Economics Faculty Database
Economics
Arts & Sciences
Duke University

 HOME > Arts & Sciences > Economics > Faculty    Search Help Login pdf version printable version 

Publications [#236191] of Vincent Conitzer

Conference articles PUBLISHED

  1. Guo, M; Conitzer, V, Strategy-proof allocation of multiple items between two agents without payments or priors, edited by Hoek, WVD; Kaminka, GA; Lespérance, Y; Luck, M; Sen, S, Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems, AAMAS, vol. 2 (January, 2010), pp. 881-888, IFAAMAS, ISBN 9781617387715 [html], [doi]
    (last updated on 2024/04/18)

    Abstract:
    We investigate the problem of allocating items (private goods) among competing agents in a setting that is both prior-free and payment-free. Specificall, we focus on allocating multiple heterogeneous items between two agents with additive valuation functions. Our objective is to design strategy-proof mechanisms that are competitive against the most efficien (first-best allocation. We introduce the family of linear increasing-price (LIP) mechanisms. The LIP mechanisms are strategy-proof, prior-free, and payment-free, and they are exactly the increasing-price mechanisms satisfying a strong responsiveness property. We show how to solve for competitive mechanisms within the LIP family. For the case of two items, we fin a LIP mechanism whose competitive ratio is near optimal (the achieved competitive ratio is 0.828, while any strategy-proof mechanism is at most 0.841-competitive). As the number of items goes to infinit, we prove a negative result that any increasing-price mechanism (linear or nonlinear) has a maximal competitive ratio of 0.5. Our results imply that in some cases, it is possible to design good allocation mechanisms without payments and without priors. Copyright © 2010, International Foundation for Autonomous Agents and Multiagent Systems (www.ifaamas.org). All rights reserved.


Duke University * Arts & Sciences * Economics * Faculty * Research * Staff * Master's * Ph.D. * Reload * Login