Value iteration in average cost Markov control processes on Borel spaces (Q1906804): Difference between revisions

This paper deals with discrete-time Markov control processes with Borel state and control spaces, allowing unbounded costs and noncompact control constraint sets. For the average cost criterion, the authors proved that, under suitable conditions, the value iteration algorithm converges to the optimal average cost and obtained an optimal policy.