Skip to content

Documentary state: DRAFT_UNACCEPTED. This draft has NOT been editorially admitted. Mathematical claims require independent checking.

OAI-26 family 092: optimal order of convex-body covering density

Editorial status: DRAFT_UNACCEPTED. NO_CERTIFICATION. This article reports claims from two pinned manuscripts without independently verifying their proofs.

The question in plain language

A convex body is a solid shape with no dents: whenever it contains two points, it also contains the line segment between them. Copies of such a body can cover all of space by translation, but the copies may overlap. Covering density measures the average amount of covering material used per unit volume.

One can restrict the translations to a lattice—a regular repeated grid—or allow arbitrary placements. The pinned OAI-26 overview claims that in high dimension, the worst-case optimal density grows on the order of the dimension times its logarithm, under either convention.

A small example

Imagine covering the plane with congruent circles. Because circles cannot tile the plane perfectly without gaps, a covering must overlap. The total circle area per unit area of the plane exceeds one. For a convex body in dimension \(n\), covering density generalizes this measure.

A lattice covering requires translates centered on a regular additive grid; an unrestricted translative covering can use irregular placements. The latter has more freedom, so its minimum density cannot be worse than the best lattice density for the same body.

The asserted high-dimensional bound is therefore significant in two directions: a lattice construction gives a universal upper bound, while some bodies demand high density even for unrestricted arrangements.

What is known

The protected Family 092 catalogue identifies two manuscripts, both dated September 23, 2026: A single-lattice covering bound of order n log n and Translative covering densities of order n log n.

The pinned publisher overview claims that every convex body in real dimension \(n\ge2\) has a lattice covering of density at most \(C n\log n\) for a universal constant \(C\).

For the reverse direction, it claims centrally symmetric convex bodies in every sufficiently large dimension for which every translative covering, whether lattice-based or not, has density at least \(c n\log n\) for a universal positive constant \(c\).

Together, if verified, these prove the worst-case order \(\Theta(n\log n)\) for lattice and unrestricted translative coverings. The lower bound applies to selected difficult bodies, not to every convex body. No numerical optimal constants or equality cases are asserted in the indexed overview.

GCL verification state

Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. The protected catalogue identifies both PDFs, and the pinned overview supplies the upper and lower covering-density claims.

This documentary pass does not independently check the lattice construction, asymptotic density definitions, or the centrally symmetric lower-bound examples. No Family 092-specific Lean proof, Comparator acceptance, source-to-formal semantic audit, or MATHCERT certificate is established.

What remains

Read both pinned manuscripts and locate their exact principal theorems, definitions of convex bodies and covering density, and treatment of the lattice determinant or unrestricted translation set. Check the universal upper bound in every stated dimension and the existence of the claimed single lattice.

For the lower bound, inspect the centrally symmetric bodies, the quantification over arbitrary translative coverings, the sufficiently-large-dimension threshold and the absolute constants. Confirm that the two estimates apply to compatible density conventions.

Complete mathematical Markdown and rendered-compendium validation, source-specific editorial criticism and protected merge/readback. Documentary acceptance is not proof certification.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. These are publisher-indexed claims, not independently verified or certified theorems.


Indexed primary manuscript paths: