${\mathbb Z}_{p}^{m}$-actions of type $(d;p,n)$
Ruben A. HIdalgo, Maximiliano Leyton-Alvarez
TL;DR
This work analyzes ${Z_p^m}$-actions of type $(d;p,n)$ on compact complex manifolds whose quotient by the action is $P^d$ with $n+1$ hyperplanes of branch order $p$. It develops a framework via generalized Fermat varieties $X_n^p(\Lambda)$ and freely acting subgroups $K\le H$, showing that any such action arises as $S\cong X_n^p(\Lambda)/K$ with $N\cong H/K$, and that the automorphism structure lifts through exact sequences, yielding normality of $N$ in ${Aut}(S)$ and rigidity $M=N$ for isomorphic actions when $(d;p,n)\notin \{(2;2,5),(2;4,3)\}$. The paper also constructs the parameter space $\Omega_{n,d}$, provides explicit equations for the generalized Fermat varieties, analyzes fixed-point loci of canonical generators, and proves non-hyperbolicity in broad parameter ranges, notably when $d+1\le m\le n\le 2d-1$. Together these results give a precise description of when ${Aut}(S)$ is finite, when the ${Z_p^m}$-action is unique up to conjugacy, and how hyperbolicity properties fail in many cases, contributing to the understanding of abelian group actions on higher-dimensional projective varieties. The findings have potential implications for rigidity, automorphism group structure, and hyperbolicity questions in complex geometry and algebraic geometry.
Abstract
A ${\mathbb Z}_{p}^{m}$-action of type $(d;p,n)$, where $2 \leq d \leq m \leq n$ are integers, is a pair $(S,N)$ where $S$ is a $d$-dimensional compact complex manifold, $N \cong {\mathbb Z}_{p}^{m}$ is a group of holomorphic automorphisms of $S$ such that the quotient orbifold $S/N$ is the $d$-dimensional projective space ${\mathbb P}^{d}$ whose branch locus consists of $n+1$ hyperplanes in general position, each one of branch order $p$. If $(d;p,n) \notin \{(2;2,5),(2;4,3)\}$ and $d+1 \leq n$, then we prove that: (i) $N$ is a normal subgroup of ${\rm Aut}(S)$ and (ii) if $(S,M)$ is a ${\mathbb Z}_{\hat{p}}^{\hat{m}}$-action of type $(d;\hat{p},\hat{n})$, then $M=N$. If, moreover, $d+1 \leq n \leq 2d-1$, then we observe that $S$ is not algebraically hyperbolic
