scientific article; zbMATH DE number 3898636

From MaRDI portal
Publication:3677537





[For the entire collection see Zbl 0549.00028.] The paper concerns stopped Markov decision models with a finite state space, compact action space, and with the average return criterion. With the help of dynamic programming arguments the form of the optimal stationary policy as well as properties of the value function are discussed. This policy is shown to be optimal also in the class of non- Markov policies (depending on the whole past history), if the quantity to be maximized is the lim sup of the utility function.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3677537)