|
|
Intelligent Commerce Research Group (ICRG)
Our research group is focused on the design and evaluation of automated negotiation for electronic commerce. This includes the following four research thrusts:
| Auctions for Electronic Commerce |
This is a general investigation of the application of auctions to electronic commerce. It involves both inventing new auction types and investigating the computational aspects of building configurable auction servers like the Michigan Internet AuctionBot.
In the fall of 2000, David Taylor developed an XML Schema to represent our auction parametrization framework. For details, see David's report and the schema files.
|
| Combinatorial Auctions |
These auctions are designed to allocate resources when agents have complementary preferences, that is, when the value that the agent receives from a combination of resources is more than additive. The Ascending k Bundle Auction (AkBA) is the mechanism we have developed for this purpose.
More recently, proxy bidding has been proposed in the context of combinatorial auctions. We have developed an algorithm for the auctioneer to use to compute the outcome entailed by the proxy bids it has received.
|
| Trading Agents |
In particular, I am interested in developing a methodology for flexible trading agents that automatically synthesis bidding strategies from the user's preferences, the auction rules, and models of the other market participants. The Trading Agent Competition presents examples of the types of problems we expect software agents to face in e-commerce. |
| General Game Playing Agents |
The General Game Playing contest poses a research challenge that involves writing a computer program that can play in arbitrary logical games. |
| Market-based Protocols |
This is the analysis of the combination of auctions and agents as a computational system. For example, we have investigated the application of market-based negotiation to distributed scheduling problems and the coordination of large numbers of users in resource-constrained environments. |
| Alphabet Soup |
The Alphabet Soup project is an open-source simulation of warehouses and manufacturing centers that use mobile robots to move inventory to workstations. It is a testbed for market-based protocols and other approaches to system control in environments with hundreds of robots. |
Current Students
If you are interested in joining our research group, I strongly recommend that you take a course in either artificial intelligence (CSC411, CSC520) or e-commerce (CSC413, CSC513) before undertaking research in this area.
Alumni
Ph.D. Graduates
- Gangshu Cai (Phd Spring 2005),
Thesis title: Flexible decision-making in sequential auctions.
Gangshu is currently an Assistant Professor at Texas A&M International University in Laredo, Texas.
- Ashish Sureka (Ph.D. spring 2005),
Thesis title: Techniques for finding Nash equilibria in combinatorial auctions.
Ashish is now working at Infosys Technologies in Bangalor, India.
M.S. Graduates:
- Neha Jain, 2006. Thesis: A Web Services Based System for Parsing and Managing Calls for Papers. Neha now works at Deloitte Touche Tohmatsu.
- Ratna Singh, 2006. Thesis: Design and Implementation of fauxBay: Test Bed for Bidding Agents in Online Auction Markets.
- Sameer Korrapati, 2004. Thesis: A Simple Auction for Early Delivery of High-Demand Products. Sameer now works at Amazon.com.
- Naveenkumar Muguda, 2004. Thesis: Non-cooperative Planning in Multi-Agent Environments with Markets for Reservations. Naveen now works at IBM in RTP.
- Erik Dahlgren, Masters Thesis officially from University of Lund, Sweden, 2003. Thesis: PackaTAC: A Conservative Trading Agent. Erik now works for NCSU.
- Shengli Bao, 2002. Thesis: A Comparison of Two Algorithms for Clearing Multi-unit Bid Double Auctions. Shengli now works at GE’s Employers Reinsurance Corporation.
- Priti P. Phadke, 2002. Thesis: An Evolutionary Approach to Finding Bidding Strategies in a Combinatorial Auction. Priti now works at Concretio, Inc.
- Harshit S. Shah, 2002. Thesis: Identifying Bidding Strategies on eBay: A Feasibility Study. Harshit now works at eBay.
- Jorge E. Prado Lozko, 2002. Thesis: Tradable Reservations in Multi-agent, Resource-constrained Environments. Jorge’s last known employer was Divine, Corp.
- Harish G. Kammanahalli, 2001. Thesis: Effects of Approximation on an Iterative Combinatorial Auction. Harish’s last known employment was with Satyam Computers, India.
- Weili Zhu, 2001. Thesis: The Application of Monte Carlo Sampling to Sequential Auction Games with Incomplete Information. Weili is working in China.
|
