Abstract
[latexpage]In this paper, we study the operation of the Rotor-Router and attempt to find a method, a process, that allows integer grouping. We begin by defining the mapping $f : \mathbb{N}^* \longrightarrow \mathbb{Z}^2,$ which associates to each integer $n \in \mathbb{N}^*$ a position, $f(n) \in \mathbb{Z}^2$, in the plane, determined by the dynamics of the Rotor-Router. The mapping $f$ thus provides a visual representation of the positive natural numbers in the plane. Filters are then applied in order to retain only the integers possessing certain characteristics, such as divisibility, congruence, primality, etc. In other words, starting from $n \in \mathbb{N}^*$, we select subsets satisfying properties such as $d \mid n, \qquad n \equiv a \pmod{m}, \qquad n \ \text{prime}, \; \dots$ The Rotor-Router mechanism produces an organization of integers whose observed regularity suggests an unexplored arithmetic distribution. In particular, the study of arithmetic progressions $4k+1$ and $4k+3$, and more generally of congruence classes $n \equiv a \pmod{b},$ has made it possible to highlight non-uniform distribution phenomena, revealing the existence of remarkable arithmetic biases in the distribution of prime numbers.
The observations and conjectures presented on this page may evolve in light of new simulations or critical feedback.
