Limit shape of random convex polygonal lines: even more universality (Q406581): Difference between revisions

A convex lattice polygonal line \(\Gamma\) is a piecewise linear continuous path on the plane starting at the origin, with vertices on the two-dimensional integer lattice \(\mathbb Z^2\) and such that the inclination of its consecutive edges is strictly increasing, staying between 0 and \(\pi/2\). Let \(\Pi\) be the set of all convex lattice polygonal lines with finitely many edges, and denote by \(\Pi_n\subset \Pi\) the subset of polygonal lines \(\Gamma\in \Pi\) with the right endpoint \(\xi_{\Gamma}= (\xi_1, \xi_2)\) fixed at \(n=(n_1, n_2)\in \mathbb Z_+^2=\{(k_1, k_2)\in \mathbb Z^2: k_j\geq 0\}\). If each space \(\Pi_n\) is endowed with a probability measure \(P_n\), then one can speak of random polygonal lines, and it is of interest to study their asymptotic statistics as \(n\to \infty\). The limit shape of random polygonal lines, whenever it exists, is defined as a planar curve \(\gamma^*\) such that, for any \(\varepsilon>0\), \(\lim_{n\to \infty} P_n\{\Gamma\in \Pi_n: d(\widetilde{\Gamma}_n, \gamma^*)\leq \varepsilon\}=1\), where \(\widetilde{\Gamma}_n= s_n(\Gamma)\), with a suitable scaling transform \(s_n:\, \mathbb R^2 \to\mathbb R^2\), and \(d(\cdot, \cdot)\) is some metric on the path space, for instance induced by the Hausdorff distance between compact sets. The limit space can depend on the family of probability laws \(P_n\). With respect to the uniform distribution on \(\Pi_n\), the problem was solved independently by \textit{A. M. Vershik} [Funct. Anal. Appl. 28, No. 1, 13--20 (1994); translation from Funkts. Anal. Prilozh. 28, No. 1, 16--25 (1994; Zbl 0848.52004)], \textit{I. Bárány} [Discrete Comput. Geom. 13, No. 3--4, 279--295 (1995; Zbl 0824.52001)] and \textit{Ya. G. Sinai} [Funct. Anal. Appl. 28, No. 2, 108--113 (1994); translation from Funkts. Anal. Prilozh. 28, No. 2, 41--48 (1994; Zbl 0832.60099)], who showed that under the natural scaling \(s_n: (x_1, x_2)\mapsto (n_1^{-1} x_1, n_2^{-1} x_2)\), \(n=(n_1, n_2)\), \(n_1>0\), \(n_2>0\), and with respect to the Hausdorff metric \(d_{{\mathcal H}}\), the limit shape \(\gamma^*\) is given by a parabola arc \(\sqrt{1- x_1} + \sqrt{x_2} =1\), \(0\leq x_1, x_2\leq 1\). Recently, the author and \textit{S. M. Zarbaliev} [Ann. Probab. 39, No. 6, 2271--2317 (2011; Zbl 1242.52007)] considered the limit shape problem for a more general class of multiplicative measures \(\{P_n\}\) (see also the author and \textit{S. M. Zarbaliev} [Dokl. Math. 79, No. 2, 197--202 (2009); translation from Dokl. Akad. Nauk, Ross. Akad. Nauk 425, No. 3, 299--304 (2009; Zbl 1177.82059)]). The goal of the present paper is to show that the limit shape \(\gamma^*\) given by the parabolic arc is universal in a much wider class of multiplicative probability measures \(\{P_n\}\). The rest of the paper is organized as follows. In Section 2 the families of probability measures are defined. In Section 3, suitable values of the parameter are chosen which implies convergence of expected polygonal lines to the limit curve \(\gamma^*\). In Section 4 higher-order moment sums are analyzed. Section 5 is devoted to the proof of the local limit theorem and the limit shape results.