In this paper, we propose a new way to model the discrete customer choice behavior through constructing a few preference orders representing the gross customer choice behavior. This approach turns the customer choices model into a stochastic programming problem. Our model descriptions show that modeling customer choices by preference orders would be more solvable in a large scale network. In addition, we show our heuristics, called backup, as the way to properly apply the seat allocation regarding highly volatile demands. Numerical results are presented to illustrate the effectiveness of the combination of preference orders and our backup heuristic.
展开▼