We study different aspects of the multiagent resource allocation problem when the objective is to find an allocation that maximizes Nash social welfare, the product of the utilities of the individual agents. The Nash solution is an important welfare criterion that combines efficiency and fairness considerations. We show that the problem of finding an optimal outcome is NP-hard for a number of different languages for representing agent preferences; we establish new results regarding convergence to Nash-optimal outcomes in a distributed negotiation framework; and we design and test algorithms similar to those applied in combinatorial auctions for computing such an outcome directly.
Additional Metadata
Keywords Nash social welfare, Multiagent resource allocation
THEME Software (theme 1), Logistics (theme 3), Energy (theme 4)
Publisher TU/e Printservice
Editor T. Calders , K. Tuyls (Karl) , M. Pechenizkiy
Series Benelux Conference on Artificial Intelligence
Conference Benelux Conference on Artificial Intelligence
Note This is an extended abstract of work presented in The Eleventh International Workshop on Agent Mediated Electronic Commerce, in Budapest, Hungary, 2009.
Citation
Ramezani, S, & Endriss, U. (2009). Nash Social Welfare in Multiagent Resource Allocation. In T Calders, K Tuyls, & M Pechenizkiy (Eds.), Benelux Conference on Artificial Intelligence (BNAIC) 2009 (pp. 359–360). TU/e Printservice.