Learning and equilibrium selection in a coordination game with heterogeneous agents

We study a modified version of the coordination game presented in [van Huyck et al., 1994], where a representative selection dynamics was proposed to explain experimental data. Assuming that the agents adjust their moves in the direction of the best response, we derive a formal analysis of the stability of the equilibria. We show by simulation that the interior equilibrium is robustly reached even when considerable heterogeneity is allowed among the agents. Our truly multi-agent game is capable of approximating quite well both the median game convergence and the experimental data.