Analysis of a risk-sensitive control problem for hidden Markov chains

In this paper the risk-sensitive control of partially observed Markov decision processes is considered. The replacement problem is analyzed in this context, and the structure of risk sensitive optimal controllers is given.