On the dimension of Kakeya sets and related maximal inequalities (Q1293232)

A Kakeya set in \(\mathbb R^d\) is a compact set \(E\subseteq\mathbb R^d\) containing a line segment in every direction. Such sets may be of measure zero. The paper concerns the conjecture saying that such sets have Hausdorff dimension \(\dim E=d\). For \(d=2\) the conjecture was proved by \textit{R. O. Davies} [Proc. Cambr. Philos. Soc. 69, 417--421 (1971; Zbl 0209.26602)]. \textit{T. Wolff} [Rev. Math. Iberoam. 11, 651--674 (1995; Zbl 0848.42015)] proved that \(\dim E\geq\frac d2+1\) and for \(d=3\) he obtained also the best estimate to this date, \(\dim E\geq\frac 52\). The author of the paper under review proves that \(\dim E\geq\frac{13}{25}d+\frac{12}{25}\) for any Kakeya set \(E\subseteq\mathbb R^d\). This result is derived from another result saying that certain compact sets \(E\) in \(\mathbb R^d\) have Minkowski dimension \(\geq\frac{13}{25}(d-1)\). The author considers also certain maximal operators associated to the Kakeya problem and a conjecture on these operators which imply the conjecture on Kakeya sets. Using this reduction he obtains the existence of a constant \(c>\frac 12\) (independent on \(d\)) for which every Kakeya set \(E\subseteq\mathbb R^d\) has Hausdorff dimension \(\geq c\cdot d\). Some combinatorial facts on sizes of sum sets and difference sets for finite sets are exploited and some new results on the reciprocals of an arithmetic progression in sum sets are obtained.