10:30 AM - 10:55 AM
A Mathematical Programming Approach to Improved Bid Prices under a Parametric Choice Model of Demand
In quantity based revenue management, one of the most powerful and simple approach to control perishable inventories, consists in assigning threshold prices (``bid prices'') to each resource. We propose a new mathematical programming approach to estimate time dependent bid prices, within the framework of customer choice-based network revenue management. In contrast with most heuristics proposed in the literature, our approach is flexible and can easily accommodate technical and practical side constraints. To solve the model, we develop a modified column generation algorithm combined with an efficient heuristic procedure for addressing the NP-Hard subproblem.
10:55 AM - 11:20 AM
A Non-Parametric Algorithm of Uncensoring Demand under Availability Constraints in RM Systems
We examine the challenge of demand forecasting in revenue management. Due to
booking limits, registered reservations do not represent the real value of demand. Usually transportation companies continue to accept reservations in a fare class until the booking limit is reached. From this point forward, the data is censored. In revenue management systems, it is desired to uncensor the observations for representing the true demand. We propose an algorithm that takes availability constraints into account via a non-parametric mathematical representation. We solve the problem by introducing a new heuristic method.
11:20 AM - 11:45 AM
A Link Based Dynamic Route Choice Model with Unrestricted Choice Set
Probabilistic route choice models are central in many transport applications. Such models are typically path based and require sampling of alternatives to define choice sets. We propose a link-based formulation that require no restriction on the choice set and no generation of paths. The model can be consistently estimated and efficiently used for prediction.
11:45 AM - 12:10 PM
Dynamic Discrete Choice Model for Railway Ticket Cancellation and Exchange Behavior
We apply dynamic discrete choice model to ticket cancellation and exchange behavior in the revenue management context. Each time period, a passenger makes the decision of whether to keep, exchange or cancel the ticket. The exchange decision allows for departure time specific choices. A 1-SL policy is adopted to approximate this dynamic programming problem.