arXiv++ Combinatorics

Given $R \subseteq \mathbb{N}$ let ${n \brace k}_R$, ${n \brack k}_R$, and $L(n,k)_R$ be the number of ways of partitioning the set $[n]$ into $k$ non-empty subsets, cycles and lists, respectively, with each block having cardinality in $R$. We refer to these as the $R$-restricted Stirling numbers of the second and first kind and the $R$-restricted Lah numbers, respectively. Note that the classical Stirling numbers of the second kind and first kind, and Lah numbers are ${n \brace k} = {n \brace k}_{\mathbb{N}}$, ${n \brack k} = {n \brack k}_{\mathbb{N}} $ and $L(n,k) = L(n,k)_{\mathbb{N}}$, respectively. The matrices $[{n \brace k}]_{n,k \geq 1}$, $[{n \brack k}]_{n,k \geq 1}$ and $[L(n,k)]_{n,k \geq 1}$ have inverses $[(-1)^{n-k}{n \brack k}]_{n,k \geq 1}$, $[(-1)^{n-k} {n \brace k}]_{n,k \geq 1}$ and $[(-1)^{n-k} L(n,k)]_{n,k \geq 1}$ respectively. The inverse matrices $[{n \brace k}_R]^{-1}_{n,k \geq 1}$, $[{n \brack k}_R]^{-1}_{n,k \geq 1}$ and $[L(n,k)_R]^{-1}_{n,k \geq 1}$ exist if and only if $1 \in R$. We express each entry of each of these matrices as the difference between the cardinalities of two explicitly defined families of labeled forests. In particular the entries of $[{n \brace k}_{[r]}]^{-1}_{n,k \geq 1}$ have combinatorial interpretations, affirmatively answering a question of Choi, Long, Ng and Smith from 2006. If $1,2 \in R$ and if for all $n \in R$ with $n$ odd and $n \geq 3$, we have $n \pm 1 \in R$, we additionally show that each entry of $[{n \brace k}_R]^{-1}_{n,k \geq 1}$, $[{n \brack k}_R]^{-1}_{n,k \geq 1}$ and $[L(n,k)_R]^{-1}_{n,k \geq 1}$ is up to an explicit sign the cardinality of a single explicitly defined family of labeled forests. Our results also provide combinatorial interpretations of the $k$th Whitney numbers of the first and second kinds of $Π_n^{1,d}$, the poset of partitions of $[n]$ that have each part size congruent to $1$ mod $d$.

For graphs $G$ and $H$, an $H$-coloring of $G$ is an adjacency preserving map from the vertices of $G$ to the vertices of $H$. $H$-colorings generalize such notions as independent sets and proper colorings in graphs. There has been much recent research on the extremal question of finding the graph(s) among a fixed family that maximize or minimize the number of $H$-colorings. In this paper, we prove several results in this area. First, we find a class of graphs ${\mathcal H}$ with the property that for each $H \in {\mathcal H}$, the $n$-vertex tree that minimizes the number of $H$-colorings is the path $P_n$. We then present a new proof of a theorem of Sidorenko, valid for large $n$, that for every $H$ the star $K_{1,n-1}$ is the $n$-vertex tree that maximizes the number of $H$-colorings. Our proof uses a stability technique which we also use to show that for any non-regular $H$ (and certain regular $H$) the complete bipartite graph $K_{2,n-2}$ maximizes the number of $H$-colorings of $n$-vertex $2$-connected graphs. Finally, we show that the cycle $C_n$ maximizes the number of proper colorings of $n$-vertex $2$-connected graphs.

We explain the notion of the {\em entropy} of a discrete random variable, and derive some of its basic properties. We then show through examples how entropy can be useful as a combinatorial enumeration tool. We end with a few open questions.

A distance graph is an undirected graph on the integers where two integers are adjacent if their difference is in a prescribed distance set. The independence ratio of a distance graph $G$ is the maximum density of an independent set in $G$. Lih, Liu, and Zhu [Star extremal circulant graphs, SIAM J. Discrete Math. 12 (1999) 491--499] showed that the independence ratio is equal to the inverse of the fractional chromatic number, thus relating the concept to the well studied question of finding the chromatic number of distance graphs. We prove that the independence ratio of a distance graph is achieved by a periodic set, and we present a framework for discharging arguments to demonstrate upper bounds on the independence ratio. With these tools, we determine the exact independence ratio for several infinite families of distance sets of size three, determine asymptotic values for others, and present several conjectures.

For a simple finite graph G denote by {G \brace k} the number of ways of partitioning the vertex set of G into k non-empty independent sets (that is, into classes that span no edges of G). If E_n is the graph on n vertices with no edges then {E_n \brace k} coincides with {n \brace k}, the ordinary Stirling number of the second kind, and so we refer to {G \brace k} as a graph Stirling number. Harper showed that the sequence of Stirling numbers of the second kind, and thus the graph Stirling sequence of E_n, is asymptotically normal --- essentially, as n grows, the histogram of ({E_n \brace k})_{k \geq 0}, suitably normalized, approaches the density function of the standard normal distribution. In light of Harper's result, it is natural to ask for which sequences (G_n)_{n \geq 0} of graphs is there asymptotic normality of ({G_n \brace k})_{k \geq 0}. Do and Galvin conjectured that if for each n, G_n is acylic and has n vertices, then asymptotic normality occurs, and they gave a proof under the added condition that G_n has no more than o(\sqrt{n/\log n}) components. Here we settle Do and Galvin's conjecture in the affirmative, and significantly extend it, replacing "acyclic" in their conjecture with "co-chromatic with a quasi-threshold graph, and with negligible chromatic number". Our proof combines old work of Navon and recent work of Engbers, Galvin and Hilyard on the normal order problem in a Weyl algebra, and work of Kahn on the matching polynomial of a graph.

Let $w$ be a word in alphabet $\{x,D\}$ with $m$ $x$'s and $n$ $D$'s. Interpreting "$x$" as multiplication by $x$, and "$D$" as differentiation with respect to $x$, the identity $wf(x) = x^{m-n}\sum_k S_w(k) x^k D^k f(x)$, valid for any smooth function $f(x)$, defines a sequence $(S_w(k))_k$, the terms of which we refer to as the {\em Stirling numbers (of the second kind)} of $w$. The nomenclature comes from the fact that when $w=(xD)^n$, we have $S_w(k)={n \brace k}$, the ordinary Stirling number of the second kind. Explicit expressions for, and identities satisfied by, the $S_w(k)$ have been obtained by numerous authors, and combinatorial interpretations have been presented. Here we provide a new combinatorial interpretation that retains the spirit of the familiar interpretation of ${n \brace k}$ as a count of partitions. Specifically, we associate to each $w$ a quasi-threshold graph $G_w$, and we show that $S_w(k)$ enumerates partitions of the vertex set of $G_w$ into classes that do not span an edge of $G_w$. We also discuss some relatives of, and consequences of, our interpretation, including $q$-analogs and bijections between families of labelled forests and sets of restricted partitions.

In the hard-core model on a finite graph we are given a parameter lambda>0, and an independent set I arises with probability proportional to lambda^|I|. On infinite graphs a Gibbs distribution is defined as a suitable limit with the correct conditional probabilities. In the infinite setting we are interested in determining when this limit is unique and when there is phase coexistence, i.e., existence of multiple Gibbs states. On finite graphs we are interested in determining the mixing time of local Markov chains. On Z^2 it is conjectured that these problems are related and that both undergo a phase transition at some critical point lambda_c approx 3.79. For phase coexistence, much of the work to date has focused on the regime of uniqueness, with the best result being recent work of Restrepo et al. showing that there is a unique Gibbs state for all lambda < 2.3882. Here we give the first non-trivial result in the other direction, showing that there are multiple Gibbs states for all lambda > 5.3646. Our proof adds two significant innovations to the standard Peierls argument. First, building on the idea of fault lines introduced by Randall, we construct an event that distinguishes two boundary conditions and always has long contours associated with it, obviating the need to accurately enumerate short contours. Second, we obtain vastly improved bounds on the number of contours by relating them to a new class of self-avoiding walks on an oriented version of Z^2. We extend our characterization of fault lines to show that local Markov chains will mix slowly when lambda > 5.3646 on lattice regions with periodic (toroidal) boundary conditions and when lambda > 7.1031 with non-periodic (free) boundary conditions. The arguments here rely on a careful analysis that relates contours to taxi walks and represent a sevenfold improvement to the previously best known values of λ.

We show that for all sufficiently large d, the uniform proper 3-coloring model (in physics called the 3-state antiferromagnetic Potts model at zero temperature) on Z^d admits multiple maximal-entropy Gibbs measures. This is a consequence of the following combinatorial result: if a proper 3-coloring is chosen uniformly from a box in Z^d, conditioned on color 0 being given to all the vertices on the boundary of the box which are at an odd distance from a fixed vertex v in the box, then the probability that v gets color 0 is exponentially small in d. The proof proceeds through an analysis of a certain type of cutset separating v from the boundary of the box, and builds on techniques developed by Galvin and Kahn in their proof of phase transition in the hard-core model on Z^d. Building further on these techniques, we study local Markov chains for sampling proper 3-colorings of the discrete torus Z^d_n. We show that there is a constant ρ\approx 0.22 such that for all even n \geq 4 and d sufficiently large, if M is a Markov chain on the set of proper 3-colorings of Z^d_n that updates the color of at most ρn^d vertices at each step and whose stationary distribution is uniform, then the mixing time of M (the time taken for M to reach a distribution that is close to uniform, starting from an arbitrary coloring) is essentially exponential in n^{d-1}.

For a graph $G$ and a positive integer $k$, the {\em graphical Stirling number} $S(G,k)$ is the number of partitions of the vertex set of $G$ into $k$ non-empty independent sets. Equivalently it is the number of proper colorings of $G$ that use exactly $k$ colors, with two colorings identified if they differ only on the names of the colors. If $G$ is the empty graph on $n$ vertices then $S(G,k)$ reduces to $S(n,k)$, the familiar Stirling number of the second kind. In this note we first consider Stirling numbers of forests. We show that if $(F^{c(n)}_n)_{n\geq 0}$ is any sequence of forests with $F^{c(n)}_n$ having $n$ vertices and $c(n)=o(\sqrt{n/\log n})$ components, and if $X^{c(n)}_n$ is a random variable that takes value $k$ with probability proportional to $S(F^{c(n)}_n,k)$ (that is, $X^{c(n)}_n$ is the number of classes in a uniformly chosen partition of $F^{c(n)}_n$ into non-empty independent sets), then $X^{c(n)}_n$ is asymptotically normal, meaning that suitably normalized it tends in distribution to the standard normal. This generalizes a seminal result of Harper on the ordinary Stirling numbers. Along the way we give recurrences for calculating the generating functions of the sequences $(S(F^c_n,k))_{k \geq 0}$, show that these functions have all real zeroes, and exhibit three different interlacing patterns between the zeroes of pairs of consecutive generating functions. We next consider Stirling numbers of cycles. We establish asymptotic normality for the number of classes in a uniformly chosen partition of $C_n$ (the cycle on $n$ vertices) into non-empty independent sets. We give a recurrence for calculating the generating function of the sequence $(S(C_n,k))_{k \geq 0}$, and use this to give a direct proof of a log-concavity result that had previously only been arrived at in a very indirect way.

It is shown that the hard-core model on ${\bf Z}^d$ exhibits a phase transition at activities above some function $λ(d)$ which tends to zero as $d\rightarrow \infty$

Write ${\cal F}$ for the set of homomorphisms from $\{0,1\}^d$ to ${\bf Z}$ which send $\underline{0}$ to 0 (think of members of ${\cal F}$ as labellings of $\{0,1\}^d$ in which adjacent strings get labels differing by exactly 1), and ${\cal F}_i$ for those which take on exactly $i$ values. We give asymptotic formulae for $|{\cal F}|$ and $|{\cal F}_i|$. In particular, we show that the probability that a uniformly chosen member ${\bf f}$ of ${\cal F}$ takes more than five values tends to 0 as $d \rightarrow \infty$. This settles a conjecture of J. Kahn. Previously, Kahn had shown that there is a constant $b$ such that ${\bf f}$ a.s. takes at most $b$ values. This in turn verified a conjecture of I. Benjamini {\em et al.}, that for each $t > 0$, ${\bf f}$ a.s. takes at most $td$ values. Determining $|{\cal F}|$ is equivalent both to counting the number of rank functions on the Boolean lattice $2^{[d]}$ (functions $f \colon 2^{[d]} \longrightarrow {\bf N}$ satisfying $f(\emptyset)=0$ and $f(A) \leq f(A \cup x) \leq f(A)+1$ for all $A \in 2^{[d]}$ and $x \in [d]$) and to counting the number of proper 3-colourings of the discrete cube (i.e., the number of homomorphisms from $\{0,1\}^d$ to $K_3$, the complete graph on 3 vertices). Our proof uses the main lemma from Kahn's proof of constant range, together with some combinatorial approximation techniques introduced by A. Sapozhenko.

For given graphs $G$ and $H$, let $|Hom(G,H)|$ denote the set of graph homomorphisms from $G$ to $H$. We show that for any finite, $n$-regular, bipartite graph $G$ and any finite graph $H$ (perhaps with loops), $|Hom(G,H)|$ is maximum when $G$ is a disjoint union of $K_{n,n}$'s. This generalizes a result of J. Kahn on the number of independent sets in a regular bipartite graph. We also give the asymptotics of the logarithm of $|Hom(G,H)|$ in terms of a simply expressed parameter of $H$. We also consider weighted versions of these results which may be viewed as statements about the partition functions of certain models of physical systems with hard constraints.

Let $\gS=(V,E)$ be a finite, $d$-regular bipartite graph. For any $λ>0$ let $π_λ$ be the probability measure on the independent sets of $\gS$ in which the set $I$ is chosen with probability proportional to $λ^{|I|}$ ($π_λ$ is the {\em hard-core measure with activity $λ$ on $\gS$}). We study the Glauber dynamics, or single-site update Markov chain, whose stationary distribution is $π_λ$. We show that when $λ$ is large enough (as a function of $d$ and the expansion of subsets of single-parity of $V$) then the convergence to stationarity is exponentially slow in $|V(\gS)|$. In particular, if $\gS$ is the $d$-dimensional hypercube $\{0,1\}^d$ we show that for values of $λ$ tending to 0 as $d$ grows, the convergence to stationarity is exponentially slow in the volume of the cube. The proof combines a conductance argument with combinatorial enumeration methods.

We study local Markov chains for sampling 3-colorings of the discrete torus $T_{L,d}={0,..., L-1}^d$. We show that there is a constant $ρ\approx .22$ such that for all even $L \geq 4$ and $d$ sufficiently large, certain local Markov chains require exponential time to converge to equilibrium. More precisely, if $\cM$ is a Markov chain on the set of proper 3-colorings of $T_{L,d}$ that updates the color of at most $ρL^d$ vertices at each step and whose stationary distribution is uniform, then the convergence to stationarity of $\cM$ is exponential in $L^{d-1}$. Our proof is based on a conductance argument that builds on sensitive new combinatorial enumeration techniques.

With a graph $G=(V,E)$ we associate a collection of non-negative real weights $\cup_{v\in V}{λ_{i,v}:1\leq i \leq m} \cup \cup_{uv \in E} {λ_{ij,uv}:1\leq i \leq j \leq m}$. We consider the probability distribution on ${f:V\rightarrow{1,...,m}}$ in which each $f$ occurs with probability proportional to $\prod_{v \in V}λ_{f(v),v}\prod_{uv \in E}λ_{f(u)f(v),uv}$. Many well-known statistical physics models, including the Ising model with an external field and the hard-core model with non-uniform activities, can be framed as such a distribution. We obtain an upper bound, independent of $G$, for the partition function (the normalizing constant which turns the assignment of weights on $\{f:V\rightarrow{1,...,m\}}$ into a probability distribution) in the case when $G$ is a regular bipartite graph. This generalizes a bound obtained by Galvin and Tetali who considered the simpler weight collection $\{λ_i:1 \leq i \leq m\} \cup \{λ_{ij}:1 \leq i \leq j \leq m\}$ with each $λ_{ij}$ either 0 or 1 and with each $f$ chosen with probability proportional to $\prod_{v \in V}λ_{f(v)}\prod_{uv \in E}λ_{f(u)f(v)}$. Our main tools are a generalization to list homomorphisms of a result of Galvin and Tetali on graph homomorphisms and a straightforward second-moment computation.

We show that if $\gS=(V,E)$ is a regular bipartite graph for which the expansion of subsets of a single parity of $V$ is reasonably good and which satisfies a certain local condition (that the union of the neighbourhoods of adjacent vertices does not contain too many pairwise non-adjacent vertices), and if $\cM$ is a Markov chain on the set of proper 3-colourings of $\gS$ which updates the colour of at most $ρ|V|$ vertices at each step and whose stationary distribution is uniform, then for $ρ\approx .22$ and $d$ sufficiently large the convergence to stationarity of $\cM$ is (essentially) exponential in $|V|$. In particular, if $\gS$ is the $d$-dimensional hypercube $Q_d$ (the graph on vertex set $\{0,1\}^d$ in which two strings are adjacent if they differ on exactly one coordinate) then the convergence to stationarity of the well-known Glauber (single-site update) dynamics is exponentially slow in $2^d/(\sqrt{d}\log d)$. A combinatorial corollary of our main result is that in a uniform 3-colouring of $Q_d$ there is an exponentially small probability (in $2^d$) that there is a colour $i$ such the proportion of vertices of the even subcube coloured $i$ differs from the proportion of the odd subcube coloured $i$ by at most $.22$. Our proof combines a conductance argument with combinatorial enumeration methods.

Let $i_t(G)$ denote the number of independent sets of size $t$ in a graph $G$. Levit and Mandrescu have conjectured that for all bipartite $G$ the sequence $(i_t(G))_{t \geq 0}$ (the {\em independent set sequence} of $G$) is unimodal. We provide evidence for this conjecture by showing that is true for almost all equibipartite graphs. Specifically, we consider the random equibipartite graph $G(n,n,p)$, and show that for any fixed $p\in(0,1]$ its independent set sequence is almost surely unimodal, and moreover almost surely log-concave except perhaps for a vanishingly small initial segment of the sequence. We obtain similar results for $p=\tildeΩ(n^{-1/2})$. We also consider the problem of estimating $i(G)=\sum_{t \geq 0} i_t(G)$ for $G$ in various families. We give a sharp upper bound on the number of independent sets in an $n$-vertex graph with minimum degree $δ$, for all fixed $δ$ and sufficiently large $n$. Specifically, we show that the maximum is achieved uniquely by $K_{δ, n-δ}$, the complete bipartite graph with $δ$ vertices in one partition class and $n-δ$ in the other. We also present a weighted generalization: for all fixed $x>0$ and $δ>0$, as long as $n=n(x,δ)$ is large enough, if $G$ is a graph on $n$ vertices with minimum degree $δ$ then $\sum_{t \geq 0} i_t(G)x^t \leq \sum_{t \geq 0} i_t(K_{δ, n-δ})x^t$ with equality if and only if $G=K_{δ, n-δ}$.

We use an entropy based method to study two graph maximization problems. We upper bound the number of matchings of fixed size $\ell$ in a $d$-regular graph on $N$ vertices. For $\frac{2\ell}{N}$ bounded away from 0 and 1, the logarithm of the bound we obtain agrees in its leading term with the logarithm of the number of matchings of size $\ell$ in the graph consisting of $\frac{N}{2d}$ disjoint copies of the complete bipartite graph $K_{d,d}$. This provides asymptotic evidence for a conjecture of S. Friedland {\it et al.}. We also obtain an analogous result for independent sets of a fixed size in regular graphs, giving asymptotic evidence for a conjecture of J. Kahn. Our bounds on the number of matchings and independent sets of a fixed size are derived from bounds on the partition function (or generating polynomial) for matchings and independent sets.

At most how many (proper) q-colorings does a regular graph admit? Galvin and Tetali conjectured that among all n-vertex, d-regular graphs with 2d|n, none admits more q-colorings than the disjoint union of n/2d copies of the complete bipartite graph K_{d,d}. In this note we give asymptotic evidence for this conjecture, giving an upper bound on the number of proper q-colorings admitted by an n-vertex, d-regular graph of the form a^n b^{n(1+o(1))/d} (where a and b depend on q and where o(1) goes to 0 as d goes to infinity) that agrees up to the o(1) term with the count of q-colorings of n/2d copies of K_{d,d}. An auxiliary result is an upper bound on the number of colorings of a regular graph in terms of its independence number. For example, we show that for all even q and fixed ε> 0 there is δ=δ(ε,q) such that the number of proper q-colorings admitted by an n-vertex, d-regular graph with no independent set of size n(1-ε)/2 is at most (a-δ)^n.

Galvin showed that for all fixed $δ$ and sufficiently large $n$, the $n$-vertex graph with minimum degree $δ$ that admits the most independent sets is the complete bipartite graph $K_{δ,n-δ}$. He conjectured that except perhaps for some small values of $t$, the same graph yields the maximum count of independent sets of size $t$ for each possible $t$. Evidence for this conjecture was recently provided by Alexander, Cutler, and Mink, who showed that for all triples $(n,δ, t)$ with $t\geq 3$, no $n$-vertex {\em bipartite} graph with minimum degree $δ$ admits more independent sets of size $t$ than $K_{δ,n-δ}$. Here we make further progress. We show that for all triples $(n,δ,t)$ with $δ\leq 3$ and $t\geq 3$, no $n$-vertex graph with minimum degree $δ$ admits more independent sets of size $t$ than $K_{δ,n-δ}$, and we obtain the same conclusion for $δ> 3$ and $t \geq 2δ+1$. Our proofs lead us naturally to the study of an interesting family of critical graphs, namely those of minimum degree $δ$ whose minimum degree drops on deletion of an edge or a vertex.