The communication cost of selfishness

Computer scientists typically measure the computational requirements of multi-agent systems assuming that the agents behave as they should according to the rules. In particular, whenever an agent is asked about his private information (e. g., his personal preferences), the assumption is that he always responds truthfully. However, for applications with economic agents, the assumption of the agents being obedient is not acceptable. E.g., we cannot expect to model e-commerce without taking into account that agents will act selfishly given the price of an object and their own private valuation of the object. In economics, the field of mechanism design describes the properties that a multi-agent system must satisfy when the system-wide behavior depends on all of the agents' selfish behaviors. Agents' incentives can be manipulated by conditional payments and by deciding what information to make available to them. If agents are selfish, some tasks are impossible to achieve. E.g. it is impossible for a multi-agent system to allocate an object to the agent having the lowest private valuation for the object. However, even when the task is still achievable with selfish agents (perhaps using conditional payments), the computational requirements of the system might be significantly higher than the requirements with obedient agents. This dissertation analyses and characterizes this increase in computational requirements due to the selfishness of the agents. We find tight upper bounds on this "cost of selfishness" whenever possible. We present cases where there is no such cost, and cases where this cost is low. Finally, we present cases where the cost is high, but bounded, and cases where it is unbounded. Our analysis of the cost of selfishness is mainly in terms of communication requirements, but we also present results that are applicable for time and memory.

[1]  Dilip Mookherjee,et al.  A theory of responsibility centers , 1992 .

[2]  Dilip Mookherjee.  Decentralization, Hierarchies, and Incentives: A Mechanism Design Perspective , 2006 .

[3]  Ronald Fadel,et al.  The communication cost of selfishness: ex post implementation , 2005, TARK.

[4]  Andrew Chi-Chih Yao,et al.  Some complexity questions related to distributive computing(Preliminary Report) , 1979, STOC.

[5]  Denise Handlarski,et al.  Green , 2007 .

[6]  A. Mas-Colell,et al.  Microeconomic Theory , 1995 .

[7]  小島 健司.  Drew Fudenberg and Jean Tirole, Game Theory , 1992 .

[8]  K. Arrow,et al.  On Partitioning a Sample with Binary-Type Questions in Lieu of Collecting Observations , 1981 .

[9]  J. A. Salvato.  John wiley & sons. , 1994, Environmental science & technology.

[10]  Noam Nisan,et al.  The communication requirements of efficient allocations and supporting prices , 2006, J. Econ. Theory.

[11]  S. Bikhchandani,et al.  Weak Monotonicity Characterizes Deterministic Dominant-Strategy Implementation , 2006 .

[12]  Joan Feigenbaum,et al.  Hardness Results for Multicast Cost Sharing , 2002, FSTTCS.

[13]  Noam Nisan,et al.  Towards a characterization of truthful combinatorial auctions , 2003, 44th Annual IEEE Symposium on Foundations of Computer Science, 2003. Proceedings..

[14]  Sang Joon Kim,et al.  A Mathematical Theory of Communication , 2006 .

[15]  Ilya Segal,et al.  Advances in Economics and Econometrics: Communication in Economic Mechanisms , 2006 .

[16]  Ilya Segal,et al.  An Efficient Dynamic Mechanism , 2013 .

[17]  Ramesh Johari,et al.  Efficiency loss in market mechanisms for resource allocation , 2004 .

[18]  Noam Nisan,et al.  Multi-player and Multi-round Auctions with Severely Bounded Communication , 2003, ESA.

[19]  Theodore Groves,et al.  Information, Incentives, and Economic Mechanisms: Essays in Honor of Leonid Hurwicz , 1987 .

[20]  M. Gribaudo,et al.  2002 , 2001, Cell and Tissue Research.

[21]  P. Herings,et al.  The Private Value Single Item Bisection Auction , 2002 .

[22]  Richard P. McLean,et al.  FULL EXTRACTION OF THE SURPLUS IN BAYESIAN AND DOMINANT STRATEGY AUCTIONS , 1988 .

[23]  David C. Parkes,et al.  Applying learning algorithms to preference elicitation , 2004, EC '04.

[24]  Noam Nisan,et al.  Auctions with Severely Bounded Communication , 2007, J. Artif. Intell. Res..

[25]  J. A. Smiley,et al.  Incentives , 2019, Handbook of Research Methods and Applications in Experimental Economics.

[26]  J. Laffont.  Aggregation and revelation of preferences , 1979 .

[27]  Drew Fudenberg,et al.  Game theory (3. pr.) , 1991 .

[28]  F. Forges,et al.  Equilibria with Communication in a Job Market Example , 1990 .

[29]  S. Reichelstein.  Incentive compatibility and informational requirements , 1984 .

[30]  E. Kushilevitz,et al.  Communication Complexity: Basics , 1996 .

[31]  Theodore Groves,et al.  Limited Communication and Incentive-Compatibility , 1987 .

[32]  Ilya Segal,et al.  The communication requirements of social choice rules and supporting budget sets , 2007, J. Econ. Theory.

[33]  Roger B. Myerson,et al.  Game theory - Analysis of Conflict , 1991 .

[34]  Ilya Segal,et al.  Communication in Economic Mechanisms , 2022 .

[35]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .