Skip to content
All apps

Triangular lattice — universal optimality

About this app

Mutually repelling particles at fixed density order themselves into a hexagonal honeycomb — for every completely monotone repulsion (Gaussian, Riesz, screened). Choose potential and range, relax from random or from a square lattice, watch defects and the circle packing; charts show that square, rectangular and rhombic lattices have more energy for every slider value. Cohn–Kumar conjecture in the plane, family 090 from github.com/openai/math (AI-generated preprint, main theorem formalised in Lean).

What it is about

Particles that repel each other should arrange themselves at fixed density so that the energy per particle is as small as possible. For circle packings, Thue and Fejes Tóth showed that the triangular lattice is densest, with density π/√12 ≈ 0.9069. In 2007 Cohn and Kumar conjectured more: the triangular lattice is universally optimal, best for every completely monotone potential f(r²) at once, such as Gaussian, Riesz 1/rˢ or screened Coulomb interactions. In dimensions 8 and 24 this is proved for E₈ and the Leech lattice; in the plane it was open. An AI-generated preprint in github.com/openai/math (family 090) claims a proof; the catalogue lists the main theorem as formalised in Lean, and the preprint is not peer-reviewed.

What the app shows

Particles in a periodic box slide downhill under the chosen repulsion and settle into a honeycomb, shown as Voronoi cells or as a circle packing. Cells with fewer or more than six neighbours are counted as defects. The energy panel compares the current energy per particle with the perfect triangular lattice and plots, across the whole slider range, the excess energy of square, rectangular and rhombic lattices on a log scale. The simulation is finite and periodic, so defects can get stuck.

Try it yourself

Choose Gaussian, Riesz or screened repulsion and set the range, exponent or screening. Start from noise or from a square lattice and watch the honeycomb form; “shake” heats the system briefly so that stuck defects can heal. In the chart, notice that the curves for the other lattices never break off for any slider value: their excess energy stays positive, even when it becomes tiny for soft, long-range repulsion.

Frequently asked questions

What does universal optimality of a lattice mean?
A lattice is universally optimal if, at fixed density, it has the lowest energy per particle for every completely monotone potential of squared distance. This includes Gaussian potentials, Riesz potentials 1/rˢ and screened Coulomb interactions. It is then the best arrangement for a whole family of interactions at once.
Why do repelling particles form hexagons?
In the triangular lattice every particle has six equally distant neighbours, and the distances are spread as evenly as possible for the given density. With repulsion this lowers the energy compared with square or rectangular arrangements. Such honeycomb patterns appear, for example, in vortex lattices in superconductors.
Has the Cohn–Kumar conjecture in the plane been proved?
An AI-generated preprint in the catalogue github.com/openai/math (family 090) claims that the triangular lattice is optimal for every non-negative completely monotone potential against all locally finite configurations. The catalogue lists this statement as formalised in Lean. The preprint has not yet been peer-reviewed.
What can I try out in this app?
You choose a potential and its parameters, start from noise, a square lattice or a triangular lattice, and watch the particles relax. You can drag particles, and “shake” heats things up briefly. The energy panel compares the lattices across the whole slider range.

Subject: Open problems, solved by AI

More from Open problems, solved by AI

A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps