On linear combinatorics. I: Concurrency---an algebraic approach

From MaRDI portal





The subject of this paper is the following Erdős-type Ramsey problem: Consider two sets \(\Phi\) and \(\Psi\) of lines (in the Euclidean plane) which are the images of affine mappings \( x \mapsto ax+b \) \( (a \not= 0) \), and a set \( \Omega(E) \) of composite mappings \(\left\{ \phi\circ\psi \mid (\phi,\psi) \in E \right\} \) or \(\left\{ \phi\circ\psi, \psi\circ\phi \mid (\phi,\psi) \in E \right\} \) (determined by some \( E \subset \Phi\times\Psi \)). The main result reads: If the size of \(\Omega(E)\) is not too large (at most \(Cn\)) compared to the sizes of \(\Phi\) and \(\Psi\) (at least \(n\), and also less than \(Cn\)), and \(E\) (at least \(cn^2\)) then there are relatively large sets \( \Phi^* \subset \Phi \) and \( \Psi^* \subset \Psi \) (i.e., such that \( \left| (\Phi^* \times \Psi^*) \cap E \right| \geq c^*n^2 \) for some \(c^*(C,c)>0\)) which are both either bundles of parallel lines, or bundles of concurrent lines.











This page was built for publication: On linear combinatorics. I: Concurrency---an algebraic approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1280269)