The recent increase of domain--specific search engines, able to discover information unknown by general--purpose search engines, leads to their federation into a single entity, called federated search engine. In this paper, we focus on how it can effectively merge sponsored search results, provided by the domain--specific search engines, into a unique list. In particular, we discuss the case in which the same ad can be provided by multiple sources, which requires information about the ad to be merged. We approach the problem of merging and sharing the revenue using mechanism design techniques. The main impossibility result we obtain points out there exists no mechanism that satisfies the customarily required properties. Thus, we present several mechanisms that violate at most one of these properties, and we experimentally analyze them using a real--world Yahoo! dataset.
[1]
Davide Martinenghi,et al.
Proximity rank join
,
2010,
Proc. VLDB Endow..
[2]
Nicholas R. Jennings,et al.
Scalable mechanism design for the procurement of services with uncertain durations
,
2010,
AAMAS.
[3]
Sarvapali D. Ramchurn,et al.
Trust-Based Mechanisms for Robust and Efficient Task Allocation in the Presence of Execution Uncertainty
,
2009,
J. Artif. Intell. Res..
[4]
Enrico Gerding,et al.
Mechanism Design for Federated Sponsored Search Auctions
,
2011,
AAAI.
[5]
Daniele Braga,et al.
Mashing Up Search Services
,
2008,
IEEE Internet Computing.