On the Markov numbers: fixed numerator, denominator, and sum conjectures (Q2048376)

Each solution \((x,y,z)\) with non negative integers of the Diophantine equation \(x^2 + y^2 + z^2 = 3xyz\) is called a Markov triple. Already in 1880, \textit{A. A. Markoff} [Math. Ann. 17, 379--400 (1880; JFM 12.0143.02)] showed that all these solutions could be generated along a binary tree called Markov tree. The Markov numbers are exactly the numbers which appear in these triples. \textit{G. Frobenius} [Berl. Ber. 1913, 458--487 (1913; JFM 44.0255.01)] conjectured that these numbers appear once and exactly once as the maximum of a triple (the unicity conjecture). It is usual to define the Farey tree to construct the set of all rational numbers between 0 and 1 that is all \(p/q\) with \(p\) and \(q\) relatively prime positive integers. This tree is defined using the Farey rules: a node \((\frac{a}{b}, \frac{a+c}{b+d}, \frac{c}{d})\) generates the two next triples \((\frac{a}{b}, \frac{a+(a+c)}{b+(b+d)}, \frac{a+c}{b+d})\) and \((\frac{a+c}{b+d}, \frac{(a+c)+c}{(b+d)+d}, \frac{c}{d})\) (the tree with only the central elements of these triplets is also known as the Stern-Brocot tree and the fractions which appear in these trees are called Farey fractions). We can index all Markov numbers by the Farey fraction which stands at the same place in the Stern-Brocot tree, this correspondence would be one to one if the unicity conjecture was true. We refer to a Markov number by its associated rational in the Farey tree namely \(m_\frac{p}{q}\). \textit{M. Aigner} [Markov's theorem and 100 years of the uniqueness conjecture. A mathematical journey from irrational numbers to perfect matchings. Cham: Springer (2013; Zbl 1276.00006)] suggested three conjectures to better understand the growth of Markov numbers according to their indices. 1. (The fixed numerator conjecture) Let \(p, q\) and \(i\) in \(\mathbb{N}\) such that \(i > 0, p <q, \gcd(p, q) = 1\) and \(\gcd(p, q + i) = 1\) then \(m_\frac{p}{q}<m_\frac{p}{q+i}\). 2. (The fixed denominator conjecture) Let \(p, q\) and \(i\) in \(N\) such that \(i > 0, p + i <q, \gcd(p, q) = 1\) and \(\gcd(p + i, q) = 1\) then \(m_\frac{p}{q}<m_\frac{p+i}{q}\). 3. (The fixed sum conjecture) Let \(p, q\) and \(i\) be positive integers such that \(i < p <q, \gcd(p, q) = 1\) and \(\gcd(p-i, q + i) = 1\) then \(m_\frac{p}{q}<m_\frac{p-i}{q+i}\). The fixed numerator conjecture was proved by \textit{M. Rabideau} and \textit{R. Schiffler} [Adv. Math. 370, Article ID 107231, 17 p. (2020; Zbl 1440.13101)] using cluster algebra and snake graphs. In this paper, the authors proved the above three conjectures are true by the tools as the words, paths on the \(\mathbb{N}^2\) grid and the Christoffel words. Along the way, they generalized Markov numbers to any couple \((p, q) \in \mathbb{N}^2\) (not only when they are relatively prime) and conjectured that the unicity is still true as soon as \(p \leq q\). Finally, the authors showed that the three conjectures are in fact true for this superset.