Electronics and Computer Science (ECS), University of Southampton

Electronics and Computer Science (ECS)

Seminar: Why (Some) Computer Scientists Care about Auctions: A Tale of Shilling, Shading, and Sniping

Speaker: Dr Enrico H Gerding
Organiser: Dr Nicholas Gibbins
Time: 30/10/2006 13:00-14:00
Location: B4/4005

Abstract

In recent years there has been a growing cross-fertilisation between computer science and economics, especially due to the advancements in the fields of multi-agent systems and electronic commerce. In this talk, I will focus on auctions and discuss how these are now being applied within multi-agent systems and, conversely, how autonomous agents are being applied to participate in existing online auctions such as eBay and Yahoo!. A popular example of the latter is the use of so-called sniping agents. These are programs which can bid on your behalf at the very last moment, just before the auction closes. Our current research is related to that and focuses on how to optimally calculate the bids of agents that participate in many, perhaps even hundreds of auctions that are selling the same item, something which is common in today’s online auctions. In that case, it is often optimal to shade your bid, i.e., bid less than what you are willing to pay for a single item.

The remainder of the talk is about the design of novel types of auctions and the use of these auctions within multi-agent systems. Auctions are essentially a principled way for allocating scarce resources between one or more parties. In addition to goods, auctions can thus similarly be used within multi-agent systems to distribute scarce resources such as time slots in agendas, Grid resources, landing slots, etc. The main challenge here is to design the auctions in such a way that they result in a desired system-level performance (e.g., that time slots are allocated in the best possible way), even when individual stake-holders act in their own best interest and have possibly conflicting goals. Furthermore, it is important to design auctions that prevent cheating, such as collusion and shilling. The latter refers to the situation where a seller (or an ‘accomplice’) participates in its own auctions pretending to be a buyer in an attempt to increase the selling price.

Speaker Biography

Dr Enrico H Gerding

Enrico obtained a M.Sc. in 1999 from the Free University in Amsterdam in Artificial Intelligence, and a Ph.D. in 2004 from the University of Eindhoven. The Ph.D. research was undertaken at the Centre for Mathematics and Computer Science (CWI) in Amsterdam. He is currently a Research Fellow in the Department of Electronics and Computer Science (ECS) at the University of Southampton, within the Intelligence, Agents, Multimedia Group.