Coherent random permutations with record statistics

Alexander Gnedin
DOI: https://doi.org/10.48550/arXiv.math/0611538
2006-11-18
Abstract:Random permutations with distribution conditionally uniform given the set of record values can be generated in a unified way, coherently for all values of $n$. Our central example is a two-parameter family of random permutations that are conditionally uniform given the counts of upper and lower records. This family interpolates between two versions of Ewens' distribution. We discuss characterisations of the conditionally uniform permutations, their asymptotic properties, constructions and relations to random partitions.
Probability,Combinatorics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the distribution characteristics of random permutations under the condition of certain statistics. Specifically, the author studied a special class of random permutations that are uniformly distributed given the number of upper and lower records. By introducing a two - parameter family of random permutation distributions \(P(\theta, \zeta)\), the author extended the existing results and explored the characteristics, construction methods of this type of permutations and their relationship with random partitions. ### Core Problems of the Paper 1. **Random Permutations with Conditional Uniform Distribution**: - Study the conditional uniform distribution of random permutations given the number of upper and lower records. - Introduce a two - parameter family of distributions \(P(\theta, \zeta)\), where \(\theta\) and \(\zeta\) control the number of lower and upper records respectively. 2. **Uniqueness of the Mixture Distribution**: - Prove that any sequence of random permutations that is conditionally uniformly distributed given the number of upper and lower records can be represented as a mixture distribution of \(P(\theta, \zeta)\). 3. **Generation Methods**: - Provide specific methods for generating this type of random permutations, including methods based on the independence of initial ranks and the Polya - Eggenberger distribution. 4. **Asymptotic Properties**: - Explore the asymptotic behavior of record values and record times when the permutation length \(n\) tends to infinity, for example, they converge to independent Gaussian variables or Poisson processes. ### Presentation of Formulas in Markdown Format - **Two - Parameter Distribution Formula**: \[ P(\theta, \zeta)_n(\pi_n)=\frac{\theta^{\ell}\zeta^{u}}{(\theta + \zeta)^{n - 1}} \] where \(\ell\) and \(u\) are the number of lower and upper records respectively. - **Binary Generating Function Formula**: \[ \sum_{\ell, u}\left[n\atop\ell + 1,u + 1\right]\theta^{\ell}\zeta^{u}=(\theta + \zeta)^{n - 1} \] - **Asymptotic Properties**: - Mean and Variance of the Number of Records: \[ E[\ell]\sim\theta\log n,\quad E[u]\sim\zeta\log n \] - Limit Distribution of Standardized Record Times: \[ \frac{t_k}{n}\to\rho_k\quad\text{a.s.} \] where \(\rho_0\) follows the beta distribution \(\text{beta}(\theta, \zeta)\) Through the study of these problems, the author not only extended the existing theoretical framework but also provided a series of new tools and methods to understand and generate random permutations with specific record statistical characteristics.