8fdaeafda56b2126.tex
1: \begin{abstract}
2: In this article, we study $G$-covers
3: of klt varieties,
4: where $G$ is a reductive group.
5: First, we exhibit an example of a klt singularity admitting a
6: $\mathbb{P}{\rm GL}_n(\kk)$-cover that is not of klt type. 
7: Then, we restrict ourselves to $G$-quasi-torsors, a special class of $G$-covers that behave like $G$-torsors outside closed subsets of codimension two. 
8: Given a $G$-quasi-torsor
9: $X\rightarrow Y$, 
10: where $G$ is a finite extension of a torus $\mathbb{T}$,
11: we show that $X$ is of klt type
12: if and only if $Y$ is of klt type.
13: We prove a structural theorem for $\mathbb{T}$-quasi-torsors over normal varieties in terms of Cox rings.
14: As an application, we show that every sequence of $\mathbb{T}$-quasi-torsors over a variety with klt type singularities is eventually a sequence of $\mathbb{T}$-torsors.
15: This is the torus version of a result due to Greb-Kebekus-Peternell regarding finite quasi-torsors of varieties with klt type singularities.
16: On the contrary, we show that in any dimension there exists a sequence of finite quasi-torsors 
17: and $\mathbb{T}$-quasi-torsors
18: over a klt type variety,
19: such that infinitely many of them are not torsors.
20: We show that every variety with klt type singularities is a quotient of a variety with canonical factorial singularities.
21: We prove that a variety with Zariski locally toric singularities
22: is indeed the quotient of a smooth variety by a solvable group. Finally, motivated by the work of Stibitz, we study the optimal class of singularities for which the previous results hold.
23: \end{abstract}
24: