|
Fang Wu
Information Dynamics Laboratory, HP Labs
Abstract
This paper models continuous-time mass bidding markets, such as
keyword auctions and market-based resource allocation systems, as a
stochastic dynamic system that fluctuates around an average value
under the influence of its users. The user's objective to maximize
his long-term average utility is formulated as a stochastic control
problem. The optimal bidding strategy is calculated both
analytically and numerically. It is shown that market fluctuations
tend to decrease expected system revenue, thus search engines like
Google and Yahoo have an incentive to create a secondary stable
market such as a futures market or a reservation market.
Full paper in PDF format
|
 |
|
Printable version
