Independence of permutation limits at infinitely many scales (Q2062759)

Let \(\pi\) be a permutation of length \(k\) (called a pattern) and \(\sigma\) be a permutation of length \(n\). The global density of \(\pi\) in \(\sigma\) is defined as the probability that \(k\) entries of \(\sigma\), chosen uniformly at random, are order-isomorphic to \(\pi\). This is also frequently called the packing density of \(\pi\) in \(\sigma\) and is studied in its own right in works such as [\textit{J. Sliačan} and \textit{W. Stromquist}, Discrete Math. Theor. Comput. Sci. 19, No. 2, Paper No. 3, 18 p. (2017; Zbl 1401.05018)] and in the context of permutons in such works as [\textit{R. Kenyon} et al., Random Struct. Algorithms 56, No. 1, 220--250 (2020; Zbl 1442.05008)]. The local density of \(\pi\) in \(\sigma\) is the probability that \(k\) consecutive entries of \(\sigma\), chosen uniformly at random, are order-isomorphic to \(\pi\). \textit{J. Borga} and \textit{R. Penaguiao} [Algebr. Comb. 3, No. 6, 1259--1281 (2020; Zbl 1462.05003)] have shown that the global and local densities are independent in the sense that given any consistent set of global densities and any consistent set of local densities, there is a sequence of permutations whose global and local densities converge to those given. Here, this result is extended to infinitely many scales simultaneously. The density of \(\pi\) in \(\sigma\) at scale \(f\) is defined to be the probability that \(k\) entries of \(\sigma\) lying within a window of length \(f\) and chosen uniformly at random are order-isomorphic to \(\pi\). Thus global density is density at scale \(n\) and local density is density at scale \(k\), but one may also consider scales such as \(\log n\) or \(\sqrt{n}\). The main result is that if any consistent combination of asymptotic pattern densities is chosen for each of a countably infinite number of suitably distinct scales, then there exists a sequence of permutations for which all the limit densities at each scale match the choices.