On random disc-polygons in a disc-polygon (Q6110565)

For \(r>0\), an \(r\)-disc-polygon \(P\) is a non-empty compact convex set in the Euclidean plane \({\mathbb R}^2\) that is bounded by finitely many circular arcs of radius \(r\) (and hence is contained in a circular disc of radius \(r\)). For a finite point set in \({\mathbb R}^2\) contained in a circular disc of radius \(r\), let \([X]_r\) be the intersection of all circular discs of radius \(r\) containing it. For given \(P\) as above, different from a circular disc of radius \(r\), and an integer \(n\), let \(x_1,\dots,x_n\) be independent random points in \(P\) with uniform distribution, and let \(P_n^r:=[\{x_1,\dots,x_n\}]_r\). This is a random \(r\)-disc polygon contained in \(P\). For the vertex number \(f_0\) and (consequently) the area \(A\), the authors prove that \[ \lim_{n\to\infty} \frac{{\mathbb E}f_0(P_n^r)}{\ln n} = \frac{2}{3}f_0(P),\qquad \lim_{n\to\infty} \frac{n{\mathbb E}A(P\setminus P_n^r)}{\ln n} = \frac{2}{3}f_0(P)A(P),\] where \({\mathbb E}\) denotes the expectation. Thus, these results are analogous to classical asymptotic relations found by Rényi and Sulanke for ordinary polygons. However, the authors had to develop different methods of proof.