Microsoft Research CORE11
''Establishing a Theory for Exchange of Multiple Indivisible Goods with Indifferences''
In this project we investigate a problem of exchanging indivisible goods where each agent is initially endowed with multiple goods and agents' preferences contain indifferences. Our main objective is to clarify the existence of exchange mechanisms that satisfy three desirable properties, namely individual rationality, strategy-proofness, and Pareto efficiency. More precisely, we would like to establish a complete characterization of agents' preferences under which the existence of such mechanisms are guaranteed. This is considered a generalization of the results by Sonoda et al (2014). If we find some cases where such mechanisms do not exist, we then develop a mechanism that is not perfectly strategy-proof but approximately strategy-proof, e.g., for most of the cases no agent has an incentive to cheat. This project is financially supported by Microsoft Research CORE11, as a part of Mt.Fuji Plan.
Keywords: algorithmic game theory; mechanism design; matching; Core assignments; top-trading-cycles; microeconomics; kidney exchange; housing market
Articles/Activities Supported by the Project:
Other Related Articles: