Learning in multilevel games with incomplete information. I

摘要

A model is presented of learning automata playing stochastic gamesnat two levels. The high level represents the choice of the gamenenvironment and corresponds to a group decision. The low levelnrepresents the choice of action within the selected game environment.nBoth of these decision processes are affected by delays in theninformation state due to inherent latencies or to the delayed broadcastnof state changes. Analysis of the intrinsic properties of this Markovnprocess is presented along with simulated iterative behavior andnexpected iterative behavior. The results show that simulation agreesnwith expected behavior for small step lengths in the iterative map. AnFeigenbaum diagram and numerical computation of the Lyapunov exponentsnshow that, for very small penalty parameters, the system exhibitsnchaotic behavior