A Game Theoretic Analysis of Agent -Mediated Resource Allocation

Developments in information technology have necessitated dynamic distributed real-time allocation of computational and network resources. We consider the use of market mechanisms to regulate a set of autonomous agents that are responsible for obtaining services. By applying game-theoretic analysis to a proportionally fair divisible auction, we show the existence of a unique Nash equilibrium in both single and multiple resource settings. Locally stable decentralized negotiation algorithms are developed for both cases. We also investigate the effects of coalition formation and show that the standard assumptions from classical cooperative game theory for determining the value of a team do not apply. Finally, we examine a larger space of mechanisms and optimize with respect to revenue generation and social welfare. This leads to the design of transparent and maximally efficient resource allocation schemes which have the minimum costs for signaling and computation.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Electrical Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Maheswaran, Rajiv Tharmeswaran

Contributors dc:contributor

• Basar, Tamer

• Economics, Theory

Language dc:language

eng

Identifier

(MiAaPQ)AAI3086130

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/80823