TY - GEN
T1 - Towards characterizing the deterministic combinatorial constrained efficient space
AU - Gonen, Rica
AU - Lerner, Anat
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.
PY - 2019
Y1 - 2019
N2 - It is often the case that economic mechanisms can have hidden characteristics with unexpected consequences in practice. This is especially true when real-world budgets come into play. For instance, the infamous German and British 3G spectrum auctions that generated revenue so successfully that cellular service providers delayed the rollout of 3G networks for lack of funds. We contribute a piece of the puzzle by characterizing the space of deterministic, dominant-strategy incentive compatible, individually rational, and Pareto-optimal combinatorial auctions where efficiency does not hold. We examine a model with two players and k nonidentical items (2k outcomes), multidimensional types, private values, nonnegative prices, and quasilinear preferences for the players with one relaxation - one of the players is subject to a publicly-known budget constraint. We show that if it is publicly known that the valuation for the largest bundle is more than the budget for at least one of the players then the following are true. (a) VCG does not fulfill the basic properties of deterministic, dominant-strategy incentive compatible, individual rationality and Pareto optimality when the all-item bundle is not arbitrarily allocated. (b) Of the dictatorial solutions, only a single family of non-trivial dictatorial mechanisms fulfills the above basic properties. (c) Weakening the public knowledge assumption results in no VCG nor dictatorship mechanisms that fulfill the properties. Our characterization of the non-efficient space for deterministic budget-constrained combinatorial auctions is similar in spirit to that of [20] for Bayesian single-item constrained efficiency auctions.
AB - It is often the case that economic mechanisms can have hidden characteristics with unexpected consequences in practice. This is especially true when real-world budgets come into play. For instance, the infamous German and British 3G spectrum auctions that generated revenue so successfully that cellular service providers delayed the rollout of 3G networks for lack of funds. We contribute a piece of the puzzle by characterizing the space of deterministic, dominant-strategy incentive compatible, individually rational, and Pareto-optimal combinatorial auctions where efficiency does not hold. We examine a model with two players and k nonidentical items (2k outcomes), multidimensional types, private values, nonnegative prices, and quasilinear preferences for the players with one relaxation - one of the players is subject to a publicly-known budget constraint. We show that if it is publicly known that the valuation for the largest bundle is more than the budget for at least one of the players then the following are true. (a) VCG does not fulfill the basic properties of deterministic, dominant-strategy incentive compatible, individual rationality and Pareto optimality when the all-item bundle is not arbitrarily allocated. (b) Of the dictatorial solutions, only a single family of non-trivial dictatorial mechanisms fulfills the above basic properties. (c) Weakening the public knowledge assumption results in no VCG nor dictatorship mechanisms that fulfill the properties. Our characterization of the non-efficient space for deterministic budget-constrained combinatorial auctions is similar in spirit to that of [20] for Bayesian single-item constrained efficiency auctions.
KW - Budget constraints
KW - Dictatorship
KW - Incentive compatibility
KW - Pareto efficiency
UR - http://www.scopus.com/inward/record.url?scp=85075637754&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-31489-7_3
DO - 10.1007/978-3-030-31489-7_3
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:85075637754
SN - 9783030314880
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 32
EP - 48
BT - Algorithmic Decision Theory - 6th International Conference, ADT 2019, Proceedings
A2 - Pekec, Saša
A2 - Venable, Kristen Brent
PB - Springer
T2 - 6th International Conference on Algorithmic Decision Theory, ADT 2019
Y2 - 25 October 2019 through 27 October 2019
ER -