Nash equilibria of \(n\)-player repeated games with semi-standard information (Q2639786)

This paper characterizes the repeated game payoffs in a class of infinitely repeated, \(n\)-player, undiscounted games with imperfect monitoring. Different methods are used to evaluate the long run average of payoffs. An upper equilibrium is a strategy profile in which each players' average profit is well defined and the lim sup of the average profit available through a unilateral deviation is less than the average profit available from the strategy profile for each agent. A lower equilibrium is obtained if lim inf replaces lim sup in the definition of upper equilibrium. It is immediate that any upper equilibrium is a lower equilibrium. The paper demonstrates that the converse is false, shows that the upper equilibrium payoff set coincides with the payoffs obtained when limits of average payoffs are evaluated using any Banach limit, and characterizes the set of upper and lower equilibrium payoffs in terms of individual rationality and feasibility conditions.