A central limit theorem for descents of a Mallows permutation and its inverse (Q2155509): Difference between revisions

This paper provides a limit theorem for some statistics of a random permutation on \(n\) elements which is not uniformly distributed but follows the Mallows distribution. This distribution is determined by a real parameter \(q> 0\) and the probability of permutation \(w \in S_n\) (where \(S_n\) denotes the symmetric group on \(n\) elements) is proportional to \(q^{\ell (w)}\), where \(\ell (w)\) denotes the number of inversions of \(w\). (Thus, \(\ell (w)\) is the number of pairs \(i<j\) such that \(w(i) > w(j)\).) The first result of the paper provides a central limit theorem for the random variable \(\mathrm{des} (w) + \mathrm{des} (w^{-1})\), where \(\mathrm{des} (w)\) is the number of descents of \(w\). This is the number of positions \(i \in [n-1]\) such that \(w(i+1) < w(i)\). In particular, the first theorem provides an upper bound on the distance between the expected value of any piecewise smooth function of the normalised \(\mathrm{des}(w) + \mathrm{des}(w^{-1})\) and the expected value of the same function applied to a standard normally distributed random variable. The second main theorem of the paper is about the asymptotic distribution of the rescaled vector \((\mathrm{des}(w), \mathrm{des}(w^{-1}))\). It shows that it converges to a normal bivariate distribution.