Cooperation and Competition: Learning of Strategies and Evolution of Preferences in Prisoners' Dilemma and Hawk-DOVE Games

By means of simulations I investigate a two-speed dynamic on strategies and preferences in prisoners' dilemmas and in hawk-dove games. Players learn strategies according to their preferences while evolution leads to a change in the preference composition. With complete information about the preferences of the opponent, cooperation in prisoners' dilemmas is achieved temporarily, with "reciprocal" preferences. In hawk-dove games, a symmetric correlated strategy profile is played that does not place any weight on mutual restraint. Among preferences only "hawkish" preferences and "selfish" preferences survive. With incomplete information, the symmetric equilibrium of the game is played. In prisoners' dilemmas only "selfish" and "reciprocal" preferences survive. In hawk-dove games all preferences are present in the medium run.