Abstract:
Dynamic allocation and pricing problems appear in numerous frameworks such as the retail of seasonal/style goods, the allocation of fixed capacities in the travel and leisure industries (e.g., airlines, hotels, rental cars, holiday resorts), the allocation of a fixed inventory of equipment in a given period of time (e.g., equipment for medical procedures, bandwidth or advertising space in online applications), and the assignment of personnel to incoming tasks. Although dynamic pricing is a very old technique (think about haggling in a bazaar!), modern Revenue Management (RM) techniques started with US Airline Deregulation Act of 1978 . A major academic textbook is The Theory and Practice of Revenue Management by K. T. Talluri and G.J. van Ryzin. According to these authors, the basic RM issues are:

1. Quantity decisions: How to allocate capacity/output to different segments, products or channels? When to withhold products from the market?
2. Structural decisions: Which selling format to choose (posted prices, negotiations, auctions, etc..)? Which features to use for particular format (segmentation, volume discounts, bundling, etc..)?
3. Pricing decisions: How to set posted prices, reserve prices? How to price differentiate? How to price over time? How to markdown over life time?

Broadly speaking, all above questions deal in fact with issues treated in the Auction/Mechanism Design literature. Nevertheless, mechanism design has not been the tool of choice in RM: instead, most papers have focused on analyzing properties of restricted classes (sometime intuitive, sometimes rather ad-hoc) of allocation/pricing schemes. One possible explanation for this gap is that the classical auction/mechanism design literature had a strong focus on static models while the emphasis in RM is on dynamics.

Thus, what is necessary for a modern theory of RM is a blend between the elegant dynamic models from the operations research, management science, computer science (with historical focus on grand, centralized optimization and/or "ad-hoc", mechanisms), and the classical mechanism design literature (with historical focus on information/incentives in static settings). There is no doubt that such a blend will be fruitful for numerous applications. Recently, this challenge is being picked by a more or less systematic body of work appearing under the heading of Dynamic Mechanism Design. The lectures shall briefly illustrate this approach, as reflected in my recent work on the subject