Abstract
In this paper we present, in the context of Diaconis’ paradigm, a general method to detect the cutoff phenomenon. We use this method to prove cutoff in a variety of models, some already known and others not yet appeared in literature, including a non-reversible random walk on a cylindrical lattice. All the given examples clearly indicate that a drift towards the opportune quantiles of the stationary measure could be held responsible for this phenomenon. In the case of birth-and-death chains this mechanism is fairly well understood; our work is an effort to generalize this picture to more general systems, such as systems having stationary measure spread over the whole state space or systems in which the study of the cutoff may not be reduced to a one-dimensional problem. In those situations the drift may be looked for by means of a suitable partitioning of the state space into classes; using a statistical mechanics language it is then possible to set up a kind of energy-entropy competition between the weight and the size of the classes. Under the lens of this partitioning one can focus the mentioned drift and prove cutoff with relative ease.
Similar content being viewed by others
References
Aldous, D., Diaconis, P.: Shuffling cards and stopping times. Am. Math. Mon. 93(5), 333–348 (1986)
Barrera, J., Bertoncini, O., Fernández, R.: Abrupt convergence and escape behavior for birth and death chains. J. Stat. Phys. 137(4), 595–623 (2009)
Bayer, D., Diaconis, P.: Trailing the dovetail shuffle to its lair. Ann. Appl. Probab. 2(2), 294–313 (1992)
Diaconis, P.: The cutoff phenomenon in finite Markov chains. Proc. Natl. Acad. Sci. 93(4), 1659 (1996)
Diaconis, P., Saloff-Coste, L.: Separation cut-offs for birth and death chains. Ann. Appl. Probab. 16(4), 2098–2122 (2006)
Diaconis, P., Graham, R.L., Morrison, J.A.: Asymptotic analysis of a random walk on a hypercube with many dimensions. Random Struct. Algorithms 1(1), 51–72 (1990)
Ding, J., Lubetzky, E., Peres, Y.: Total variation cutoff in birth-and-death chains. Probab. Theory Relat. Fields 146(1), 61–85 (2010)
Erdős, P., Rényi, A.: On a classical problem of probability theory. Magy. Tud. Akad. Mat. Kut. Intéz. Közl. 6, 215–220 (1961)
Feller, W.: An Introduction to Probability Theory and Applications, vol. 1. Wiley, New York (1968)
Feller, W.: An Introduction to Probability Theory and Applications, vol. 2, 2nd edn. Wiley, New York (1968)
Jerrum, M.: Counting, Sampling and Integrating: Algorithms and Complexity. Birkhäuser, Heidelberg (2003)
Levin, D.A., Peres, Y., Wilmer, E.L.: Markov Chains and Mixing Times. American Mathematical Society, Providence (2006)
Levin, D.A., Luczak, M.J., Peres, Y.: Glauber dynamics for the mean-field Ising model: cut-off, critical power law, and metastability. Probab. Theory Relat. Fields 146(1), 223–265 (2010)
Martinez, S., Ycart, B.: Decaying rates and cutoff for convergence and hitting times of Markov chains. Adv. Appl. Probab. 33(1), 188–205 (2001)
Presutti, E.: Scaling Limits in Statistical Mechanics and Microstructures in Continuum Mechanics. Springer, Berlin (2008)
Acknowledgements
We want to thank Roberto Fernández, Alex Gaudillere and Elisabetta Scoppola for useful discussions and comments. Also we thank an anonymous referee for many helpful comments, leading to an improvement of the paper. This paper is dedicated to the memory of Roberta Dal Passo.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A: Mean Value and Variance of \({\zeta_{n}^{1}}\) for the Mean-Field Ising Model
In this appendix we present in full details the estimates for \({\mathbb{E}[{\zeta_{n}^{1}}]}\) and \(\text{Var}[{\zeta_{n}^{1}}]\) we have used to apply Corollary 1.3 to the magnetization chain in Sect. 3.6. Since for β=0 the magnetization chain reduces to the Ehrenfest chain, the following estimates hold as well for the Ehrenfest Urn model presented in Sect. 3.3.
Standard formulas (see e.g. [2]) give
where ζ k→k−1 is the first time the chain visits k−1 after visiting k and
Let us begin rewriting the ratio of the two binomial coefficients as
Next, note that for any of the values of triple (i,j,k) involved in the calculations
So we find handy the following two easy lemmas.
Lemma A.1
For x∈[0,1]
Lemma A.2
For 0≤y≤x≤1
In virtue of Lemma A.2 we can bound line (A.5) as follows:
Thus, for \({\frac{1}{2}\sqrt{\frac{n}{1-\beta}}}\leq k \leq\frac {n}{2}-\log n\),
Therefore we obtain the following upper bounds:
and
From previous computations, noticing that
we have that
and then by summation
From (A.2), using (A.11) and (A.22)–(A.23), we can easily bound the variance of ζ k→k−1 as follows:
Therefore \(\text{Var}[{\zeta_{n}^{1}}] = \sum_{k={\frac{1}{2}\sqrt {\frac{n}{1-\beta}}}+1}^{\frac{n}{2}} \text{Var}[\zeta_{k\to k-1}] \) grows at most as O(n 2).
Eventually, let us bound from below the expectation \({\mathbb{E}[{\zeta_{n}^{1}}]}\). From (A.1) and (A.5) we have
Then we have
where ε 1 tends to 0 exponentially fast in n.
Remark A.1
The error ε 1 gives a negligible contribution to \({\mathbb{E}[{\zeta_{n}^{1}}]}\) being exponentially small, for this reason we will henceforth drop it.
The right-hand in (A.32) can be rewritten as follows
with ε 2=o(n −1). Now set \(\varphi= -\frac{4k^{2}}{n^{2}}+\frac{2}{n}-\frac{4k}{n^{2}}\), then (A.35) can be rewritten as follows
where ε=O(log−2(logn)) and \(\varepsilon_{3} = O ( n^{-\frac{1}{2}}\log\log n )\).
Therefore
where
and
Now, γ can be rewritten as
Therefore γ tends asymptotically to 0.
The right-hand in (A.39) now becomes
from which we see that, to the leading order in n
Appendix B: Mean Value and Variance of \({\zeta_{n}^{1}}\) for the Partially Diffusive Random Walk
Standard formulas (see e.g. [2]) give
where ζ k→k−1 is the first time the chain visits k−1 after visiting k. By means of (3.46)–(3.48) and reversibility,
Using the properties of the exponential integral we get
and therefore
Similarly, for n sufficiently large we have that
From (B.7) we see that for n sufficiently large \({\mathbb{E}[{\zeta_{n}^{1}}-{\zeta_{n}^{\theta}}]}\) grows as n 2ε at most.
To compute \(\text{Var}[{\zeta_{n}^{1}}]\) we use the following formulas
Then we estimate the sum from below as its first term
and from above as
From (B.10) and (B.15) we see that \(\text{Var}[\zeta_{k\to k-1}] = O ( \frac{n^{2}}{k^{2}} )\) and therefore, to the leading order, \(\text{Var}[{\zeta_{n}^{1}}] = O ( n^{2-\varepsilon} )\).
Rights and permissions
About this article
Cite this article
Lancia, C., Nardi, F.R. & Scoppola, B. Entropy-Driven Cutoff Phenomena. J Stat Phys 149, 108–141 (2012). https://doi.org/10.1007/s10955-012-0584-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10955-012-0584-9