue management for airline seat reservation as well as its extensions to multi-period models by other researchers, it is assumed that the demands in each period are independent and that customers whose request for reservation are once rejected disappear immediately. This assumption does not reflect a common behavior of customers in practice such that they tend to seek booking again at higher fare. The theoretical treatment of such buy-up behavior is not simple because of the resultant mutual dependence of the demands over several periods. Some literature heuristically (incorrectly) treats the two-class model with customers' buy-up behavior. In a series of papers, we study the optimal booking limits in the two-and more-period static revenue management models with customers' buy-up behavior. In the RIMS Proceedings of Symposium 2158 (paper 19) [9], we showed an analysis for the three-period model with numerical results. In the present paper, we proceed to study the four-period model (without numerical results) based on a newly found technique to derive the expressions for the first partial derivatives of the expected number of reserved seats with respect to the booking limits in each period.