arXiv++ Combinatorics

Consider a random permutation of $kn$ objects that permutes $n$ disjoint blocks of size $k$ and then permutes elements within each block. Normalizing its cycle lengths by $kn$ gives a random partition of unity, and we derive the limit law of this partition as $k,n \to \infty$. The limit may be constructed via a simple square cutting procedure that generalizes stick breaking in the classical case of random permutations ($k=1$). The expected size of the largest part of this square cutting distribution is approximated to be $0.40$, in contrast with the Golomb-Dickman constant around $0.624$ describing the longest cycle of a uniform random permutation as well as the largest prime factor of a random integer. The distribution function of this largest part is shown to also be the mean of a certain multiplicative function. Along the way we give the first extension of the Erdős-Turán law to a proper permutation subgroup.

Counting permutations of $[n]$ by the number of records, i.e. left-to-right maxima, is a classic problem in combinatorial enumeration. In the first volume of ``The Art of Computer Programming", Donald Knuth demonstrated its relevance for analysis of average case complexity of a basic algorithm for determining a maximum in a linear list of numbers. It is well known that the expected, and likely, number of those records in a {\it uniformly\/} random permutation is asymptotic to $\log n$. Cyril Banderier, Rene Beier, and Kurt Mehlhorn studied the case of a non-uniform random permutation, which is obtained from a generic permutation of $[n]$ by selecting its elements one after another independently with probability $p$, and permuting the selected elements uniformly at random. They proved that $E_n(p)$, the largest expected number of the maxima, is between $\text{const}\sqrt{n/p}$ and $O\bigl(\sqrt{(n/p)\log n}\bigr)$ if $p$ is fixed. For $p\gg 1/n$ and simultaneously $1-p\ge \text{const }n^{-1/2}\log n$, we prove that $E_n(p)$ is exactly of order $(1-p)\sqrt{n/p}$.

We study positivity and probabilistic properties arising from the Young--Fibonacci lattice $\mathbb{YF}$, a 1-differential poset on binary (Fibonacci) words of 1's and 2's, graded by digit sum. Building on Okada's theory of clone Schur functions (Trans. Amer. Math. Soc. 346 (1994), 549--568), we define clone coherent measures on $\mathbb{YF}$ that generate random Fibonacci words of increasing length; unlike for the Young lattice (powered by the classical Schur functions), clone coherent measures are generally not extremal on $\mathbb{YF}$. Our first main result is a complete characterization of Fibonacci positive specializations -- parameter sequences which yield positive clone Schur functions on $\mathbb{YF}$. Second, we connect Fibonacci positivity with: (i) total positivity of tridiagonal matrices; (ii) Stieltjes moment sequences; (iii) the combinatorics of set partitions; and (iv) families of univariate orthogonal polynomials from the (q-)Askey scheme. We further link moment sequences of orthogonal polynomials to combinatorial structures on Fibonacci words, a connection that may be of independent interest. Third, we analyze scaling limits of the induced random words, obtaining stick-breaking-type limits (linked to GEM laws), new dependent stick-breaking limits, and limits supported on the discrete part of the Martin boundary of $\mathbb{YF}$. These results significantly extend the asymptotics of the Plancherel measure on $\mathbb{YF}$ proved by Gnedin--Kerov (Math. Proc. Camb. Philos. Soc. 129 (2000), 433--446). Finally, we prove Cauchy-type identities for clone Schur functions with quadridiagonal-determinant right-hand side (in contrast to the product form for classical Schur functions), and construct models of random permutations and involutions from Fibonacci-positive specializations together with a Robinson--Schensted correspondence adapted to $\mathbb{YF}$.

Permutons, which are probability measures on the unit square $[0, 1]^2$ with uniform marginals, are the natural scaling limits for sequences of (random) permutations. We introduce a $d$-dimensional generalization of these measures for all $d \ge 2$, which we call $d$-dimensional permutons, and extend -- from the two-dimensional setting -- the theory to prove convergence of sequences of (random) $d$-dimensional permutations to (random) $d$-dimensional permutons. Building on this new theory, we determine the random high-dimensional permuton limits for two natural families of high-dimensional permutations. First, we determine the $3$-dimensional permuton limit for Schnyder wood permutations, which bijectively encode planar triangulations decorated by triples of spanning trees known as Schnyder woods. Second, we identify the $d$-dimensional permuton limit for $d$-separable permutations, a pattern-avoiding class of $d$-dimensional permutations generalizing ordinary separable permutations. Both high-dimensional permuton limits are random and connected to previously studied universal 2-dimensional permutons, such as the Brownian separable permutons and the skew Brownian permutons, and share interesting connections with objects arising from random geometry, including the continuum random tree, Schramm--Loewner evolutions, and Liouville quantum gravity surfaces.

Lecture notes of a master course given at Orsay between 2019-2024. Topics covered include Part I: One-dimensional random walks, cycle lemma and Bienaymé--Galton--Watson random trees. Part II: Erdös--Rényi random graphs, three proofs of the emergence of the giant component. Part III: Random recursive tree, random permutations and continuous time embedding techniques. Intended for publication.

We extend Friedman's theorem to show that, for any fixed $r>1$, a random $2r$--regular Schreier graph associated with the action of $r$ uniformly random permutations of $[n]$ on $k_{n}$--tuples of distinct elements in $[n]$ has a near--optimal spectral gap with high probability, provided $k_{n}\leq n^{\frac{1}{20}-ε}.$ Previously this was known only for $k$--tuples where $k$ is fixed. In fact, we prove the stronger result of strong convergence of random permutations in irreducible representations of quasi--exponential dimension. Along the way, we give a new bound for the expected stable irreducible character of a random permutation obtained via a word map, showing that $\mathbb{E}\left[χ^μ\left(w(σ_{1},\dots,σ_{r})\right)\right]=O\left(\frac{1}{\dimχ^μ}\right)=O\left(n^{-k}\right)$, where $k$ is the number of boxes outside the first row of the Young diagram $μ,$ solving one aspect of a conjecture of Hanany and Puder. We obtain this bound using an extension of Wise's $w$--cycle conjecture.

We introduce and analyze a natural game formulated as follows. In this one-person game, the player is given a random permutation $A=(a_1,\dots, a_n)$ of a multiset $M$ of $n$ reals that sum up to $0$, where each of the $n!$ permutation sequences is equally likely. The player only knows the value of $n$ beforehand. The elements of the sequence are revealed one by one and the player can stop the game at any time. Once the process stops, say, after the $i$th element is revealed, the player collects the amount $\sum_{j=i+1}^{n} a_j$ as his/her payoff and the game is over (the payoff corresponds to the unrevealed part of the sequence). Three online algorithms are given for maximizing the expected payoff in the binary case when $M$ contains only $1$'s and $-1$'s. $\texttt{Algorithm 1}$ is slightly suboptimal, but is easier to analyze. Moreover, it can also be used when $n$ is only known with some approximation. $\texttt{Algorithm 2}$ is exactly optimal but not so easy to analyze on its own. $\texttt{Algorithm 3}$ is the simplest of all three. It turns out that the expected payoffs of the player are $Θ(\sqrt{n})$ for all three algorithms. In the end, we address the general problem and deal with an arbitrary zero-sum multiset, for which we show that our $\texttt{Algorithm 3}$ returns a payoff proportional to $\sqrt{n}$, which is worst case-optimal.

Let $S_n$ denote the set of permutations of $[n]$ and let $σ=σ_1\cdotsσ_n\in S_n$. For a subsequence $\{σ_{i_j}\}_{j=1}^k$ of $\{σ_i\}_{i=1}^n$ of length $k\ge2$, construct the ``up/down'' sequence $V_1\cdots V_{k-1}$ defined by $$ V_j=\begin{cases} U,\ \text{if}\ σ_{i_j+1}-σ_{i_j}>0;\\ D,\ \text{if}\ σ_{i_j+1}-σ_{i_j}<0.\end{cases} $$ Consider now a fixed up/down pattern: $V_1\cdots V_l$, where $l\in\mathbb{N}$ and $V_j\in\{U, D\},\ j\in[l]$. Given a permutation $σ\in S_n$, consider the length of the longest subsequence of $σ$ that repeats this pattern. For example, consider $l=3$ and $V_1V_2V_3=UUD$. Then for the permutation $342617985\in S_9$, the length of the longest subsequence that repeats the pattern $UUD$ is 7; it is obtained by 3461798 and 3461785. The above framework includes two well-known cases. The pattern $U$ is the celebrated case of the longest increasing subsequence. The pattern $UD$ (or $DU$) is the case of the longest alternating subsequence. These have been studied both under the uniform distribution on $S_n$ as well as under the uniform distribution on those permutations in $S_n$ which avoid a particular pattern of length three. In this paper, we consider the patterns $UUD$ and $UUUD$ under the uniform distribution on those permutations in $S_n$ which avoid the pattern $132$. We prove that the expected value of the longest increasing subsequence following the pattern $UUD$ is asymptotic to $\frac37n$ and the expected value of the longest increasing subsequence following the pattern $UUUD$ is asymptotic to $\frac4{11}n$. (For $UD$ (alternating subsequences) it is known to be $\frac12n$.) This leads directly to appropriate corresponding results for permutations avoiding any particular pattern of length three.

One method to generate random permutations involves using Gaussian elimination with partial pivoting (GEPP) on a random matrix $A$ and storing the permutation matrix factor $P$ from the resulting GEPP factorization $PA=LU$. We are interested in exploring properties of random butterfly permutations, which are generated using GEPP on specific random butterfly matrices. Our paper highlights new connections among random matrix theory, numerical linear algebra, group actions of rooted trees, and random permutations. We address the questions of the longest increasing subsequence (LIS) and number of cycles for particular uniform butterfly permutations, with full distributional descriptions and limit theorems for simple butterfly permutations. We also establish scaling limit results and limit theorems for nonsimple butterfly permutations, which include certain $p$-Sylow subgroups of the symmetric group of $N=p^n$ elements for prime $p$. For the LIS, we establish power law bounds on the expected LIS of the form $N^{α_p}$ and $N^{β_p}$ where $\frac12 < α_p < β_p < 1$ for each $p$ with $α_p = 1 - o_p(1)$, showing distinction from the typical $O(N^{1/2})$ expected LIS frequently encountered in the study of random permutations (e.g., uniform permutations). For the number of cycles scaled by $(2-1/p)^n$, we establish a full CLT to a new limiting distribution depending on $p$ with positive support we introduce that is uniquely determined by its positive moments that satisfy explicit recursive formulas; this thus determines a CLT for the number of cycles for any uniform $p$-Sylow subgroup of $S_{p^n}$.

We study classical pattern counts in Mallows random permutations with parameters $(n,q_n)$, as $n\to\infty$. We focus on three different regimes for the parameter $q = q_n$. When $n^{3/2}(1-q)\to0$, we use coupling techniques to prove that pattern counts in Mallows random permutations satisfy a central limit theorem with the same asymptotic mean and variance as in uniformly random permutations. When $q\to1$ and $n(1-q)\to\infty$, we use results on the displacements of permutation points to find the order of magnitude of pattern counts. When $q\in(0,1)$ is fixed, we use the regenerative property of the Mallows distribution to compare pattern counts with certain $U$-statistics, and establish central limit theorems. We also construct a specific Mallows process, that is a coupling of Mallows distributions with $q$ ranging from $0$ to $1$, for which the process of pattern counts satisfies a functional central limit theorem.

We study random permutations arising from reduced pipe dreams. Our main model is motivated by Grothendieck polynomials with parameter $β=1$ arising in K-theory of the flag variety. The probability weight of a permutation is proportional to the principal specialization (setting all variables to 1) of the corresponding Grothendieck polynomial. By mapping this random permutation to a version of TASEP (Totally Asymmetric Simple Exclusion Process), we describe the limiting permuton and fluctuations around it as the order $n$ of the permutation grows to infinity. The fluctuations are of order $n^{\frac13}$ and have the Tracy-Widom GUE distribution, which places this algebraic (K-theoretic) model into the Kardar-Parisi-Zhang universality class. We also investigate non-reduced pipe dreams and make progress on a recent open problem on the asymptotic number of inversions of the resulting permutation. Inspired by Stanley's question for the maximal value of principal specializations of Schubert polynomials, we resolve the analogous question for $β=1$ Grothendieck polynomials, and provide bounds for general $β$.

A combinatorial object is said to be quasirandom if it exhibits certain properties that are typically seen in a truly random object of the same kind. It is known that a permutation is quasirandom if and only if the pattern density of each of the twenty-four 4-point permutations is close to 1/24, which is its expected value in a random permutation. In other words, the set of all twenty-four 4-point permutations is quasirandom-forcing. Moreover, it is known that there exist sets of eight 4-point permutations that are also quasirandom-forcing. Breaking the barrier of linear dependency of perturbation gradients, we show that every quasirandom-forcing set of 4-point permutations must have cardinality at least five.

Let $\mathbb{S}_n$ denote the symmetric group on $[n]=\{1,\ldots,n\}$ with the uniform probability measure. For a permutation $π\in \mathbb{S}_n$ let $X_π$ denote the simplicial complex on the vertex set $[n]$ whose simplices are all $\{i_0,\ldots, i_m\} \subset [n]$ such that $i_0<\cdots<i_m$ and $π(i_0)<\cdots < π(i_m)$. For $r \geq 0$ let $p_r(n)$ denote the probability that $X_π$ is not topologically $r$-connected for $π\in \mathbb{S}_n$. It is shown that for fixed $r \geq 0$ there exist constants $0<C_r, C_r' < \infty$ such that \[ C_r \frac{(\log n)^r}{n} \leq p_r(n) \leq C_r' \frac{(\log n)^{2r}}{n}. \]

It is often convenient to represent a process for randomly generating a graph as a graphon. (More precisely, these give \emph{vertex exchangeable} processes -- those processes in which each vertex is treated the same way.) Other structures can be treated by generalizations like hypergraphons, permutatons, and, for a very general class, theons. These representations are not unique: different representations can lead to the same probability distribution on graphs. This naturally leads to questions (going back at least to Hoover's proof of the Aldous--Hoover Theorem on the existence of such representations) that ask when quasirandomness properties on the distribution guarantee the existence of particularly simple representations. We extend the usual theon representation by adding an additional datum of a random permutation to each tuple, which we call a $\ast$-representation. We show that if a process satisfies the \emph{unique coupling} property UCouple[$\ell$], which says roughly that all $\ell$-tuples of vertices ``look the same'', then the process is $\ast$-$\ell$-independent: there is a $\ast$-representation that does not make use of any random information about $\ell$-tuples (including tuples of length $<\ell$). Simple examples show that the use of $\ast$-representations is necessary. This resolves a question of Coregliano and Razborov, since it easily follows that UCouple[l] implies Independence[\ell'] (the existence of an $\ell'$-independent ordinary representation) for $\ell'<\ell$.

This paper studies permutation statistics that count occurrences of patterns. Their expected values on a product of $t$ permutations chosen randomly from $Γ\subseteq S_{n}$, where $Γ$ is a union of conjugacy classes, are considered. Hultman has described a method for computing such an expected value, denoted $\mathbb{E}_Γ(s,t)$, of a statistic $s$, when $Γ$ is a union of conjugacy classes of $S_{n}$. The only prerequisite is that the mean of $s$ over the conjugacy classes is written as a linear combination of irreducible characters of $S_{n}$. Therefore, the main focus of this article is to express the means of pattern-counting statistics as such linear combinations. A procedure for calculating such expressions for statistics counting occurrences of classical and vincular patterns of length 3 is developed, and is then used to calculate all these expressions. The results can be used to compute $\mathbb{E}_Γ(s,t)$ for all the above statistics, and for all functions on $S_{n}$ that are linear combinations of them.

Chatteerjee and Diaconis have recently shown the asymptotic normality for the joint distribution of the number of descents and inverse descents in a random permutation. A noteworthy point of their results is that the asymptotic variance of the normal distribution is diagonal, which means that the number of descents and inverse descents are asymptotically uncorrelated.The goal of this paper is to go further in this analysis by proving a large deviation principlefor the joint distribution. We shall show that the rate function of the joint distributionis the sum of the rate functions of the marginal distributions, which also means that the number of descents and inverse descents are asymptotically independent at the large deviation level. However,we are going to prove that they are finely dependent at the sharp large deviation level.

By using the matrix formulation of the two-step approach to distributions of patterns in random sequences, recurrence and explicit formulas for the generating functions of successions in random permutations of arbitrary multisets are derived. Explicit formulas for the mean and variance are also obtained.

Continuous-time Mallows processes are processes of random permutations of the set $\{1, \ldots, n\}$ whose marginal at time $t$ is the Mallows distribution with parameter $t$. Recently Corsini showed that there exists a unique Markov Mallows process whose left inversions are independent counting processes. We prove that this process admits a global and a local limit as $n \to \infty$. The global limit, obtained after suitably rescaling space and time, is an explicit stochastic process on $[0,1]$ whose description is based on the permuton limit of the Mallows distribution, analyzed by Starr. The local limit is a process of permutations of $\mathbb{Z}$ which is closely related to the construction of the Mallows distribution on permutations of $\mathbb{Z}$ due to Gnedin and Olshanski. Our results demonstrate an analogy between the asymptotic behavior of Mallows processes and the recently studied limiting properties of random sorting networks.

Graph independence (also known as $ε$-independence or $λ$-independence) is a mixture of classical independence and free independence corresponding to graph products or groups and operator algebras. Using conjugation by certain random permutation matrices, we construct random matrix models for graph independence with amalgamation over the diagonal matrices. This yields a new probabilist,ic proof that graph products of sofic groups are sofic.

Binary search trees (BST) are a popular type of data structure when dealing with ordered data. Indeed, they enable one to access and modify data efficiently, with their height corresponding to the worst retrieval time. From a probabilistic point of view, binary search trees associated with data arriving in a uniform random order are well understood, but less is known when the input is a non-uniform random permutation. We consider here the case where the input comes from i.i.d. random points in the plane with law $μ$, a model which we refer to as a permuton sample. Our results show that the asymptotic proportion of nodes in each subtree depends on the behavior of the measure $μ$ at its left boundary, while the height of the BST has a universal asymptotic behavior for a large family of measures $μ$. Our approach involves a mix of combinatorial and probabilistic tools, namely combinatorial properties of binary search trees, coupling arguments, and deviation estimates.