Dynamic revenue management of airline and production systems

Abstract

This dissertation investigates revenue management of airline seat inventory control problems and stock rationing of production/inventory problem. In a booking horizon, in order to maximize the total revenue through selling a finite number of seats of a plane, management must repeatedly decide whether to accept a booking request or reserve the seat for later higher-fare bookings. Correspondingly, for a production/inventory system, in terms of minimizing the penalty cost, it is often important to satisfy a class of demand with higher priority than a lower one. Therefore, inventories could be rationed accordingly. In this dissertation, Markov Decision Processes (MDP) has been used to formulate the problems of flight seat inventory control and rationing of the make-to-stock production/inventory system. Optimal policies for these problems have been completely or partially characterized. For each of the problems, numerical studies verify our analytical results.