Mathematical programming of airline revenue management with passenger choice behavior

Abstract

Mathematical programming models of airline seat inventory control tend to protect more seats for the high fare class. In order to further study the properties of booking policy based on mathematical programming models, we propose the deterministic and stochastic models that incorporate passenger choice behavior and develop efficient genetic algorithm(GA) to solve the stochastic programming model. In the experiments, we make an evaluation between the mathematical programming models and the decision rules based on traditional EMSR and EMSRb models in three aspects: the percentage of demand diversion, the number of fare classes and the demand level. The results show that, mathematical programming models' tendency to overprotect high-fare demand can make them perform better when adopted to control seat inventory with passenger demand diversion in some situations.