Prof. Emo Welzl and Prof. Bernd Gärtner
|Mittagsseminar Talk Information|
Date and Time: Tuesday, April 05, 2005, 12:15 pm
Duration: This information is not available in the database
Location: This information is not available in the database
Speaker: Piotr Krysta (Univ. Dortmund)
This work deals with the design of efficiently computable incentive compatible, or truthful, mechanisms for combinatorial optimization problems with multi-parameter agents. Our focus lies on the design of approximation algorithms for NP-hard mechanism design problems. In order to ensure incentive compatibility, approximation algorithms need to satisfy certain monotonicity properties.
The most efficient way to solve packing integer programs (PIPs) is LP-based randomized rounding. Unfortunately, this method is not monotone. We show that monotone primal-dual greedy algorithms are capable of achieving almost the same approximation ratios for PIPs as randomized rounding. We demonstrate this technique based on two intensively studied applications from mechanism design, namely combinatorial auctions and unsplittable flow (routing). In both cases, we can significantly improve on the previously known approximation ratios for incentive compatible mechanisms. For example, we obtain the first incentive compatible mechanism for routing in general networks that is capable of optimizing the network utilizations up to constant factors, if edge capacities are sufficiently large. The best previous mechanism only achieved approximation factors logarithmic in the size of the network.
(Joint work with Berthold Voecking)
Automatic MiSe System Software Version 1.4803M | admin login