Muddling through: Noisy equilibrium selection

This paper examines an evolutionary model in which the primary source of ''noise'' that moves the model between equilibria is not arbitrarily improbable mutations but mistakes in learning. We model strategy selection as a birth-death process, allowing us to find a simple, closed-form solution for the stationary distribution. We examine equilibrium selection by considering the limiting case as the population gets large, eliminating aggregate noise. Conditions are established under which the risk-dominant equilibrium in a 2 x 2 game is selected by the model as well as conditions under which the payoff-dominant equilibrium is selected. (C) 1997 Academic Press.