Speaker: William E. Walsh, Computer Science and Engineering, U. Michigan
Market Protocols for Decentralized Supply Chain Formation
Abstract: Much of the current work in electronic commerce has focused on technology for automating existing modes of decision making, such as facilitating bilateral exchange between customers and merchants or managing resource flows between suppliers and manufacturers in pre-existing commercial relationships. But to achieve the oft-expressed visions of dynamically forming and dissolving business interactions (e.g., the rhetoric of ``virtual corporations'') requires automated support for supply chain formation, the process of bottom-up assembly of complex production and exchange relationships.
Supply chain formation presents difficult coordination issues for negotiation protocols.Ý Agents (i.e. companies, individuals, or their computational representatives) must simultaneously negotiate production relationships at multiple levels, with important interdependencies among inputs and outputs at each level. Decentralization constraints, due to autonomous, self-interested agents with conflicting objectives and limited information, further complicate the problem.Ý I have formalized a supply chain formation problem that captures these features, and have designed and analyzed market protocols for the problem.
I qualitatively distinguish market protocols along two dimensions: the scope of concern of the market mechanism and the assumed agent behavior. I present market protocols at different points in this space---a distributed, progressive protocol with non-strategic behavior and a combinatorial protocol with both non-strategic and strategic behavior---and evaluate them by economic metrics.
Short Bio: William E. Walsh is a Ph.D. candidate in Computer Science and Engineering at the University of Michigan.Ý His research focuses on automated negotiation technologies for electronic commerce and market approaches to decentralized problems with complex dependencies. Recent projects include decentralized supply chain formation, constraint satisfaction, and scheduling.Ý He has been one of the primary developers of the Michigan Internet AuctionBot and the ICMAS 2000 Trading Agent Competition.Ý He is a NASA/Jet Propulsion Laboratory graduate student researcher.
Hosts: P.
Wurman, Computer Science, NCSU