Hexagon is a repository for the preservation, dissemination, study, and evaluation of research-level works in the mathematical sciences and theoretical computer science, including both original and expository texts. Hexagon does not perform peer review. Every deposit is screened automatically and, during the beta, also reviewed by a human moderator before it appears; acceptance is not a judgment of correctness, quality, or novelty.

Browse by subject

Newest works

To replace this with a different default view, perform an advanced search for the parameters you want, then select “Set this search as my default view.”

math.CO — Combinatorics

A strengthening of some results of the paper "Structured sunflowers and canonical Ramsey properties"

Contributed by Rob Sullivan

In this short note, we give an exposition of an argument (potentially LLM-generated by Budiyanto) which strengthens two theorems of the author and Winkel from the recent paper “Structured sunflowers and canonical Ramsey properties”, also proving the main conjecture of that paper.

Primarily human-written textHuman understanding: all parts
Nov 19, 20263 pages
PDFView work

math.NT — Number Theory

Exact Sumsets of Primitive Quadratic Norms

Contributed by Roy van Rijn

We determine the exact binary sumsets of primitive quadratic norm values at discriminants −3,−7,−8,−11-3,-7,-8,-11. The local congruence conditions are sufficient in three cases; at discriminant −7-7 the complete additional exception set is {7,77,203}\{7,77,203\}. Stronger split-prime core identities separate the ramified dilations. The proofs combine quaternionic norm shells, precisely normalized classical Hecke bounds, and dyadic and ramified local filters. Explicit infinite estimates, rational interval certificates, complete finite sumset scans and multiplicative descent cover every positive target. We supply standalone verification code and explain the finite reductions that make the certificates complete. All sumsets concern numerical values, without coordinate multiplicities.

Primarily AI-generated textHuman understanding: no partsEisenstein integersHecke operatorsbinary quadratic formsexact sumsetsfinite certificatesprimitive quadratic norms
Nov 19, 202634 pages
PDFView work

math.NT — Number Theory

Pointwise Correlations of Quadratic Form Values and Splitting Semigroups

Contributed by Roy van Rijn

Under the stated analytic inputs, we establish the fixed-shift pair asymptotics predicted by Smilansky for the ordinary values of x2+dy2x^2+dy^2, d=1,2,3,4,7d=1,2,3,4,7. In particular, the number of n≤Xn\le X for which both nn and n+1n+1 are sums of two squares is asymptotic to X/(2log⁡X)X/(2\log X). For primitive positive definite binary forms of fixed discriminants Δi=fi2D\Delta_i=f_i^2D sharing one negative fundamental discriminant DD, we count values admitting coprime coordinates. Replacing either form within its respective genus preserves the leading binary count, with a relative logarithmic saving through an independent classical route. Under the recent analytic inputs we give an explicit mixed leading term with relative o(1)o(1), including different fixed conductors in that one field, and no class-number divisor. The central mechanism is a shared comparison between sparse indicators and arithmetic models retaining every surviving congruence character. Following Walsh's proposed indicator extension, we supply the models, local factors and a uniform reflected-target argument, using the recent zero-free and weighted-dilation theorems of OpenAI. Applications to fixed subgroup restrictions and prime-coset self-sums give target-uniform asymptotics with logarithmic savings and exact eventual congruence criteria. They include prime support 1(mod12)1\pmod{12} on targets 2(mod12)2\pmod{12} and prime support 3(mod4)3\pmod4 on even targets. All counts concern numerical values without coordinate or ideal multiplicities.

Primarily AI-generated textHuman understanding: no partsarithmetic modelsbinary quadratic formsgenus invariancepointwise correlationssplitting semigroupssums of two squares
Nov 19, 202669 pages
PDFView work

math.SP — Spectral Theory

Maximizing the First Positive Neumann Eigenvalue of Simple Pentagons

Contributed by Lüwei Dong, Hongyi Wang

We prove that among all simple pentagons of prescribed area—including non-convex pentagons and pentagons with straight interior angles—the regular pentagon uniquely maximizes the first positive Neumann eigenvalue of the Laplacian up to congruence. The proof separates naturally into two regimes. For pentagons having at least one reflex or straight interior angle, the number of strictly convex vertices is at most four; conformal mapping from the unit disk together with a Schwarz–Christoffel and Blaschke-product estimate yields the uniform bound ∣P∣μ1(P)<10.34|P|\mu_1(P) < 10.34, strictly below the regular-pentagon value ∣R∣μ1(R)>10.35335|R|\mu_1(R) > 10.35335. For strictly convex pentagons, normalizing a maximum-area vertex triangle reduces the problem to a four-dimensional shape chart coupled with a unimodular metric matrix. Combining covariance bounds, explicit polynomial trial functions, conic duality for pairs of trial functions, and Bernstein-basis branch-and-bound, we confine every potential maximizer to an explicit neighborhood of the regular pentagon in vertex Fourier coordinates. Because the first positive eigenspace of the regular pentagon is two-dimensional, we analyze both eigenfunctions simultaneously together with their first-order corrections to show that every nontrivial local deformation of the similarity class strictly decreases ∣P∣μ1(P)|P|\mu_1(P). All finite-dimensional inequalities invoked in the proof are accompanied by independently verifiable rational and interval arithmetic certificates.

Author comments: 24 pages. The ancillary files include the numerical certificates and verification programs. Reproduction instructions for this submission are in anc/verification.json.

A mix of human-written and AI-generated textHuman understanding: all parts
Nov 19, 202624 pages
PDFView work

math.CO — Combinatorics

Sharp packing bounds for Poly-5 and Handy-Octs

Contributed by William Howe, Grant Huffman, Jensen Kohlmeyer, Liam Kruer

On the 89-square Poly-5 board, distinct pentominoes can cover at most 26 of the 28 boundary squares and 56 of the 60 exterior unit edges. Both maxima occur in a full tiling. For the five Handy-Octs octiamonds, the least area of a convex triangular-grid enclosure is 43 unit triangles. There is one optimal enclosure and it has precisely five packings, up to lattice symmetries. We prove the bounds by finite enumeration.

Primarily AI-generated textHuman understanding: some partsexact coveroctiamondspentominoestilings
Nov 19, 20264 pages
PDFView work

math.MG — Metric Geometry

Gradient structures on metric measure spaces

Contributed by Fabrice Baudoin

This work develops a theory of finite-rank gradients on metric measure spaces where the reference measure and the measure used to integrate gradients may be mutually singular. Examples include local trees, RCD spaces, sub-Riemannian manifolds, the harmonic Sierpiński gasket, and the Liouville Brownian motion.

Primarily AI-generated textHuman understanding: all parts
Nov 19, 2026200 pages
PDFView work

math.AG — Algebraic Geometry

Schubert Cox Rings and Root Operators

Contributed by Jiarui Fei

We identify the Cox ring of every Schubert variety in a symmetrizable Kac--Moody partial flag variety with an upper cluster algebra whose frozen variables are not inverted. The construction also applies to Bott--Samelson varieties for arbitrary words and to their normal projective birational contractions that preserve the standard affine chart. A line-bundle recursion determines the boundary coefficients of the exchange relations. In this mutation class, each nonzero positive simple-root operator and each nonzero stabilizing negative simple-root operator acts in a suitable seed as a monomial times a partial derivative in one frozen variable. Consequently, a single theta basis describes the root filtrations on all divisorial section spaces. We compute the facets of its regular cone from the word and the surviving divisors. The leading terms of the root operators yield tropical formulas for finite-type Levi crystals and polyhedral formulas for branching multiplicities. We obtain formulas for adjoint sections and extend the adapted root charts to projective horospherical varieties with free class group under a simply connected semisimple group.

Author comments: 39 pages, comments are welcome

A mix of human-written and AI-generated textHuman understanding: some parts
Nov 19, 202639 pages
PDFView work

math.NT — Number Theory

Non-existence of non-Eulerian irregular Diophantine quadruples

Contributed by Andrej Dujella, Sinisa Slijepcevic

A Diophantine quadruple {a,b,c,d}\{a,b,c,d\} with a<b<c<da<b<c<d is irregular if d≠a+b+c+2abc+2(ab+1)(ac+1)(bc+1)d\ne a+b+c+2abc+2\sqrt{(ab+1)(ac+1)(bc+1)}. It is conjectured that there are none, and He and Togb\'e have recently given a proof that there are only finitely many, with c<10338c<10^{338}. We study the non-Eulerian irregular quadruples, those whose triple {a,b,c}\{a,b,c\} is not an Euler triple, that is, c≠a+b+2ab+1c\ne a+b+2\sqrt{ab+1}, and reduce them to nine explicit families. The families are defined by the degree of {a,b,c}\{a,b,c\} under canonical descent, by the position of the descent partner d−(a,b,c)d_-(a,b,c) relative to aa and bb, by the shape of the iterated descent, and by explicit windows for the index of z=cd+1z=\sqrt{cd+1} in two Pell sequences. In particular c<1099c<10^{99}, and c<1079.71c<10^{79.71} when zz has an odd-index representation; for triples of degree one, c<1.81⋅1047c<1.81\cdot10^{47}. He and Togb\'e obtained c<10162c<10^{162} for degree one and c<10338c<10^{338} for higher degree. With the bound c<4⋅1048c<4\cdot10^{48} for Euler triples of a companion paper, every irregular Diophantine quadruple has c<1099c<10^{99}. For each family we give exact finite tests, in integer arithmetic or with certified rational enclosures, whose success shows that a given triple of the family has no irregular extension. Running these tests over the families is a finite computation. The numerical verification of these remaining cases is in progress, and its results will be appended to the final version of the paper.

Primarily AI-generated textHuman understanding: some parts
Nov 19, 2026203 pages
PDFView work

cs.DS — Data Structures and Algorithms

Exact minimum-cost arborescences under a single arc failure

Contributed by Yilin Liu

Every rooted weighted digraph on n vertices has a subgraph with O(n3/2)O(n^{3/2}) arcs that contains a minimum-cost spanning arborescence after any feasible single arc deletion, answering a question of Dey and Kavitha. An oracle of the same size reports an optimal arborescence or a shortest list of arc replacements in time linear in the output. The construction uses shortest cycles in a weighted matroid exchange graph and admits deterministic subcubic preprocessing on simple digraphs. A reduction from directed distance preservation gives an Ω(n4/3log⁡n)\Omega(n^{4/3} \log n)-bit lower bound for reporting on simple digraphs with input arc identifiers, even with constant per-query error.

Primarily AI-generated textHuman understanding: no partsdistance preserversfault-tolerant sparsificationminimum-cost arborescencessingle-arc failuresweighted matroid intersection
Nov 19, 20268 pages
PDFView work

Browse all works →