Paul Milgrom (Stanford University)
The US incentive auction raised a series of unprecedented research challenges for economists and computer scientists. The government settled on a simplified design in which there are "only" 600,000 constraints to be respected during the course of the auction. The problem difficulty and the need to make the auction understandable, and even "obvious" for the bidders affected not only the computational strategy, but also the auction and even the products offered. This talk will summarize the challenges and some of the research ideas that will play a role in the final design.
Light refreshments will be served before the lecture at 3:30 p.m.