Bayesian estimation of the mean holding time in average semi-Markov control processes
Tom 42 / 2015
Applicationes Mathematicae 42 (2015), 205-218
MSC: Primary 93E10; Secondary 90C40.
DOI: 10.4064/am42-2-7
Streszczenie
We consider semi-Markov control models with Borel state and action spaces, possibly unbounded costs, and holding times with a generalized exponential distribution with unknown mean $\theta $. Assuming that such a distribution does not depend on the state-action pairs, we introduce a Bayesian estimation procedure for $\theta $, which combined with a variant of the vanishing discount factor approach yields average cost optimal policies.
