Social Reinforcement: Cascades, Entrapment, and Tipping
研究了不同主体行为相互强化的现象，如网络效应和从众效应，将其建模为具有递增差异的博弈，证明均衡临界点、级联和俱乐部陷阱是相互强化的自然结果，并刻画了最小临界点集合。
The actions of different agents sometimes reinforce each other. Examples are network effects and the threshold models used by sociologists as well as (Harvey) Leibenstein's “bandwagon effects.” We model such situations as a game with increasing differences, and show that tipping of equilibria, cascading, and clubs with entrapment are natural consequences of this mutual reinforcement. If there are several equilibria, one of which Pareto dominates, then the inefficient equilibria can be tipped to the efficient one, a result of interest in the context of coordination problems. We characterize the smallest tipping set.