The PRISE-Team

Dr. Oliver Wendt

e-mail: wendt@wiwi.uni-frankfurt.de

phone: 069/798-28124

Project Manager

 

 

Dipl.-Wirtsch.-Ing. Michael Schwind

e-mail: schwind@wiwi.uni-frankfurt.de

phone: 069/798-28679

Research Field:

A solution for the task allocation problem of distributed systems should be strived by the use of reinforcement-learning agents under the constraint of optimal pricing. Showing an relationship between this allocation problem and simulations in agent based computational economics (ACE) is another objective of the project. A comparison of the classical yield-management solution using stochastic dynamic programing with the reinfocement-learning approach, as well as an game theoretic interpretation of learning individuals, is to be drawn on a theoretical level. Implementation in the practical part of the research project should be based on the agent simulation platform SWARM.

 
Dipl.- Kfm. Stefan Seibel  
e-mail: seibel@wiwi.uni-frankfurt.de phone: 069/798-22751

Research Field:

As an alternative method of resolution combinatorial auctions mechanism will be analyzed with regard to their suitability for auctioning bundles of information product ressources.
First point of view is to discuss the generals aspects of combinatorial auctions including mathematic methods for solving these integer and linear programming problems. As second step one will show and extend the game theoretic features of this subject in order to find a method which creates incentives for the bidders in an combinatorial auction to only submit bids for bundled and single items which correspondend to ther real values of the items, i.e. to externalize their real payment reserves.
As final step of the project an agent-based simulation of combinatorial auctions for information products should be developed and implemented on the basis of SWARM.