Table of Contents
Fetching ...

An almost linear time algorithm testing whether the Markoff graph modulo $p$ is connected

Colby Austin Brown

Abstract

The Markoff graph modulo $p$ is known to be connected for all but finitely many primes $p$ (see Eddy, Fuchs, Litman, Martin, Tripeny, and Vanyo [arxiv:2308.07579]), and it is conjectured that these graphs are connected for all primes. In this paper, we provide an algorithmic realization of the process introduced by Bourgain, Gamburd, and Sarnak [arxiv:1607.01530] to test whether the Markoff graph modulo $p$ is connected for arbitrary primes. Our algorithm runs in $o(p^{1 + ε})$ time for every $ε> 0$. We demonstrate this algorithm by confirming that the Markoff graph modulo $p$ is connected for all primes less than one million.

An almost linear time algorithm testing whether the Markoff graph modulo $p$ is connected

Abstract

The Markoff graph modulo is known to be connected for all but finitely many primes (see Eddy, Fuchs, Litman, Martin, Tripeny, and Vanyo [arxiv:2308.07579]), and it is conjectured that these graphs are connected for all primes. In this paper, we provide an algorithmic realization of the process introduced by Bourgain, Gamburd, and Sarnak [arxiv:1607.01530] to test whether the Markoff graph modulo is connected for arbitrary primes. Our algorithm runs in time for every . We demonstrate this algorithm by confirming that the Markoff graph modulo is connected for all primes less than one million.
Paper Structure (4 sections, 12 theorems, 37 equations, 7 figures, 2 tables, 3 algorithms)

This paper contains 4 sections, 12 theorems, 37 equations, 7 figures, 2 tables, 3 algorithms.

Key Result

Lemma 1

Let $z \in \mathbb{Z} / p\mathbb{Z}$. Then there are $x, y \in \mathbb{Z} / p\mathbb{Z}$ such that if and only if either $z \neq 2$ or $-1$ is a quadratic residue modulo $p$.

Figures (7)

  • Figure 1: The Markoff tree $\mathcal{G^\times}$.
  • Figure 2: The Markoff graph $\mathcal{G}^\times_5$.
  • Figure 3: The factor trie for $n = 60$.
  • Figure 4: Total number of bad triples for all primes less than $1,000,000$ (top) and random primes less than $110,000,000$ (bottom).
  • Figure 5: Maximum number of orbits checked during any iteration of Step \ref{['step:cosets']} for all primes less than $1,000,000$ (top) and random primes less than $110,000,000$ (bottom).
  • ...and 2 more figures

Theorems & Definitions (29)

  • Lemma 1
  • proof
  • Conjecture 1
  • Conjecture 2
  • Lemma 2
  • proof
  • Theorem 1
  • Lemma 3
  • proof
  • Lemma 4
  • ...and 19 more