Modeling Airline Overbooking
Diana Iercosan
Faculty Advisor: T. Lengyel

     The airline industry uses mathematical models in order to increase its profits. Airlines face two major constraints: fixed capacity and the fact that seats on a certain flight represent perishable-assets, as they are not available to customers after the flight’s departure.
     Airlines can maximize revenue by adopting the discounted seat strategy with a single overbooking factor at cabin level. This policy is based on the possibility to segment the market on price sensitivity, which allows companies to charge customers, who book in advance, lower fares. The authorization levels for each fare class are computed in order to obtain an optimal allocation that minimizes the lost opportunities to sell seats at higher-fares and produce a full aircraft. The Expected Marginal Seat Revenue and improved methods are used to decide the booking limits in a deterministic as well as dynamic form.
     Additionally, estimates of the optimal overbooking factor are determined in order to maximize the expected number of passengers on a flight and as a result the probability of flying at full capacity. Overbooking takes into account the costs incurred from having empty seats, but also the penalty paid to extra passengers, who are denied boarding. Simplistically, a linear bumping compensation has been used to produce an appropriate booking limit. A more realistic approach, which deals with chain effects by using a nonlinear bumping compensation.

Support provided by: Ford Research Fellowship