The economics analysis of a Q -learning model of cooperation with punishment and risk taking preferences

摘要

The aim of this paper is to better understand how cooperation mechanisms work in the context of a Q-learning model. We apply a learning reinforcement model to analyse the conditions needed to have a stable cooperative equilibrium when people take part in a common project and could take advantages of free-riding. Our results show that a stable equilibrium can be reached thank to mechanisms of punishment, but the final result strongly depends on the risk-taking individuals' preferences. In particular, we find that the penalties will be effective only with people having high exploration rates,namely with people able to adapt their strategies and learn to cooperate. Otherwise, it is possible to have an unstable equilibrium with cooperation until individuals have a very high intrinsic motivation to cooperate, whatever the others do.