Average optimality for risk-sensitive control with general state space (Q2455059)

A discrete-time Markov control process on a general state space is considered. The aim of the paper is to establish the optimality inequality for risk-sensitive dynamic programming and derive an optimal stationary policy. A similar result was obtained by Hernández-Hernández and Marcus under the assumption that there exists a stationary policy which induces a finite average cost that is equal some constant in each state. Here, instead of this assumption, the author assumes that a certain family of functions is bounded which makes the process reach ``good states'' sufficiently fast. For related papers see: [\textit{D. Hernández-Hernández} and \textit{S. I. Marcus}, Appl. Math. Optim. 40, 273--285 (1999; Zbl 0937.90115)].