Journées de l'optimisation 2007 Optimizations days

In this paper, we investigate the use of distribution free approach as an approximation scheme for pricing in Revenue management. The approach determines both the pricing and quantity jointly to best maximize the revenue. It is first addressed to the standard newsvendor problem. The analysis is extended to consider: i) Newsvendor problem with shortage cost penalty; ii) Newsvendor problem extended to shortage and holding cost; iii) Recourse case, where there is a second purchasing opportunity; iv) Random yield case; and v) Multiple item case, where more than one item compete for a scare resource such as budget. We show that the distribution free estimates on revenue in most case studies is quasi-concave, and therefore results global optimal control of pricing and capacity such that the revenue is maximized. The distribution free estimates on revenue are found near optimal in a numerical experimentation. A statistical analysis is also reported to compare the standard newsvendor problem with its extensions.

Profit maximization relies heavily on the integration of logistics activities with an improved management of the revenues. Therefore, the operational policies chosen by the carrier have an important impact on the network yield and thus on global profitability. This talk focuses on an operational bilevel mathematical formulation which encompasses pricing decisions and network planning policies such as car blocking and routing as well as train makeup and scheduling. An exact solution approach based on a mixed integer formulation adapted to the problem structure will be presented.

This paper examines how discrete choice models can be used in the airline industry for forecasting passenger demand. Increases in data processing capabilities over the past years have promoted the use of these models. Based on a individual’s revealed choice when facing several alternatives, they allow for the introduction of product related attributes and capture the weight that each attribute carries in the individual’s decision making process. Attributes are ticket price, time in transit, level and class of service, etc. From the carrier’s perspective, they provide valuable insight on how passengers will react to different scheduling and pricing policies. The carrier can then use this information in its revenue maximization quest.

Nous nous intéressons à un problème de tarification dans un réseau. Un modèle de choix discret de type Logit est employé afin de modéliser le comportement des usagers. Suite à une analyse conjointe de la fonction de revenu et de la topologie du réseau, la structure d'une tarification localement optimale est caractérisée. Par ailleurs, la substitution d'une distribution discrète à la loi de probabilité des variables aléatoires modélisant la désutilité des routes permet d'approximer l'affectation probabiliste par un flot déterministe multiclasse pour lequel des outils de résolution globale sont disponibles.