🔢 mathematics

Hilbert's 16th problem for arrangements of curves on a surface

This paper introduces a combinatorial framework using intersection numbers, Dyck words, and rooted trees to classify arrangements of curves on real surfaces, successfully applying it to achieve a complete classification of three lines and a cubic, and a partial classification of three lines and a quartic, as a generalization of Hilbert's 16th problem.

Original authors: Giacomo Maletto

Published 2026-06-23

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Giacomo Maletto

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 Big Picture: A Topological Puzzle

Imagine you are an artist working on a giant, flat, round canvas (which mathematicians call the real projective plane, or $\mathbb{P}^2(\mathbb{R})$ ). You have a set of rules:

You draw three straight lines across the canvas.
You then draw a curvy shape (like a circle, a figure-eight, or a more complex blob) that crosses those lines but never touches them at the same spot twice, and never has any "kinks" or sharp points.

The question the paper asks is: How many different ways can these lines and the curvy shape be arranged relative to each other?

In math, "different" means you can't stretch, squish, or twist the canvas to turn one arrangement into another without cutting or gluing. This is the modern version of a famous 100-year-old puzzle called Hilbert's 16th Problem, which originally asked how to classify these shapes when there is only one curve. This paper expands that puzzle to look at a whole family of curves (lines + one big curve) at the same time.

The Solution: Turning Shapes into Lego Instructions

The author's main breakthrough is a new way to describe these messy, curvy arrangements using a simple, rigid code. Think of it like turning a complex 3D sculpture into a set of Lego instructions.

Instead of trying to draw the curve, the author breaks the canvas down into a grid (like a map) created by the three lines. Then, he describes the curvy shape using three specific pieces of data, which he calls a "Combinatorial Curve":

The Count ( $n$ ): How many times does the curvy shape cross each line segment of the grid? (e.g., "It crosses the left line 2 times, the right line 0 times...").
The Dance Steps ( $W$ ): As you walk around the edge of a specific section of the canvas, does the curve enter and leave? The author uses a special code called a Dyck word (a sequence of opening ( and closing ) parentheses) to record this.
- Analogy: Imagine walking around a room. Every time you see a new piece of the curve enter the room, you write "(". Every time you see it leave, you write ")". This ensures the curve doesn't get "stuck" or cross itself in impossible ways.
The Nesting Map ( $T$ ): Sometimes the curve makes little loops (ovals) that float in the empty spaces between the lines without touching them. The author uses a Rooted Tree to map these.
- Analogy: Think of a family tree. The "root" is the big empty space. If there is a small loop inside that space, it's a child. If there is a loop inside that loop, it's a grandchild. This tree tells you exactly how the loops are nested inside each other.

The Golden Rule (Theorem A): The paper proves that if you have this code ( $n, W, T$ ), you know the exact shape of the arrangement. No two different shapes share the same code, and no shape is left out.

The Computer Lab: The "NWT" Library

Because there are thousands of possible codes, the author didn't just do this on paper. He built a software library called NWT (written in the Julia programming language).

Think of this library as a massive digital filing cabinet.
It can generate millions of these "Lego instructions."
It can check if a specific instruction set is "legal" (mathematically possible) or "illegal" (impossible to draw).
It can even draw the picture of the curve based on the code.

The Detective Work: Finding the Real Ones

Just because a code exists doesn't mean a real algebraic curve (a curve defined by a specific math equation) can actually make that shape. The author had to filter the list of all possible codes to find the ones that are "realizable."

He used two main detective tools:

The "Bézout" Rule: This is a mathematical law of conservation. It says that if you have a line and a curve of a certain complexity, they can only cross each other a specific maximum number of times. If a code says they cross 10 times but the math says the maximum is 4, that code is a fake. The author used this to throw out thousands of impossible codes.
Viro's Patchworking: This is a construction technique. Imagine building a complex sculpture by gluing together small, simple pieces. The author used this method to prove that many of the codes can actually be built.

The Results: What Did They Find?

The author applied this system to two specific puzzles:

Three Lines + A Cubic Curve (Degree 3):
- Result: He found exactly 119 unique ways these can be arranged.
- Bonus: He found 3 specific arrangements that were "missing" from previous construction methods (Viro's patchworking) but proved they still exist by finding the actual math equations for them.
Three Lines + A Quartic Curve (Degree 4):
- This is much more complex.
- Result: He completely classified the cases where the curve has exactly one loop (oval). There are 619 of these.
- The "Floatless" Cases: He also looked at cases where the curve doesn't have any "floating" loops (loops that don't touch the lines). He narrowed this down to a range of 1,834 to 1,883 possibilities.
- The Unknown: For the general case (all possible quartic curves), he couldn't finish the list. He knows the answer is somewhere between 9,426 and 17,624, but the list is too huge to check every single one yet.

Summary

In short, this paper takes a very hard geometry problem—figuring out all the ways lines and curves can twist around each other—and solves it by turning the shapes into a computer-readable code.

The Problem: How many ways can you arrange 3 lines and 1 curvy shape?
The Tool: A code made of numbers, parentheses, and trees.
The Outcome: A complete list for simpler shapes (cubics) and a massive, partially complete list for more complex shapes (quartics), all verified by a custom-built computer program.

The paper doesn't claim to solve medical problems or build bridges; it is a pure math achievement that organizes the "zoo" of possible curve shapes, proving exactly which ones exist and which ones are impossible.

Technical Summary: Hilbert's 16th Problem for Arrangements of Curves on a Surface

Problem Statement
This work addresses a generalization of Hilbert's 16th problem to arrangements of algebraic curves on the real projective plane $\mathbb{P}^2(\mathbb{R})$ . Specifically, it investigates the topological classification of $k$ -arrangements $(C_1, \dots, C_k)$ , defined as $k$ -tuples of curves where all pairs meet transversely and no three share a common point. The primary focus is on determining which topological types of such arrangements are $(d_1, \dots, d_k)$ -realizable, meaning they can be represented by nonsingular algebraic curves of specified degrees $d_i$ . The paper concentrates on the specific cases of arrangements involving three lines and a cubic curve ( $(1,1,1,3)$ -realizability) and three lines and a quartic curve ( $(1,1,1,4)$ -realizability).

Methodology
The paper establishes a bridge between the geometric topology of curve arrangements and combinatorial structures through the following framework:

Combinatorial Encoding: The authors introduce a combinatorial structure $(n, W, T)$ , termed a "combinatorial curve," to encode the topological type of a curve $D$ transverse to a fixed cellular arrangement of curves $(C_1, \dots, C_k)$ . This structure consists of:
- Intersection Numbers ( $n$ ): Integers representing the number of intersections between the transverse curve and the edges of the cellular decomposition induced by the base arrangement.
- Dyck Words ( $W$ ): Words encoding the sequence of intersections along the boundaries of the faces of the cellular decomposition, tracking the "ground part" of the curve (components intersecting the base lines).
- Rooted Trees ( $T$ ): Trees encoding the arrangement of the "floating part" of the curve (components not intersecting the base lines) within the regions defined by the ground part.
Theoretical Foundation:
- Theorem A: Proves a bijection between topological types of arrangements $(C_1, \dots, C_k, D)$ and combinatorial curves $(n, W, T)$ modulo $i$ -automorphisms of the underlying cellular complex.
- Obstructions (Section 5): The paper develops necessary conditions for realizability based on Bézout's theorem. These include "basic Bézout" (bounding intersection numbers), "Bézout at order 0" (ensuring any two regions can be connected by a combinatorial line with specific intersection properties), and higher-order Bézout criteria.
- Constructions (Section 6): The authors utilize Viro's Patchworking Construction to generate realizable cases from regular triangulations of the Newton polygon. They also employ a "translation" technique, where the combinatorial data of a curve is transformed by shifting the curve relative to the base lines, generating new realizable configurations from existing ones.
Computational Implementation:
- The authors developed the NWT library in the Julia programming language to handle large databases of combinatorial curves.
- The classification process involves enumerating regular triangulations, applying Viro's patchworking and translations, and filtering the resulting combinatorial curves against the Bézout obstructions.
- The algorithm distinguishes between "Admissible" curves (those satisfying necessary Bézout conditions) and "Realized" curves (those constructible via Viro's method and translations).

Key Contributions and Results

Complete Classification for Cubics: The paper provides a complete classification of arrangements of three unlabeled lines and a cubic curve.
- Theorem B: There are exactly 119 topological types of such arrangements.
- The authors identified 3 specific admissible arrangements that were not initially generated by Viro's patchworking or translations but were subsequently proven realizable by constructing explicit polynomial equations.
Partial Classification for Quartics: For arrangements of three unlabeled lines and a quartic curve, the paper offers partial results due to increased complexity.
- Theorem C: There are exactly 619 topological types of arrangements consisting of three lines and a quartic with a single oval.
- Bounds: The total number of topological types for three lines and a quartic is bounded between 9,426 and 17,624.
- Floatless Cases: Restricting to "floatless" arrangements (where every oval of the quartic meets at least one line), the number of types is bounded between 1,834 and 1,883.
- Discrepancy: In the floatless case, there are 49 admissible arrangements (satisfying Bézout conditions) that have not yet been realized by the authors' construction methods.
Two Lines and a Quartic: The authors also classify arrangements of two lines and a quartic, finding that all but 39 cases are accounted for.

Significance and Claims
The paper claims to lay the foundations for the study of topological types of curve arrangements on real compact surfaces. Its significance lies in:

Generalization: Extending the classical Hilbert 16th problem (which concerns single curves) to arrangements of multiple curves, framing the problem in terms of combinatorial objects (Dyck words and trees) rather than just algebraic schemes.
Algorithmic Approach: Providing a systematic, computationally verified method to enumerate and classify these topological types, moving beyond manual classification which becomes intractable for higher degrees.
Bridging Geometry and Combinatorics: Demonstrating that the topological type of an arrangement is fully determined by the combinatorial data $(n, W, T)$ up to automorphism, and that realizability can be tested via combinatorial obstructions derived from Bézout's theorem.
Completeness for Low Degrees: Achieving a complete classification for the $(1,1,1,3)$ case and providing the most comprehensive bounds and partial classifications to date for the $(1,1,1,4)$ case.

The authors note that while they have identified admissible cases that are not yet realized, they lack a systematic method to produce realizations for these specific cases or a new obstruction to exclude them, highlighting an open direction for future research within the scope of their framework.