SearcharxivSearch

arXiv subjects

Lingsen Meng

Publications and source records attributed to Lingsen Meng.

2 recordsLinked to original sources

Residual structure and growing inversion-monotonicity regions for 1324-avoiding permutations

Let $a(n,k)$ be the number of $1324$-avoiding permutations of length $n$ with $k$ inversions. Linusson and Verkama proved $a(n,k)\le a(n+1,k)$ for $k\le2n-7$. We study the obstruction beyond that line: the residuals $\mathcal R_{\delta,n}$, namely the indecomposable, non-almost-decomposable avoiders at defect $\delta=k-2n+7$. Contracting maximal increasing consecutive runs reduces residuality to a quadratic equation on a finite family of skeletons. It follows that, for every fixed $\delta$, the eventual count has the form $|\mathcal R_{\delta,n}|=A_\delta n^2+B_\delta n+C_\delta$. Our central uniform result determines the quadratic coefficient at every defect: with $P(q)=\prod_{j\ge1}(1-q^j)^{-1}$, $\sum_{\delta\ge0}A_\delta q^\delta=4q^3(1+q)P(q)^2/(1-q)^2$. This is a formula for the leading coefficient of the residual count, not for the full count. The same structural estimates give computer-assisted proofs of $a(n,k)\le a(n+1,k)$ for every $n\ge1$ and $k\le2n+6$, and of regions whose width grows with $n$: for $n\ge2^{16},2^{18},2^{20}$ the defect may be as large as $\lfloor\sqrt n/4\rfloor$, $\lfloor\sqrt n/3\rfloor$, $\lfloor\sqrt n/2\rfloor$, respectively. More generally, every fixed $c<\sqrt2\log(5)/\log(68)$ is admissible for all sufficiently large $n$. The three added fixed defects $11,12,13$ use complete catalogue and rational-sum certificates supplied in the accompanying archival supplement. The leading-coefficient theorem is obtained from a complete finite classification of marked rank-three cores and all-parameter extension lemmas. We also determine the exact rank-three stabilization onset, while keeping it separate from the still unknown onset of the complete residual count. The unrestricted Claesson--Jel'inek--Steingr'imsson conjecture, the full residual polynomials, and the sharp global base-length bound remain open.

math.CO

Trees of odd order with at most two vertices of degree two are edge-graceful

A graph G with q edges and p vertices is edge-graceful if some bijection f from E(G) onto {1,...,q} makes the induced vertex sums f^+(v), the sum of f(e) over the edges e incident to v, distinct modulo p. Lee conjectured in 1989 that every tree of odd order is edge-graceful; the broadest general result we have located, due in equivalent form to Kaplan, Lev and Roditty, covers trees of odd order with at most one vertex of degree two. We prove that every tree of odd order with at most two vertices of degree two is edge-graceful. The proof combines zero-sum block partitions of Z_n with perfect and hooked Langford sequences; the residual case analysis is verified symbolically for all odd n <= 5001 and holds uniformly beyond, and the construction was executed and independently re-checked on all 2,245,070 trees of odd order at most 25 with exactly two vertices of degree two. Since an edge-graceful graph is antimagic, the theorem also enlarges the family of trees known to be antimagic.

math.CO