arXiv subjects

Yuval Peres

Publications and source records attributed to Yuval Peres.

At least 253 records · Page 14Linked to original sources

arXiv2006

Random-Turn Hex and other selection games

The game of Hex has two players who take turns placing stones of their respective colors on the hexagons of a rhombus-shaped hexagonal grid. Black wins by completing a crossing between two opposite edges, while White wins by completing a crossing between the other pair of opposite edges. Although ordinary Hex is famously difficult to analyze, Random-Turn Hex--in which players toss a coin before each turn to decide who gets to place the next stone--has a simple optimal strategy. It belongs to a general class of random-turn games--called selection games--in which the expected payoff when both players play the random-turn game optimally is the same as when both players play randomly. We also describe the optimal strategy and study the expected length of the game under optimal play for Random-Turn Hex and several other selection games.

math.PR↗

arXiv2006

Valleys and the maximum local time for random walk in random environment

Let $ξ(n, x)$ be the local time at $x$ for a recurrent one-dimensional random walk in random environment after $n$ steps, and consider the maximum $ξ^*(n) = \max_x ξ(n,x)$. It is known that $\limsup ξ^*(n)/n$ is a positive constant a.s. We prove that $\liminf_n (\log\log\log n)ξ^*(n)/n$ is a positive constant a.s.; this answers a question of P. Révész (1990). The proof is based on an analysis of the {\em valleys /} in the environment, defined as the potential wells of record depth. In particular, we show that almost surely, at any time $n$ large enough, the random walker has spent almost all of its lifetime in the two deepest valleys of the environment it has encountered. We also prove a uniform exponential tail bound for the ratio of the expected total occupation time of a valley and the expected local time at its bottom.

math.PR↗

arXiv2006

Late points for random walks in two dimensions

Let $\mathcal{T}_n(x)$ denote the time of first visit of a point $x$ on the lattice torus $\mathbb {Z}_n^2=\mathbb{Z}^2/n\mathbb{Z}^2$ by the simple random walk. The size of the set of $α$, $n$-late points $\mathcal{L}_n(α)=\{x\in \mathbb {Z}_n^2:\mathcal{T}_n(x)\geq α\frac{4}π(n\log n)^2\}$ is approximately $n^{2(1-α)}$, for $α\in (0,1)$ [$\mathcal{L}_n(α)$ is empty if $α>1$ and $n$ is large enough]. These sets have interesting clustering and fractal properties: we show that for $β\in (0,1)$, a disc of radius $n^β$ centered at nonrandom $x$ typically contains about $n^{2β(1-α/β^2)}$ points from $\mathcal{L}_n(α)$ (and is empty if $β<\sqrtα $), whereas choosing the center $x$ of the disc uniformly in $\mathcal{L}_n(α)$ boosts the typical number of $α, n$-late points in it to $n^{2β(1-α)}$. We also estimate the typical number of pairs of $α$, $n$-late points within distance $n^β$ of each other; this typical number can be significantly smaller than the expected number of such pairs, calculated by Brummelhuis and Hilhorst [Phys. A 176 (1991) 387--408]. On the other hand, our results show that the number of ordered pairs of late points within distance $n^β$ of each other is larger than what one might predict by multiplying the total number of late points, by the number of late points in a disc of radius $n^β$ centered at a typical late point.

math.PR↗

arXiv2005

Critical percolation on certain non-unimodular graphs

An important conjecture in percolation theory is that almost surely no infinite cluster exists in critical percolation on any transitive graph for which the critical probability is less than 1. Earlier work has established this for the amenable cases Z^2 and Z^d for large d, as well as for all non-amenable graphs with unimodular automorphism groups. We show that the conjecture holds for the basic classes of non-amenable graphs with non-unimodular automorphism groups: for decorated trees and the non-unimodular Diestel-Leader graphs. We also show that the connection probability between two vertices decay exponentially in their distance. Finally, we prove that critical percolation on the positive part of the lamplighter group has no infinite clusters.

math.PR↗

arXiv2005

Tail Bounds for the Stable Marriage of Poisson and Lebesgue

Let Ξbe a discrete set in R^d. Call the elements of Ξcenters. The well-known Voronoi tessellation partitions R^d into polyhedral regions (of varying volumes) by allocating each site of R^d to the closest center. Here we study allocations of R^d to Ξin which each center attempts to claim a region of equal volume α. We focus on the case where Ξarises from a Poisson process of unit intensity. It was proved in math.PR/0505668 that there is a unique allocation which is stable in the sense of the Gale-Shapley marriage problem. We study the distance X from a typical site to its allocated center in the stable allocation. The model exhibits a phase transition in the appetite α. In the critical case α=1 we prove a power law upper bound on X in dimension d=1. It is an open problem to prove any upper bound in d\geq 2. (Power law lower bounds were proved in math.PR/0505668 for all d). In the non-critical cases α<1 and α>1 we prove exponential upper bounds on X.

math.PR↗

arXiv2005

Fluctuation of planar Brownian loop capturing large area

We consider a planar Brownian loop $B$ that is run for a time $T$ and conditioned on the event that its range encloses the unusually high area of $πT^2$, with $T$ being large. We study the deviation of the range of the conditioned process $X$ from a circle of radius $T$, as a model for the fluctuation of a phase boundary. This deviation is measured by means of the inradius and outradius of the region enclosed by the range of $X$. We prove that in a typical realization of the conditioned measure, each of these quantities differs from $T$ by at most $T^{2/3 + ε}$.

math.PR↗

arXiv2005

Scaling limits of the uniform spanning tree and loop-erased random walk on finite graphs

Let x and y be chosen uniformly in a graph G. We find the limiting distribution of the length of a loop-erased random walk from x to y on a large class of graphs that include the discrete torus in dimensions 5 and above. Moreover, on this family of graphs we show that a suitably normalized finite-dimensional scaling limit of the uniform spanning tree is a Brownian continuum random tree.

math.PR↗

arXiv2005

Two new Markov order estimators

We present two new methods for estimating the order (memory depth) of a finite alphabet Markov chain from observation of a sample path. One method is based on entropy estimation via recurrence times of patterns, and the other relies on a comparison of empirical conditional probabilities. The key to both methods is a qualitative change that occurs when a parameter (a candidate for the order) passes the true order. We also present extensions to order estimation for Markov random fields.

math.ST↗

arXiv2005

Extra heads and invariant allocations

Let Πbe an ergodic simple point process on R^d and let Π^* be its Palm version. Thorisson [Ann. Probab. 24 (1996) 2057-2064] proved that there exists a shift coupling of Πand Π^*; that is, one can select a (random) point Y of Πsuch that translating Πby -Y yields a configuration whose law is that of Π^*. We construct shift couplings in which Y and Π^* are functions of Π, and prove that there is no shift coupling in which Πis a function of Π^*. The key ingredient is a deterministic translation-invariant rule to allocate sets of equal volume (forming a partition of R^d) to the points of Π. The construction is based on the Gale-Shapley stable marriage algorithm [Amer. Math. Monthly 69 (1962) 9-15]. Next, let Γbe an ergodic random element of {0,1}^{Z^d} and let Γ^* be Γconditioned on Γ(0)=1. A shift coupling X of Γand Γ^* is called an extra head scheme. We show that there exists an extra head scheme which is a function of Γif and only if the marginal E[Γ(0)] is the reciprocal of an integer. When the law of Γis product measure and d\geq3, we prove that there exists an extra head scheme X satisfying E\exp c\|X\|^d<\infty; this answers a question of Holroyd and Liggett [Ann. Probab. 29 (2001) 1405-1425].

math.PR↗

arXiv2005

Bootstrap percolation on infinite trees and non-amenable groups

Bootstrap percolation on an arbitrary graph has a random initial configuration, where each vertex is occupied with probability p, independently of each other, and a deterministic spreading rule with a fixed parameter k: if a vacant site has at least k occupied neighbors at a certain time step, then it becomes occupied in the next step. This process is well-studied on Z^d; here we investigate it on regular and general infinite trees and on non-amenable Cayley graphs. The critical probability is the infimum of those values of p for which the process achieves complete occupation with positive probability. On general trees, we find the following discontinuity: if the branching number of a tree is strictly smaller than k, then the critical probability is 1, while it is 1-1/k on the k-ary tree. A related result is that in any rooted tree T, there is a way of erasing k children of the root, together with all their descendants, and repeating this for all remaining children, and so on, such that the remaining tree T' has branching number \br(T') \leq \max (\br(T)-k, 0). We also prove that on any 2k-regular non-amenable graph, the critical probability for the k-rule is strictly positive.

math.PR↗

arXiv2005

Anchored expansion, percolation and speed

Benjamini, Lyons and Schramm [Random Walks and Discrete Potential Theory (1999) 56-84] considered properties of an infinite graph G, and the simple random walk on it, that are preserved by random perturbations. In this paper we solve several problems raised by those authors. The anchored expansion constant is a variant of the Cheeger constant; its positivity implies positive lower speed for the simple random walk, as shown by Virag [Geom. Funct. Anal. 10 (2000) 1588-1605]. We prove that if G has a positive anchored expansion constant, then so does every infinite cluster of independent percolation with parameter p sufficiently close to 1; a better estimate for the parameters p where this holds is in the Appendix. We also show that positivity of the anchored expansion constant is preserved under a random stretch if and only if the stretching law has an exponential tail. We then study a simple random walk in the infinite percolation cluster in Cayley graphs of certain amenable groups known as ``lamplighter groups.'' We prove that zero speed for a random walk on a lamplighter group implies zero speed for random walk on an infinite cluster, for any supercritical percolation parameter p. For p large enough, we also establish the converse.

math.PR↗

arXiv2005

Fast simulation of new coins from old

Let S\subset (0,1). Given a known function f:S\to (0,1), we consider the problem of using independent tosses of a coin with probability of heads p (where p\in S is unknown) to simulate a coin with probability of heads f(p). We prove that if S is a closed interval and f is real analytic on S, then f has a fast simulation on S (the number of p-coin tosses needed has exponential tails). Conversely, if a function f has a fast simulation on an open set, then it is real analytic on that set.

math.PR↗

arXiv2005

Zeros of the i.i.d. Gaussian power series: a conformally invariant determinantal process

Consider the zero set of the random power series f(z)=sum a_n z^n with i.i.d. complex Gaussian coefficients a_n. We show that these zeros form a determinantal process: more precisely, their joint intensity can be written as a minor of the Bergman kernel. We show that the number of zeros of f in a disk of radius r about the origin has the same distribution as the sum of independent {0,1}-valued random variables X_k, where P(X_k=1)=r^{2k}. Moreover, the set of absolute values of the zeros of f has the same distribution as the set {U_k^{1/2k}} where the U_k are i.i.d. random variables uniform in [0,1]. The repulsion between zeros can be studied via a dynamic version where the coefficients perform Brownian motion; we show that this dynamics is conformally invariant.

math.PR↗

arXiv2005

Spherical Asymptotics for the Rotor-Router Model in Z^d

The rotor-router model is a deterministic analogue of random walk invented by Jim Propp. It can be used to define a deterministic aggregation model analogous to internal diffusion limited aggregation. We prove an isoperimetric inequality for the exit time of simple random walk from a finite region in Z^d, and use this to prove that the shape of the rotor-router aggregation model in Z^d, suitably rescaled, converges to a Euclidean ball in R^d.

math.PR↗

arXiv2005

Evolving sets, mixing and heat kernel bounds

We show that a new probabilistic technique, recently introduced by the first author, yields the sharpest bounds obtained to date on mixing times of Markov chains in terms of isoperimetric properties of the state space (also known as conductance bounds or Cheeger inequalities). We prove that the bounds for mixing time in total variation obtained by Lovasz and Kannan, can be refined to apply to the maximum relative deviation $|p^n(x,y)/π(y)-1|$ of the distribution at time $n$ from the stationary distribution $π$. We then extend our results to Markov chains on infinite state spaces and to continuous-time chains. Our approach yields a direct link between isoperimetric inequalities and heat kernel bounds; previously, this link rested on analytic estimates known as Nash inequalities.

math.PR↗

arXiv2005

Universal finitary codes with exponential tails

In 1977, Keane and Smorodinsky showed that there exists a finitary homomorphism from any finite-alphabet Bernoulli process to any other finite-alphabet Bernoulli process of strictly lower entropy. In 1996, Serafin proved the existence of a finitary homomorphism with finite expected coding length. In this paper, we construct such a homomorphism in which the coding length has exponential tails. Our construction is source-universal, in the sense that it does not use any information on the source distribution other than the alphabet size and a bound on the entropy gap between the source and target distributions. We also indicate how our methods can be extended to prove a source-specific version of the result for Markov chains.

math.PR↗

arXiv2005

Absolute continuity for random iterated function systems with overlaps

We consider linear iterated function systems with a random multiplicative error on the real line. Our system is $\{x\mapsto d_i + λ_i Y x\}_{i=1}^m$, where $d_i\in \R$ and $λ_i>0$ are fixed and $Y> 0$ is a random variable with an absolutely continuous distribution. The iterated maps are applied randomly according to a stationary ergodic process, with the sequence of i.i.d. errors $y_1,y_2,...$, distributed as $Y$, independent of everything else. Let $h$ be the entropy of the process, and let $χ= E[\log(λY)]$ be the Lyapunov exponent. Assuming that $χ< 0$, we obtain a family of conditional measures $ν_y$ on the line, parametrized by $y = (y_1,y_2,...)$, the sequence of errors. Our main result is that if $h > |χ|$, then $ν_y$ is absolutely continuous with respect to the Lebesgue measure for a.e. $y$. We also prove that if $h < |χ|$, then the measure $ν_y$ is singular and has dimension $h/|χ|$ for a.e. $y$. These results are applied to a randomly perturbed IFS suggested by Y. Sinai, and to a class of random sets considered by R. Arratia, motivated by probabilistic number theory.

math.DS↗

arXiv2005

An LIL for cover times of disks by planar random walk and Wiener sausage

Let R_n be the radius of the largest disk covered after n steps of a simple random walk. We prove that almost surely limsup_{n \to \infty}(log R_n)^2/(log n log_3 n) = 1/4, where log_3 denotes 3 iterations of the log function. This is motivated by a question of Erdős and Taylor. We also obtain the analogous result for the Wiener sausage, refining a result of Meyre and Werner.

math.PR↗