Overbooking models based on personal data analysis

Overbooking policy, which is part of revenue management, is applied in many sectors today where unused capacity cannot be stored. The purpose of this application is to maximize the amount of earnings while reducing spoilage costs. Especially in the service sectors where the competition is high, such as the airline industry, hotel and restaurant management, it is important that the overbooking limit brings the company closer to the maximum gain. There are different approaches in the literature to determine the optimal value of the overbooking limit. In the classical approach, the show probability of each customer is assumed to be the same. In this study, unlike the classical approach, the show probability of each reservation is estimated on a personal basis. Personal probabilities needed for the suggested approach are obtained from historical booking information using forecasting algorithms. The fact that each customer has different show probability makes it difficult to calculate the probabilities of different number of bookings in the classical way. For this reason, Monte Carlo simulation is used to calculate the optimal booking limit when show probability of each customer is taken differently on a personal basis. The optimal reservation limit and expected revenue are calculated with the total show probabilities obtained as a result of the simulation. As a result of the study, it is seen that with the proposed dynamic models higher gains can be achieved compared to the classical approach. Accuracy and conformity rates of the forecasting algorithms are analyzed for performance measurement and the changes in the results of the proposed models with respect to the changes in parameter values are researched through a detailed sensitivity analysis.