Evolutionary Social Learning Through Networks

We study evolutionary learning processes through social networks. We show that network heterogeneity, in particular scale-freeness, enhances social learning by studying a model in which an agent set on a network imitates the strategy only from the neighborhood. Further, in order to determine whether it is the heterogeneity of the game network (which determines who plays with whom) or the heterogeneity of the learning network (which determines who learns from whom) that enhances social learning, we study another model wherein a network of learning exists but a game does not exist. We show that network heterogeneity enhances social learning not because the network of learning is heterogeneous but because the network of games is heterogeneous. Moreover, we also study a coordination game on a network. We derive the condition that a Pareto optimal strategy spreads over time. We show that Pareto optimal strategy spreads if the players taking the Pareto optimal strategy are on hub vertices in the beginning