Learning efficient Nash equilibria in distributed systems

► We analyze learning rules that depend only on a playerʼs own realized payoffs. ► Rules of this type select efficient Nash equilibria in large classes of games. ► They optimize the performance of distributed systems using minimal information.