Mechanism Design with Exchangeable Allocations

Abstract: We investigate mechanism design without payments when agents have different types of preferences. Contrary to most settings in the literature where agents have the same preference, e.g. in the facility location games all agents would like to stay close to (or away from) the facility, we demonstrate the limitation of mechanism design without payments when agents have different preferences by introducing exchanging phases. We consider two types of exchanging phases. The first model is called central exchanges where the exchanges are performed by a central authority. The other model is called individual exchanges where agents exchange their outcomes by themselves. By using facility location games as an example, we provide a truthful mechanism that optimizes social welfare in central exchanges. We also provide a universally truthful randomized mechanism that achieves at least a half of the optimal social welfare in individual exchanges.