On the existence of optimal meshes in every convex domain on the plane (Q1711820): Difference between revisions

Let \(P_n^d\) be the set of all real algebraic polynomials in \(d\) variables and total degree at most \(n\). A compact subset \(K\subset\mathbb R^d\) has an \textit{optimal mesh} if there exist discrete sets \(Y_n\), \(n\in\mathbb N\), with \(\operatorname{card}{Y}_n\leq{B}n^d\) such that \[ {\|p\|}_{K}\leq{A}{\|p\|}_{Y_n} \] holds for any \(p\in{P}_n^d\), with some \(A,B>0\) depending only on \(K\), where \({\|\cdot\|}_K\) denotes the usual supremum norm on~\(K\). For \(d=2\), the author is able to prove that any convex domain \(K\subset\mathbb R^2\) has an optimal mesh. More precisely, he proves that for \(\varepsilon\in(0,1)\) and \(n\in\mathbb N\) there exist \(Y_n\) such that \({\|p\|}_{K}\leq(1+\varepsilon){\|p\|}_{Y_n}\) for \(p\in{P}_n^2\) and \(\operatorname{card}{Y}_n\leq4\cdot10^5n^2/\varepsilon^2\).