Exact Power Estimates for Countable Markov Chains
1995, v.1, Issue 1, 57-78
We develop new methods for studying non-exponential asymptotics of the stationary probabilities and the convergence rates for countable Markov chains. For some well-known cases like zero drift random walks on the quadrant we get coinciding lower and upper bounds.
Keywords: Markov chains classification,stationary probabilities,Lyapunov functions,martingales,convergence rate