Overbooking is one of the most important techniques in revenue management which sells goods or services in excess of the available capacity because there is a possibility that some customers might cancel or not show up. In air cargo overbooking problem, overbooking model is more complex as compared to others because of the two-dimensional characteristic, i.e., weight and volume of the booking requests. This study presents two-dimensional air cargo overbooking models to find the optimal overbooking level with an objective of minimizing the total cost. Booking request and show-up rate are random variables with known distributions. This research proposes a new way of modeling the total cost that has not yet been presented before. There are four main parts in this research. First, the full mathematical model is formulated. Second, computational experiments are used to observe the impacts of important factors on the optimal overbooking level. Third, the other two models are proposed to reduce the complexity of the full model: 1) a regression model with interactions and 2) a naive method model. Last, the results of these three models are compared to identify which situation is suitable for each. The results showed that the optimal overbooking level obtained from the simplified regression model is close to the optimal overbooking level obtained from the full model (R-sq(adj) = 98.3%). The naive method is the simplest method presented in this study; however, it does not always give the appropriate optimal overbooking level. For example, the naive method predicts the optimal overbooking level close to the real optimal overbooking level when the ratio between the offloading and spoilage costs are low because it does not consider this factor.

Thesis (M.Eng.)--Chulalongkorn University, 2014