Automorphisms of the canonical double cover of a toroidal grid

TL;DR

This paper proves Wilson's conjecture that most toroidal and triangular grids are stable in terms of automorphisms, except for specific known unstable cases, thereby clarifying the symmetry properties of these grid structures.

Contribution

It establishes the stability of all non-degenerate toroidal and triangular grids, confirming Wilson's conjectures and extending understanding of their automorphism groups.

Findings

Most toroidal grids are stable except for known unstable families.

All non-degenerate triangular grids are stable.

Wilson's conjectures on grid stability are proven.

Abstract

The Cartesian product of two cycles (of length m and length n) has a natural embedding on the torus, such that each face of the embedding is a 4-cycle. The toroidal grid Qd(m,n,r) is a generalization of this in which there is a shift by r when traversing the meridian of length m. In 2008, Steve Wilson found two interesting infinite families of (nonbipartite) toroidal grids that are unstable. (By definition, this means that the canonical bipartite double cover of the grid has more than twice as many automorphisms as the grid has.) It is easy to see that bipartite grids are also unstable, because the canonical double cover is disconnected. Furthermore, there are degenerate cases in which there exist two different vertices that have the same neighbours. This paper proves Wilson's conjecture that Qd(m,n,r) is stable for all other values of the parameters. In addition, we prove an analogous conjecture of Wilson for the triangular grids Tr(m,n,r) that are obtained by adding a diagonal to each face of Qd(m,n,r) (with all of the added diagonals parallel to each other).