Bisector energy and few distinct distances (Q312146): Difference between revisions

The famous distinct distances conjecture of \textit{P. Erdős} [Am. Math. Mon. 53, 248--250 (1946; Zbl 0060.34805)] asserts that any \(n\) points in the plane determine \(\Omega(n /\sqrt{\log n})\) distinct distances. He showed that an \(\sqrt{n}\times \sqrt{n}\) grid has only \(\Omega(n /\sqrt{\log n})\) distinct distances, and also conjectured that [Discrete Math. 60, 147--153 (1986; Zbl 0595.52013)] any point set point set with so few distinct distances has a ``grid structure'', e.g. there is line with \(\Omega(\sqrt{n})\) points. A recent celebrated paper of \textit{L. Guth} and \textit{N. H. Katz} [Ann. Math. (2) 181, No. 1, 155--190 (2015; Zbl 1310.52019)] essentially solved the distinct distances conjecture proving that any \(n\) points in the plane determine \(\Omega(n /{\log n})\) distinct distances. The paper under review takes some steps towards establishing a ``grid structure'': if \(n\) points determine only \(\Omega(n /\sqrt{\log n})\) distinct distances, then for every \(0<\alpha\leq 1/4\), there exists a line or a circle that contains at least \(n^\alpha\) of the points, or there exist \(\Omega(n^{8/5-12\alpha/5-\epsilon})\) distinct lines that each contain \(\Omega({\log n})\) of the points. This result is based on a new upper bound on the number of isosceles trapezoids among \(n\) points: \(O(M^{2/5}n^{(12+\epsilon)/5}+ Mn^2)\), where \(\epsilon>0\) is arbitrary fixed number, and \(M\) is a strict upper bound on the number points on any line or circle among the given \(n\) points.