A systematic approach to Diophantine equations: open problems

Imagine you are a detective trying to solve a massive, infinite library of math puzzles. These aren't just any puzzles; they are Diophantine equations. In plain English, these are algebraic recipes where you can only use whole numbers (like 1, 2, -5, 0) as your ingredients.

The goal is simple: Find the whole numbers that make the recipe balance (equal zero).

The problem? Some of these recipes look incredibly simple to write down, but they are so stubborn that no human (or computer) has been able to crack them yet.

This paper, written by Bogdan Grechuk, is essentially a "Most Wanted" list for these stubborn puzzles. It organizes them by how "small" or "simple" they are, helping mathematicians know exactly where to focus their energy.

Here is a breakdown of the paper's concepts using everyday analogies:

1. The "Size" of a Puzzle (The H-Measure)

How do you decide which puzzle is the "smallest" or simplest? You can't just count the number of letters.

The Analogy: Imagine every variable (like $x$ , $y$ , $z$ ) is a brick, and every number coefficient (like the 2 in $2x$) is a weight attached to that brick.
The Rule: To calculate the "size" ( $H$ ) of a puzzle, you pretend every variable is the number 2, take the absolute value of every weight, and do the math.
Why it matters: This gives us a ruler. We can say, "Okay, we've solved all puzzles with a size of 10. Now, let's look at the ones with a size of 11." This paper lists the smallest ones that are still unsolved.

2. The Different Types of "Unsolved"

The paper doesn't just ask "Does a solution exist?" It asks different levels of questions, like a detective asking different types of questions about a crime scene.

Level 1: The "Recipe" Question (Polynomial Parametrization)

The Question: Can we write a single master formula (a recipe) that generates every possible solution?
The Analogy: Imagine a vending machine. If you press button A, you get a soda. If you press button B, you get a chip. A "polynomial family" is like a machine where you just turn a dial (the variable $u$ ), and it spits out a valid solution every time.
The Open Problem: For some equations (like $x^2 + y^2 + zt + 1 = 0$ ), we don't know if such a "master dial" exists. We might find a few solutions by luck, but we can't write a rule to find them all.

Level 2: The "Description" Question

The Question: If we can't write a master recipe, can we at least describe the solutions in a "reasonable" way? Maybe using a list of starting points and a set of rules to jump from one solution to the next?
The Analogy: Think of a maze. Sometimes you can't draw the whole map at once, but you can say, "Start here, and if you hit a wall, turn left."
The Open Problem: For equations like $y^2 + z^2 = x^3 + 1$ , we don't even know if there's a simple way to describe the path through the maze.

Level 3: The "Existence" Question (Hilbert's 10th Problem)

The Question: Does any solution exist at all?
The Analogy: You are looking for a needle in a haystack. You don't need to find the needle; you just need to know if the haystack contains one.
The Open Problem: For some very short equations (like $y(x^3 - z^2) = z$ ), we genuinely do not know if a whole-number solution exists. It's like flipping a coin that has landed on its edge—we just can't tell which side is up.

3. The Categories of Puzzles

The author sorts these puzzles into different "bins" to make the list manageable:

Symmetric: The puzzle looks the same no matter how you swap the variables (like a circle).
Cyclic: The puzzle looks the same if you rotate the variables (like a necklace).
Homogeneous: Every term in the equation has the same "degree" (complexity), like a cake where every layer is the same height.
Independent: The variables don't mix; they stay in their own lanes.

4. The "Length" vs. "Size"

The paper introduces a second way to measure difficulty called Length ( $l$ ).

The Analogy: If "Size" is the weight of the puzzle, "Length" is the number of words in the sentence.
Why it matters: Some puzzles are very short to write (few words) but have huge numbers in them. The paper lists the shortest sentences that are still unsolved. For example, an equation with only 8 "words" (like $y(x^3 - z^2) = z$ ) is still a mystery!

5. The "Updates" (The Living Document)

The paper is a "living" document. It's not a static book; it's a wiki that gets updated.

The Analogy: Imagine a "Wanted" poster. Sometimes, a detective solves a case, and the police cross that name off the list.
What happened here: The author mentions that between versions of this paper, several equations were "solved" (crossed off) because other mathematicians found the answers. Some were solved by humans, and interestingly, one was solved by an AI (ChatGPT)!

Summary

This paper is a roadmap for the unknown.
It tells us: "We have solved all the tiny, simple puzzles. Here are the smallest, simplest puzzles that are still mysteries. If you want to be a math hero, start here."

It highlights that even in the world of simple whole numbers, there are still deep, beautiful, and baffling secrets waiting to be discovered.

Based on the provided document, here is a detailed technical summary of the paper "A systematic approach to Diophantine equations: open problems" by Bogdan Grechuk.

1. Problem Statement

The paper addresses the fundamental challenge in number theory of classifying and solving polynomial Diophantine equations (equations of the form $P(x_1, \dots, x_n) = 0$ with integer coefficients). While Hilbert's 10th problem established that no general algorithm exists to determine if any Diophantine equation has a solution, the specific difficulty of individual equations varies wildly.

The core problem is to identify the smallest Diophantine equations (in terms of a defined complexity measure) for which specific solution properties remain unknown. The author aims to create a living, systematic list of these "smallest open problems" to guide future research and track progress in solving them.

2. Methodology

A. Complexity Metric: Size $H(P)$

To order equations systematically, the author utilizes the size metric $H(P)$ , defined in prior work [1]. For a polynomial $P$ with $k$ non-similar monomials $a_i x_1^{k_{i1}} \dots x_n^{k_{in}}$ of degrees $d_i$ , the size is:
$H(P) = \sum_{i=1}^k |a_i| 2^{d_i}$
This metric effectively substitutes $2 $for every variable and takes the absolute value of coefficients, creating a hierarchy where equations with smaller$ H $are considered "simpler." The author notes that for any fixed size$ H$, there are only finitely many distinct equations (up to equivalence).

B. Equivalence Classes

To avoid redundancy, equations are grouped into equivalence classes. Two equations are equivalent if they can be transformed into one another via:

Multiplication by a non-zero constant.
Substitution $x_i \to -x_i$ .
Renaming or permuting variables.
Only one representative from each class is listed.

C. Categorization

The paper organizes open problems by:

Variable count: Unrestricted, 2-variable, 3+ variables.
Structure: Homogeneous, symmetric, cyclic, independent monomials.
Degree: Cubic, quartic, etc.
Alternative Metric: A "length" metric $l(P)$ based on the logarithm of the product of coefficients and powers of 2, used to find the shortest equations by character count.

D. Problem Definitions

The paper defines a hierarchy of specific questions (Problems 1–7) regarding these equations:

Problem 1: Can all integer solutions be described as a finite union of polynomial families?
Problem 2: Does the equation have arbitrarily large integer solutions (i.e., $\min(|x_i|) \ge k$ for any $k$ )? If not, describe all solutions.
Problem 3: (For homogeneous equations) Does a non-trivial integer solution exist where all variables are non-zero?
Problem 4: Is the set of integer solutions finite or infinite? If finite, list them.
Problem 5: (For homogeneous equations) Does a non-zero integer solution exist?
Problem 6: Does an integer solution exist at all? (Hilbert's 10th context).
Problem 7: Does a solution exist in positive integers?

3. Key Contributions

Systematic Cataloging: The paper provides the first comprehensive, ordered list of the smallest open Diophantine equations across multiple categories (symmetric, cyclic, homogeneous, etc.) based on the $H(P)$ metric.
Identification of Thresholds: It pinpoints the exact size $H$ where specific problems transition from "solved" to "open." For example, it identifies that for $H \le 31$ , all 2-variable equations have finite solution sets (though listing them is open), while at $H=32$ , the genus 2 equation $x^4 - x^2 + xy + y^3 = 0$ becomes open.
Dynamic Updates: The document serves as a version-controlled record (versions 1 through 5), explicitly documenting which equations have been solved in the interim by other researchers or AI (e.g., ChatGPT 5.4 Pro) and removing them from the open lists.
Alternative Metrics: It introduces the "length" metric $l(P)$ to identify the shortest equations by character count, revealing that some equations with small $H$ are actually quite long, and vice versa.

4. Key Results and Open Equations

The paper lists numerous specific equations that remain open. Notable examples include:

Polynomial Parametrization (Problem 1): The smallest open equations have size $H=13$ (e.g., $x^2 + y^2 + zt + 1 = 0$ ).
Describing Solutions (Problem 2/4):
- The smallest open equations for describing integer solutions are $y^2 + z^2 = x^3 \pm 1$ ( $H=17$ ).
- The smallest open 2-variable equations (genus 3) appear at $H=28$ (e.g., $x^4 - y^3 - x + y = 0$ ).
- The smallest open symmetric equation is $x^3 + y^3 + z^3 = 3$ ( $H=27$ ).
Existence of Solutions (Problem 6):
- The smallest open 2-variable equations are at $H=32$ (e.g., $y^3 + xy = x^4 + 4$ ).
- The smallest open cubic equation is $2 + 2x + x^3 - y^2 - xy^2 + xz^2 = 0 $($ H=34$).
Homogeneous Equations (Problem 5):
- The smallest open homogeneous equation is $x^4 + x^3y - y^4 + y^3z + z^4 = 0$ ( $H=80$ ).
- For the form $ax^d + by^d + cz^d = 0$ , the smallest open case is $4x^5 + 4y^5 + 11z^5 = 0 $($ H=608$).
Positive Integers (Problem 7): The smallest open equations are $y(x^3 - z^2) = z$ and $x^2y^2 + x = z^3$ ( $H=26$ ).

5. Significance

Research Roadmap: By ordering equations by size, the paper provides a clear "to-do list" for number theorists. It suggests that if a general theory exists for a certain class of equations, it should likely be discoverable by solving the smallest open instances first.
Benchmarking: The list serves as a benchmark for testing new algorithms, heuristics, and even AI systems (as evidenced by the removal of equations solved by ChatGPT in Version 5).
Clarification of Difficulty: It highlights that "small" equations are not necessarily "easy." Equations with very low $H$ (e.g., $H=13$ or $H=17$ ) remain stubbornly open, indicating deep structural gaps in current number theory.
Community Resource: The document acts as a centralized, up-to-date repository, preventing duplicate efforts and ensuring the community focuses on the current frontier of unsolved problems.

In summary, this paper transforms the chaotic landscape of Diophantine equations into a structured, ordered hierarchy, identifying the precise boundaries of current mathematical knowledge and providing a dynamic target for future breakthroughs.