A model for airline seat control considering revenue uncertainty and risk

Most revenue management (RM) models rely on the concept of marginal seat revenue and the assumption of risk neutrality to develop an expected-revenue maximisation policy for airline seat-inventory control. For the single-leg dynamic RM problem, this study developed a modified seat control policy to take risk into consideration by discounting the marginal seat revenue and relaxing the optimality condition in the dynamic programming model. The effectiveness of the modified policy was examined by a series of simulation experiments. In particular, another risk-aversion RM model with the objective of utility maximisation was tested. The simulation analysis shows that both models can balance expected revenue and revenue variation in a systematic way so as to address risk-aversion preferences of different airlines.