🔢 mathematics

An algebraic approach to circle packing

This paper demonstrates that circle packings realizing certain surface triangulations can be determined by solving a symmetric system of polynomial equations associated with triangle corners, which generalizes the Descartes circle theorem and bridges the authors' previous spinorial approach with classical Euclidean geometry.

Original authors: Daniel V. Mathews, Orion Zymaris

Published 2026-08-05

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel V. Mathews, Orion Zymaris

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Geometry of Sticker Patches

Imagine you are an architect trying to build a perfect model of a curved surface, like a soccer ball or a donut, using only flat, triangular pieces of paper. You want to glue these triangles together so that they fit perfectly, but there's a catch: you want to place a circle in the middle of every corner where the triangles meet, and these circles must touch their neighbors exactly, like a tightly packed mosaic of bubbles. This is the world of circle packing, a branch of mathematics that sits at the intersection of geometry (the study of shapes and spaces) and topology (the study of how things are connected).

For a long time, mathematicians have known that these circle patterns are incredibly powerful. They act like a secret code that can describe the shape of a surface, whether it's a flat sheet, a sphere, or a torus (a donut shape). The challenge has always been figuring out exactly how big each circle needs to be to make the whole puzzle fit together without gaps or overlaps. Usually, solving this requires complex, messy calculations that are hard to crack. But what if there was a simpler way? What if the rules for fitting these circles together could be written down as a neat set of algebraic equations, like a recipe for baking a cake? This is the question that drives the research in the paper you are about to read.

The Algebraic Recipe for Circle Puzzles

In their paper, "An Algebraic Approach to Circle Packing," Daniel V. Mathews and Orion Zymaris have discovered a surprisingly elegant way to solve these circle puzzles. They show that for many different shapes, you don't need to guess and check to find the right circle sizes. Instead, you can write down a specific system of polynomial equations (mathematical sentences involving variables raised to powers, like $x^2$ or $xy$) that, when solved, give you the exact sizes and positions of the circles.

Think of the surface you want to build as a giant jigsaw puzzle made of triangles. In this paper, the authors treat every single "corner" of every triangle as a variable in a giant math equation. They call these variables $m$ . The authors prove that if you can find a set of positive numbers for these $m$ variables that satisfy their specific rules, you have automatically found a valid circle packing. It's like having a magic key: if the numbers fit the lock (the equations), the door opens to reveal a perfect geometric pattern.

The paper breaks these rules down into three main types of "glue" that hold the puzzle together:

Triangle Equations: These ensure that the three circles inside a single triangle fit together correctly, like three friends holding hands in a circle.
Edge Equations: These make sure that when two triangles share a side, the circles on that shared edge agree on how big they should be relative to each other.
Vertex Equations: These are the most complex. They ensure that when you look at a point where many triangles meet (a vertex), all the circles around it fit together in a full circle without leaving a gap or overlapping too much.

The authors show that for simple shapes like a flat disk, these equations are unique and have no extra, redundant rules. However, for more complex shapes like a sphere (a ball) or a torus (a donut), the rules depend on a few extra choices you make, like picking a specific triangle to be the "north pole" or choosing a path to walk around the donut. Even with these choices, the math holds up.

One of the most exciting parts of their discovery is a new, more symmetrical version of a famous old rule called the Descartes Circle Theorem. You might know the classic version, which relates the sizes of four mutually touching circles. The authors have generalized this to work for any number of circles arranged in a flower-like pattern around a center. Their new formula is more balanced and symmetrical than previous attempts, making it easier to work with. They call these variables $m$ "spinorial" because they are deeply connected to a concept in physics and advanced math called spinors, which are essentially "square roots" of geometry. In simple terms, these numbers capture the essence of the angles in the triangles in a way that makes the whole system much easier to solve.

The paper also tackles the tricky issue of "branching." Sometimes, when you pack circles around a point, they might wrap around more than once, like a spiral staircase. The authors provide a special set of "unbranched" equations that ensure the circles wrap around exactly once, creating a smooth surface without any twists or folds. They prove that if you solve these equations, you are guaranteed to get a valid circle packing, and conversely, every valid circle packing corresponds to a solution of these equations.

What makes this approach so powerful is that it turns a geometric problem (drawing circles) into an algebraic one (solving equations). This means mathematicians can use powerful computer tools to find solutions that would be impossible to draw by hand. The authors demonstrate this with examples, showing how to calculate the exact sizes of circles for a tetrahedron (a pyramid with a triangular base) or a standard torus. In the case of the torus, they show that the equations force the circles to form a perfect, repeating hexagonal pattern, just like the cells in a honeycomb.

Ultimately, Mathews and Zymaris have provided a new toolkit for understanding the geometry of surfaces. By showing that circle packings are the solutions to a specific set of polynomial equations, they have opened the door to using algebra to solve geometric problems. Whether you are a mathematician trying to understand the shape of the universe or a curious teenager wondering how to fit circles together perfectly, this paper suggests that the answer lies in the elegant, symmetrical language of algebra. The results are presented as rigorous proofs, meaning the connection between the equations and the circle packings is mathematically certain, not just a guess or a simulation. This work doesn't just find a new way to draw circles; it reveals that the rules of geometry are, at their heart, a beautiful and solvable algebraic puzzle.

Technical Summary: An Algebraic Approach to Circle Packing

Problem Statement
The paper addresses the problem of finding circle packings that realize specific triangulations of surfaces (specifically $\Delta$ -complexes $K$ ). While circle packing theory is well-established, often relying on variational or transcendental methods, this work seeks an algebraic formulation. The goal is to determine whether circle packings realizing a given triangulation $K$ on Euclidean, spherical, or toroidal surfaces can be found by solving a system of polynomial equations. The authors focus on cases where $K$ triangulates a simply connected plane surface, a sphere, or a torus.

Methodology
The authors introduce a system of variables and polynomial equations, termed "circle packing equations," associated with the corners of triangles in the complex $K$ .

Variables: For each corner $(\Delta, v)$ of a triangle $\Delta$ with vertex $v$ , a variable $m_{\Delta, v}$ is defined. Geometrically, in a realization $\Phi$ , these variables correspond to the cotangent of half the angle at that corner: $m_{\Delta, v} = \cot(\theta_{\Delta, v}/2)$ .
Equations: The system comprises three primary types of equations, plus specific conditions for non-planar topologies:
1. Triangle Equations: For each face $\Delta$ with corners $u, v, w$ , the variables satisfy $m_u m_v m_w = m_u + m_v + m_w$ . This encodes the condition that the angles of a Euclidean triangle sum to $\pi$ .
2. Edge Equations: For each interior edge $e$ shared by triangles $\Delta_1, \Delta_2$ with endpoints $v_1, v_2$ , the variables satisfy $m_{\Delta_1, v_1} m_{\Delta_2, v_2} = m_{\Delta_1, v_2} m_{\Delta_2, v_1}$ . This ensures consistency in the ratio of circle curvatures across the edge.
3. Vertex Equations: For each interior vertex $v$ of degree $d$ , the variables satisfy a generalized Descartes circle theorem equation. This is a polynomial equation of degree $d-1$ derived from the condition that the sum of angles around $v$ is a multiple of $2\pi$ .
4. Holonomy Equations (Torus only): For tori, two additional equations are imposed corresponding to generators of the first homology group. These ensure that the "rotation" accumulated along closed loops is trivial, allowing the local Euclidean structure to glue into a global torus.
5. Sphere-closing Equations (Sphere only): For spheres, a specific triangle is designated (e.g., the "north pole"), and its corner variables are fixed to constants ( $-\sqrt{3}$ ) to enforce the spherical geometry via stereographic projection.

The authors also define "unbranched" versions of these equations, where the angle sum conditions are restricted to exactly $2\pi$ (for vertices) or $0$ (for holonomy), rather than integer multiples.

Key Contributions and Results

Symmetric Generalised Descartes Circle Theorem (Theorem 1.2): The authors present a new generalization of the Descartes circle theorem for an $n$ -flower (a central circle surrounded by $n$ tangent petal circles). They derive a symmetric polynomial equation relating the variables $m_j$ :
$\prod_{j=1}^n (m_j + i) = \prod_{j=1}^n (m_j - i)$
This equation is of degree $n-1$ and is more symmetric than previous formulations. It is shown to be equivalent to the vanishing of the imaginary part of the product, yielding a real polynomial equation.
Bijective Correspondence (Theorem 1.1): The central result establishes a bijection between the positive real solutions of the circle packing equations for $K$ and the conformal classes of Euclidean (or good spherical) circle packings realizing $K$ .
- For simply connected plane surfaces, the solution set is a real semialgebraic set.
- For spheres and tori, the equations depend on auxiliary choices (a distinguished triangle or homology basis), but any such choice yields a system whose positive solutions correspond to the conformal classes of packings.
- The paper proves that solutions to these equations can be constructively used to build the circle packing, proving the converse direction of the correspondence.
Unbranched Correspondence (Theorem 1.20): A similar bijection is established for unbranched circle packings using the unbranched vertex and holonomy equations.
Redundancy and Rigidity: The authors analyze the redundancy in the equation systems. For spheres and tori, they define "reduced" systems of equations (removing specific edge and vertex equations) that are equivalent to the full systems but have a variable-to-equation count that reflects the rigidity of the solution space (dimension 0 for spheres and tori with fixed boundary/branching data).
Connection to Spinors and Geometry: The work connects the algebraic variables $m$ to the spinorial approach of previous work (Mathews [13]). The variables $m$ are identified as the real parts of products of spinors and are shown to be cotangents of half-angles. The equations are described as "spinorial square roots of geometry," linking complex number geometry (arguments of $m+i$ ) to the angle sum conditions.

Significance and Claims
The paper claims to provide a clean, algebraic framework for circle packing problems, analogous to Thurston's gluing and completeness equations for hyperbolic 3-manifolds.

Algebraic Nature: Unlike variational approaches that minimize energy functionals, this approach reduces the existence and classification of circle packings to solving systems of polynomial equations.
Generality: The method applies to finite and infinite triangulations of discs, spheres, and tori.
Geometric Interpretation: The authors explicitly link the algebraic variables to classical Euclidean geometry (cotangents of half-angles) and Soddy circles, providing a transparent geometric meaning to the polynomial constraints.
Constructive Proof: The proof of the main theorems is constructive, demonstrating how to build the circle packing from a solution to the equations, thereby confirming that the algebraic solutions are not merely formal but correspond to actual geometric realizations.

The authors note that while the equations for spheres and tori are not unique (depending on choices of auxiliary structures), the resulting solution sets consistently map to the correct conformal classes of circle packings. The work is presented as a development of the spinorial approach, offering a more symmetric and algebraic perspective on the Descartes circle theorem and circle packing realizations.

The Geometry of Sticker Patches

The Algebraic Recipe for Circle Puzzles

More like this