On the vanishing discount factor approach for Markov decision processes with weakly continuous transition probabilities (Q2264001): Difference between revisions

This paper deals with average cost Markov decision processes with Borel state and control spaces, possibly unbounded costs and non-compact action subsets under the assumption of weak continuity of the transition law. It provides a simplified and somewhat elementary proof of the existence of average cost for stationary optimal policies under the same assumptions as in the paper by \textit{E. A. Feinberg} et al. [Math. Oper. Res. 37, No. 4, 591--607 (2012; Zbl 1297.90173)]. The prove is based on the concept of lower semicontinuous envelope of functions and an elementary result on the interchange of limits and minima in lieu of a Fatou's lemma for varying measures.