|
Titre :
|
Actualizing behaviors through a learning process
|
|
Auteurs :
|
A. Jean Marie ;
M. Tidball ;
LABORATOIRE MONTPELLIERAIN D'ECONOMIE THEORIQUE ET APPLIQUEE
|
|
Type de document :
|
rapport
|
|
Editeur :
|
Montpellier : Faculté des sciences économiques, 2002
|
|
Format :
|
26 p. + ann.
|
|
Note générale :
|
Sigle : LAMETA
Diffusion tous publics
|
|
Langues:
|
= Anglais
|
|
Catégories :
|
Methodologie
|
|
Mots-clés:
|
THEORIE DU JEU
|
|
Résumé :
|
This paper concerns the concepts of reactions and conjectures of players in a dynamic context. We propose a framework where players form conjectures on how the other players react (or would react) to their actions, and update this belief according to observations. They take decisions based on this conjecture, and to not know profit functions of other agents. Players somehow "learn" about the behavior (including the conjectures) of the other players. We show that some optimality and the local stability of the strategies obtained in the long run. Two examples are analyzed in detail: Cournot's oligopoly and Bertrand's duopoly.
|
