Searcharxiv⌕ Search

arXiv · 2610.26763

Small subgraphs of large minimum degree

Abstract

In 1991, Erdős, Faudree, Rousseau and Schelp asked to estimate the smallest $u=u(n,d,s)$ such that every $n$-vertex graph with average degree at least $d$ has a subgraph on at most $u$ vertices with minimum degree at least $s$. We resolve this problem up to polylogarithmic factors for all even values of $s$. More precisely, we prove that every $n$-vertex graph with minimum degree $d\geq s$ has a subgraph on $O_s\left(\max\left\{\frac{n}{d^{s/(s-2)}}(\log n)^{180},\log n\right\}\right)$ vertices with minimum degree at least $s$, whenever $s\ge 4$ is even. This improves a result of Janzer, Sudakov and Tomon, who obtained a similar conclusion for subgraphs of average degree at least $s$. We also prove that for every positive integer $s$ and every $\varepsilon>0$, there exists a finite family $\mathcal{H}$ of graphs with minimum degree $s$ such that $\operatorname{ex}(n,\mathcal{H})=O(n^{2-2/s+\varepsilon})$. This is tight and is the first result of similar kind that applies to odd values of $s$ greater than 3.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Zach Hunter, Barnabás Janzer, Oliver Janzer, Benny Sudakov. 2026-10-19. Small subgraphs of large minimum degree. https://arxiv.org/abs/2610.26763

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Notes on the Divisibility of Divisor Sums of Catalan Numbers

We investigate divisibility questions for the sum-of-divisors function $σ(C_n)$ of the Catalan numbers $C_n=(n+1)^{-1}\binom{2n}{n}$. Using the prime factors occurring to exponent one in $C_n$, we give explicit divisors of $σ(C_n)$ and obtain, for each fixed integer $m\geq 2$, a power $m^{R_m(n)}$ dividing $σ(C_n)$, with $R_m(n)\sim\frac{2\log 2}{φ(m)}\frac{n}{\log n}$. We further prove uniform divisibility for all moduli up to fixed powers of $\log n$ and, with the same leading constant $2\log 2$, for almost all moduli up to $\sqrt{n}$ apart from logarithmic factors. We also establish a complete classification of the integers $m\geq 2$ for which $m\midσ(mk-1)$ for every $k\geq 1$, identify precisely the exceptional coprimality relations among the associated linear forms, and record a twin-prime product divisor of $σ(C_n)$. The arguments use standard prime-distribution results and do not assume any conjecture concerning twin primes.

math.CO↗

Sign-coherence and tropical sign pattern for rank $3$ real cluster-cyclic exchange matrices

The sign-coherence about $c$-vectors was conjectured by Fomin-Zelevinsky and proved by Gross-Hacking-Keel-Kontsevich for integer skew-symmetrizable case. We prove this conjecture associated with $c$-vectors for rank 3 real cluster-cyclic skew-symmetrizable case. Simultaneously, we establish their self-contained recursion and monotonicity. Then, these $c$-vectors are proved to be roots of certain quadratic equations. Based on these results, we prove that the corresponding exchange graphs of $C$-pattern and $G$-pattern are $3$-regular trees. We also study the structure of tropical signs and equip the dihedral group $\mathrm{D}_6$ with a cluster realization via certain mutations.

math.CO↗

Subexponential upper bound on the number of rich words

Let $R_q(n)$ denote the number of rich words of length $n$ over a fixed alphabet of size $q\geq2$. It is known that $R_q(n)^{1/n}\to1$. We establish a nonrecursive subexponential upper bound on the number of rich words, with its asymptotic growth quantified in terms of an iterated logarithm. For every $λ\in(1/2,1)$ and $γ>1$, there exists $c>0$ such that, for every positive function $ϕ$ satisfying $ϕ(n)\to\infty$ and $n/ϕ(n)\to\infty$, the functions \[ f(n)=\left[c\ln^{*}\!\left(\frac{n}{ϕ(n)}\ln q\right)\right]^{1/γ}, \qquad B(n)=q^{\frac{n}{ϕ(n)}+\frac{n}{(2λ)^{f(n)-1}}} \] satisfy $R_q(n)\leq B(n)$ for all sufficiently large $n$ and $B(n)^{1/n}\to1$. Here $\ln^{*}$ is the iterated natural logarithm, with value zero on $(0,1]$. The proof iterates a counting inequality based on palindromic factorization and controls the resulting multiplicative constants using the iterated logarithm.

math.CO↗