Les modèles de comportements adaptatifs appliqués à l'oligopole de Cournot

This paper examines a recurrent oligopoly game, where firms imitate the most successful behaviors in their memory. We intend to characterize stochastically stable states. Several models are considered which differ in the specification of players' memory. We note that the specification of memory plays a key role in the selection of the stochastically stable states. Moreover, in most cases, the Cournot-Nash equilibrium is not the unique stochastically stable state, and there exist cases where this equilibrium does not belong to the set of stochastically stable states.