Solving the Airline Overbooking Problem Using Fuzzy Optimisation Techniques

It is essential to cut costs and increase revenue wherever is possible in an airline business. To increase revenue, revenue management (yield management) is used in airline industry to maximise the profit through seven elements; customer behaviour and demand forecasting, control system, revenue factors, variable cost factors, fare products, problem scale, and problem interfaces. In control system, one factor is the overbooking concept which is selling tickets over the capacity to increase capacity utilisation in case of no-show probability of ticketed passengers. However, risks are involved when implementing an overbooking strategy. The main risk is that it is unknown how many passengers will show up or not at the time of departure for the flight. If the number of passengers show up for the flight exceeds the capacity, the airline has to deny passengers and have to compensate the denied passengers which will incur as a cost, decreasing the profit. It has been noted that most of the previous works used estimation or random probability on deciding the number of go-show or no-show passengers. This is impractical as in most cases there is no defined/known probabilistic behaviour that could model most of the aspects of oversold tickets problems. The aim of this research is to solve overbooking problem in the airline industry to maximize revenue using mathematical models based on fuzzy optimisation. This will assist in deciding how many overbooks are optimal per single-leg flight.