Abstract
This paper treats a Markov decision problem with an infinite planning horizon and no discounting. This model is analyzed by application, perhaps repeated, of a simple linear program.