## Electronic Journal of Probability

### On sensitivity of mixing times and cutoff

#### Abstract

A sequence of chains exhibits (total variation) cutoff (resp., pre-cutoff) if for all $0<\epsilon < 1/2$, the ratio $t_{\mathrm{mix} }^{(n)}(\epsilon )/t_{\mathrm{mix} }^{(n)}(1-\epsilon )$ tends to 1 as $n \to \infty$ (resp., the $\limsup$ of this ratio is bounded uniformly in $\epsilon$), where $t_{\mathrm{mix} }^{(n)}(\epsilon )$ is the $\epsilon$-total variation mixing time of the $n$th chain in the sequence. We construct a sequence of bounded degree graphs $G_n$, such that the lazy simple random walks (LSRW) on $G_n$ satisfy the “product condition” $\mathrm{gap} (G_n) t_{\mathrm{mix} }^{(n)}(\epsilon ) \to \infty$ as $n \to \infty$, where $\mathrm{gap} (G_n)$ is the spectral gap of the LSRW on $G_n$ (a known necessary condition for pre-cutoff that is often sufficient for cutoff), yet this sequence does not exhibit pre-cutoff.

Recently, Chen and Saloff-Coste showed that total variation cutoff is equivalent for the sequences of continuous-time and lazy versions of some given sequence of chains. Surprisingly, we show that this is false when considering separation cutoff.

We also construct a sequence of bounded degree graphs $G_n=(V_{n},E_{n})$ that does not exhibit cutoff, for which a certain bounded perturbation of the edge weights leads to cutoff and increases the order of the mixing time by an optimal factor of $\Theta (\log |V_n|)$. Similarly, we also show that “lumping” states together may increase the order of the mixing time by an optimal factor of $\Theta (\log |V_n|)$. This gives a negative answer to a question asked by Aldous and Fill.

#### Article information

Source
Electron. J. Probab., Volume 23 (2018), paper no. 25, 34 pp.

Dates
Accepted: 27 February 2018
First available in Project Euclid: 15 March 2018

https://projecteuclid.org/euclid.ejp/1521079338

Digital Object Identifier
doi:10.1214/18-EJP154

Mathematical Reviews number (MathSciNet)
MR3779818

Zentralblatt MATH identifier
1387.60112

#### Citation

Hermon, Jonathan; Peres, Yuval. On sensitivity of mixing times and cutoff. Electron. J. Probab. 23 (2018), paper no. 25, 34 pp. doi:10.1214/18-EJP154. https://projecteuclid.org/euclid.ejp/1521079338

#### References

• [1] L. Addario-Berry and M. I. Roberts, Mixing Time Bounds via Bottleneck Sequences, Journal of Statistical Physics, (2017): 1–27.
• [2] D. Aldous and J. Fill, Reversible Markov chains and random walks on graphs, (2002) Unfinished manuscript. Available at http://www.stat.berkeley.edu/~aldous/RWG/book.html.
• [3] N. Alon and V. Milman, $\lambda _1$ isoperimetric inequalities for graphs, and superconcentrators. Journal of Combinatorial Theory, Series B 38.1 (1985): 73–88.
• [4] N. Alon, $\lambda _1$ Eigenvalues and expanders, Combinatorica 6.2 (1986): 83–96.
• [5] R. Basu, J. Hermon, and Y. Peres, Characterization of cutoff for reversible Markov chains, Ann. Probab. 45.3 (2017): 1448–1487.
• [6] I. Benjamini, Instability of the Liouville property for quasi-isometric graphs and manifolds of polynomial volume growth, Journal of Theoretical Probability 4.3 (1991): 631–637.
• [7] G. Y. Chen and L. Saloff-Coste, Comparison of cutoffs between lazy walks and Markovian semigroups, Journal of Applied Probability 50.4 (2013): 943–959.
• [8] P. Diaconis and L. Saloff-Coste, Logarithmic Sobolev inequalities for finite Markov chains, The Annals of Applied Probability, 6(3) (1996): 695–750.
• [9] J. Ding, E. Lubetzky and Y. Peres, Total variation cutoff in birth-and-death chains, Probability Theory and Related Fields 146.1-2 (2010): 61–85.
• [10] J. Ding and Y. Peres, Sensitivity of mixing times, Electronic Communications in Probability 18.88 (2013): 1–6.
• [11] N. Fountoulakis and B. A. Reed, Faster mixing and small bottlenecks, Probability Theory and Related Fields 137.3 (2007): 475–486.
• [12] S. Goel, R. Montenegro and P. Tetali, Mixing time bounds via the spectral profile, Electron. J. Probab 11.1 (2006): 1–26.
• [13] J. Hermon, On sensitivity of uniform mixing times, Annales de l’Institut Henri Poincaré Probabilités et Statistiques 54.1 (2018): 234–248.
• [14] J. Hermon, H. Lacoin and Y. Peres, Total Variation and Separation Cutoffs are not equivalent and neither one implies the other, Electron. J. Probab., 21(44) (2016): 1–36.
• [15] G. Kozma, On the precision of the spectral profile, ALEA Lat. Am. J. Probab. Math. Stat. 3 (2007): 321–329.
• [16] D. Levin and Y. Peres, Markov chains and mixing times, American Mathematical Soc. (2017). With contributions by Elizabeth L. Wilmer and a chapter by James G. Propp and David B. Wilson.
• [17] E. Lubetzky and A. Sly, Explicit expanders with cutoff phenomena, Electronic Journal Probability 16.15 (2011): 419–435.
• [18] A. Lubotzky. R. Phillips and P. Sharnak, Ramanujan Graphs, Combinatorica 8.3 (1988): 261–277.
• [19] M. Morgenstern, Existence and explicit constructions of q+ 1 regular Ramanujan graphs for every prime power q, Journal of Combinatorial Theory, Series B 62.1 (1994): 44-62.
• [20] B. Morris and Y. Peres, Evolving sets, mixing and heat kernel bounds, Probability Theory and Related Fields 133.2 (2005): 245–266.
• [21] Y. Peres. American Institute of Mathematics (AIM) research workshop “Sharp Thresholds for Mixing Times” (Palo Alto, December 2004). Summary available at http://www.aimath.org/WWN/mixingtimes.
• [22] Y. Peres and P. Sousi, Mixing times are hitting times of large sets. Journal of Theoretical Probability 28.2 (2013): 1–32.
• [23] C. Pittet and L. Saloff-Coste, On the stability of the behavior of random walks on groups. Journal of Geometric Analysis, 10(4) (2000): 713–737.
• [24] A. Sinclair and M. Jerrum, Approximate counting, uniform generation and rapidly mixing Markov chains, Information and Computation 82.1 (1989): 93–133.