Abstract: Propp and Wilson showed how to generate a Markov chain which in a finite number of steps gives a sample from the stationary distribution supported by a countable set. In this talk we show that this is feasible if the chain satisfies the Doeblin recurrence condition.
