上海科技大学知识管理系统

This paper studies one-sided matching with initial endowments and the social connections between participants are specifically considered (their social network). Different from the traditional setting, the matching starts with a group of initial participants, and the others need their invitation to join the matching game. Thus, the incentive compatibility (IC) notion here considers both reporting preferences and inviting others via their social connections. This is challenging because inviters and invitees might compete in the game. Besides IC, stability and optimality are two properties extensively studied in matching, but they both are incompatible with the new IC. In the new setting, compatible stability has been discussed, but no discussion about compatible optimality yet. We complete this and prove the theoretical boundaries regarding both stability and optimality in the new setting. We then propose a mechanism called Connected Trading Cycles to satisfy all the desirable properties for the first time. Unlike the previous solutions that add restrictions on participants' matching choices to achieve IC, we allow participants to choose anyone in the game, which in principle improves everyone's satisfiability in the matching. Finally, we give the first characterization of IC in the network setting to facilitate IC mechanism design.