Abstract:The cutoff phenomenon describes a case where a Markov chain exhibits a sharp transition in its convergence to stationarity. Diaconis [Proc Natl Acad Sci USA 93(4):1659–1664, 1996] surveyed this phenomenon, and asked how one could recognize its occurrence in families of finite ergodic Markov chains. Peres [American Institute of Mathematics (AIM) Research Workshop, Palo Alto. http://www.aimath.org/WWN/mixingtimes , 2004] noted that a necessary condition for cutoff in a family of reversible chains is that the product of the mixing-time and spectral-gap tends to infinity, and conjectured that in many settings, this condition should also be sufficient. Diaconis and Saloff-Coste [Ann Appl Probab 16(4):2098–2122, 2006] verified this conjecture for continuous-time birth-and-death chains, started at an endpoint, with convergence measured in separation . It is natural to ask whether the conjecture holds for these chains in the more widely used total-variation distance. In this work, we confirm the above conjecture for all continuous-time or lazy discrete-time birth-and-death chains, with convergence measured via total-variation distance. Namely, if the product of the mixing-time and spectral-gap tends to infinity, the chains exhibit cutoff at the maximal hitting time of the stationary distribution median, with a window of at most the geometric mean between the relaxation-time and mixing-time. In addition, we show that for any lazy (or continuous-time) birth-and-death chain with stationary distribution π , the separation 1 − p t ( x , y )/ π ( y ) is maximized when x , y are the endpoints. Together with the above results, this implies that total-variation cutoff is equivalent to separation cutoff in any family of such chains.

Cutoff for a class of auto‐regressive models with vanishing additive noise

Limit Theory for Random Coefficient Autoregressive Process under Possibly Infinite Variance Error Sequence

Characterization of cutoff for reversible Markov chains

Minimax Optimal rates of convergence in the shuffled regression, unlinked regression, and deconvolution under vanishing noise

Asymptotics for A Class of Dependent Random Variables

Total-variation Cutoff in Birth-and-death Chains

Cutoff for mixtures of permuted Markov chains: reversible case

Random walk on sparse random digraphs

Geometric ergodicity of trans-dimensional Markov chain Monte Carlo algorithms

Discrete sticky couplings of functional autoregressive processes

Cutoff phenomenon in nonlinear recombinations

On the least squares estimator in a nearly unstable sequence of stationary spatial AR models

Lower bounds on the rate of convergence for accept-reject-based Markov chains in Wasserstein and total variation distances

Information-theoretic classification of the cutoff phenomenon in Markov processes

Almost sure convergence rates of adaptive increasingly rare Markov chain Monte Carlo

Systemic Infinitesimal Over-dispersion on Graphical Dynamic Models

Simultaneous Cutoff on the Multitype Configuration Model

The varentropy criterion is sharp on expanders

Strong convergence rates of semi-discrete splitting approximations for stochastic Allen--Cahn equation

Cutoff for mixtures of permuted Markov chains: general case

A Self-Normalized Central Limit Theorem for Markov Random Walks