[Paper Review] Approximate groups, II: the solvable linear case

Emmanuel Breuillard, Ben Green|ArXiv.org|Jul 6, 2009

Finite Group Theory Research7 references4 citations

TL;DR

This paper establishes that $K$-approximate subgroups in solvable subgroups of $\mathrm{GL}_n(\mathbb{C})$ are efficiently controlled by nilpotent progressions of bounded step and dimension polynomial in $K$. By combining structural results on solvable linear groups with prior work on torsion-free nilpotent groups, the authors show that such approximate subgroups are $e^{K^{C_n}}$-controlled by nilpotent progressions of dimension at most $K^{C_n}$ and step at most $n-1$, providing polynomial bounds where previous results had only weak or exponential dependencies.

ABSTRACT

We describe the structure of "K-approximate subgroups'' of solvable subgroups of GL_n(C), showing that they have a large nilpotent piece. By combining this with the main result of our recent paper on approximate subgroups of torsion-free nilpotent groups, we show that such approximate subgroups are efficiently controlled by nilpotent progressions.

Motivation & Objective

• To understand the structure of $K$-approximate subgroups within solvable subgroups of $\mathrm{GL}_n(\mathbb{C})$.
• To extend the classification of approximate subgroups from torsion-free nilpotent groups to the broader class of solvable linear groups.
• To establish polynomial bounds in the approximation parameter $K$ for control by nilpotent progressions, improving upon earlier results with exponential or unspecified bounds.
• To demonstrate that the nilpotent structure of such approximate subgroups is stable and efficiently representable via nilpotent progressions.

Proposed method

• Leveraging Mal’cev’s theorem to conjugate solvable subgroups into upper-triangular matrices $\mathrm{Upp}_n(\mathbb{C})$, enabling structural analysis.
• Applying Jordan decomposition to split any element into semisimple and unipotent parts, and proving that the semisimple and unipotent subgroups commute and form a direct product.
• Using the nonabelian Ruzsa covering lemma to relate the size and control of sets in the group, particularly in the context of approximate groups.
• Reducing the problem to the universal cover of the unipotent part, where the group is isomorphic to $\mathbb{R}^n \times \Gamma$ with $\Gamma$ a simply-connected nilpotent Lie group.
• Applying the main result from Part I of the series [3] on torsion-free nilpotent groups to show that a subset of the approximate group lifts to a nilpotent progression in the universal cover.
• Projecting the nilpotent progression back down to the original group to obtain a control set of exponential size in $K^{C_n}$, yielding the final control result.

Experimental results

Research questions

• RQ1 How can the structure of $K$-approximate subgroups in solvable subgroups of $\mathrm{GL}_n(\mathbb{C})$ be characterized?
• RQ2 Can such approximate subgroups be efficiently controlled by nilpotent progressions, and with what quantitative bounds?
• RQ3 What is the role of the unipotent and semisimple components in the Jordan decomposition of elements of solvable linear groups for the structure of approximate subgroups?
• RQ4 How does the dependence on the approximation parameter $K$ compare to previous results in the literature, particularly in terms of bound quality?
• RQ5 To what extent can the structure of approximate subgroups in solvable groups be reduced to that of nilpotent groups via algebraic decomposition?

Key findings

• Every $K$-approximate subgroup $A$ of a solvable subgroup of $\mathrm{GL}_n(\mathbb{C})$ is $K^{C_n}$-controlled by a $K^{C_n}$-approximate subgroup $B$ that generates a nilpotent group of step at most $n-1$ inside a conjugate of $\mathrm{Upp}_n(\mathbb{C})$.
• The same approximate subgroup $A$ is $e^{K^{C_n}}$-controlled by a nilpotent progression of dimension at most $K^{C_n}$ and step at most $n-1$.
• The control bounds are polynomial in $K$ for the intermediate nilpotent control, and exponential in $K^{C_n}$ for the final nilpotent progression control, with the exponent $C_n$ depending only on $n$.
• The proof relies on the Jordan decomposition of elements in $\mathrm{GL}_n(\mathbb{C})$, showing that the semisimple and unipotent parts commute and form a direct product decomposition of the group.
• The unipotent part is simply-connected and its universal cover is isomorphic to $\mathbb{R}^n \times \Gamma$ with $\Gamma$ a simply-connected nilpotent Lie group, enabling application of prior results on nilpotent progressions.
• By lifting a large subset of $A$ to the universal cover and applying the main result from [3], a nilpotent progression is constructed that controls the original approximate group with exponential dependence on $K^{C_n}$.