Multiplicatively badly approximable matrices up to logarithmic factors

Abstract

Let $\|x\|$ denote the distance from $x\in\mathbb{R}$ to the nearest integer. In this paper, we prove an existence and density statement for matrices $\boldsymbol{A}\in\mathbb{R}^{m\times n}$ satisfying $$\liminf_{|\boldsymbol{q}|_{\infty}\to +\infty}\prod_{j=1}^{n}\max\{1,|q_{j}|\}\log\left(\prod_{j=1}^{n}\max\{1,|q_{j}|\}\right)^{m+n-1}\prod_{i=1}^{m}\|A_{i}\boldsymbol{q}\|>0,$$ where the vector $\boldsymbol{q}$ ranges in $\mathbb{Z}^{n}$ and $A_{i}$ are the rows of the matrix $\boldsymbol{A}$. This result extends a previous result of Moshchevitin for $2$-dimensional vectors to arbitrary dimension. The estimates needed to apply Moshchevitin's method to the case $m>2$ are not currently available. We therefore develop a substantially different method, that allows us to overcome this issue. We also generalise this existence result to the inhomogeneous setting. Matrices with the above property appear to have a very small sum of reciprocals of fractional parts. This fact helps us to shed light on a question raised by L\^e and Vaaler, thereby proving some new estimates for such sums in higher dimension.