A characterization of bad approximability
arXiv:1707.00771 · doi:10.1088/1361-6544/aaca8e
Abstract
We show that badly approximable vectors are exactly those that cannot, for any inhomogeneous parameter, be inhomogeneously approximated at every monotone divergent rate. This implies in particular that Kurzweil's Theorem cannot be restricted to any points in the inhomogeneous part. Our results generalize to weighted approximations, and to higher irrationality exponents.
19 pages; v2: minor changes, not of a mathematical nature; v3: minor changes, not mathematical