Probabilistic Stirling numbers associated with sequences

Dae San Kim, Taekyun Kim · 2025

Read Probabilistic Stirling numbers associated with sequences by Dae San Kim, Taekyun Kim (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-14568.txt.

Introduction

Dae San Kim and Taekyun Kim redefine probabilistic Stirling numbers linked to a random variable. Stirling numbers count partitions and permutations, and recent work added a probabilistic version based on moments. The earlier definitions failed to satisfy key orthogonality and inverse relations. Those relations are central to how the two kinds of numbers interact.

The authors define the first kind through the cumulant generating function in a corrected way. With that fix, both kinds satisfy the crucial relations again. They also study degenerate counterparts of both kinds. Explicit formulas are computed for the numbers. The paper restores a clean algebraic structure for later combinatorial use.

Catalog record

CREATORDae San Kim, Taekyun Kim
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-14568

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Probabilistic Stirling numbers associated with sequences Dae San Kim, Taekyun Kim Abstract . Let Y Y be a random variable whose moment generating function exists in a neighborhood of the origin. Recently, probabilistic Stirling numbers of the first kind, s Y ​ ( n , k ) s_{Y}(n,k) , and of the second kind, S 2 Y ​ ( n , k ) S_{2}^{Y}(n,k) , associated with Y Y have been introduced. However, s Y ​ ( n , k ) s_{Y}(n,k) , based on the cumulant generating function of Y Y , and S 2 Y ​ ( n , k ) S_{2}^{Y}(n,k) do not satisfy orthogonality and inverse relations. This paper aims to redefine the probabilistic Stirling numbers of the first kind associated with Y Y , denoting them by S 1 Y ​ ( n , k ) S_{1}^{Y}(n,k) , such that S 1 Y ​ ( n , k ) S_{1}^{Y}(n,k) and S 2 Y ​ ( n , k ) S_{2}^{Y}(n,k) do satisfy these crucial relations. Furthermore, we investigate their degenerate counterparts, the probabilistic degenerate Stirling numbers of both kinds. We explicitly compute S 2 Y ​ ( n , k ) , S 1 Y ​ ( n , k ) S_{2}^{Y}(n,k),S_{1}^{Y}(n,k) , and their degenerate versions for several discrete and continuous random variables Y Y . As an application, we express arbitrary polynomials as linear combinations of probabilistic Euler and probabilistic degenerate Euler polynomials. 2010 Mathematics Subject Classification 11B73; 11B83; 60-08 1. Introduction The Stirling number of the second, denoted by S_{2}(n,k) , counts the number of ways to partition a set of n objects into k nonempty subsets. The signed Stirling number of the first kind, denoted by, S_{1}(n,k) is defined such that the number of permutations of n elements with exactly k cycles is given by |S_{1}(n,k)|=(-1)^{n-k}S_{1}(n,k) (see, for example, [7,23,24,26]). The degenerate Stirling numbers of the second kind, S_{2,\lambda}(n,k) (defined in (3)), and of the first kind, S_{1,\lambda}(n,k) (defined in (5)), are degenerate versions of S_{2}(n,k) and S_{1}(n,k) , respectively (see [13,14,22]). These degenerate Stirling numbers frequently arise in the study of degenerate versions of various special polynomials and numbers (see [8,9,20]). The study of degenerate versions of special polynomials and numbers has seen a resurgence of interest and a corresponding increase in research activity. These investigations have employed a diverse range of mathematical tools, including quantum mechanics, generating functions, combinatorial methods, p -adic analysis, umbral calculus, probability theory, differential equations, analytic number theory, and operator theory. This exploration of degenerate versions originated with Carlitz’s work on degenerate Bernoulli and Euler polynomials (see [4,5]). Notably, this area of study has expanded beyond polynomials and numbers to encompass transcendental functions (see [15,16]) and umbral calculus (see [10]). The Stirling numbers of both kinds and the degenerate Stirling numbers of both kinds are generalized as follows (see [11]). Let {\bf{P}}=\{p_{n}(x)\}_{n=0}^{\infty} be a sequence of polynomials such that deg p_{n}(x)=n,\,p_{0}(x)=1 . The Stirling numbers of the second kind associated with \bf{P} , denoted by S_{2}(n,k;{\bf{P}}) , are defined as the coefficients in the expansion of p_{n}(x) in terms of the falling factorials (x)_{k} : p_{n}(x)=\sum_{k=0}^{n}S_{2}(n,k;{\bf{P}})(x)_{k} . Note that S_{2}(n,k;{\bf{P}}) reduces to the ordinary Stirling numbers of the second kind, S_{2}(n,k) , when {\bf{P}}=\{x^{n}\}_{n=0}^{\infty} (see ( 1 )), and to the degenerate Stirling numbers of the second kind, S_{2,\lambda}(n,k) , when {\bf{P}}=\{(x)_{n,\lambda}\}_{n=0}^{\infty} (see ( 3 )). The Stirling numbers of the first kind associated with \bf{P} , denoted by S_{1}(n,k;{\bf{P}}) , are defined by the expansion (x)_{n}=\sum_{k=0}^{n}S_{1}(n,k;{\bf{P}})p_{k}(x) . The Stirling numbers of both kinds associated with \bf{P} have been investigated using umbral calculus (see [11]). Analogously, the central factorial numbers of both kinds associated with \bf{P} can be studied by replacing the falling factorials (x)_{n} with the central factorials x^{[n]} , where x^{[n]}=x(x+\frac{1}{2}n-1)_{n-1},\,\,(n\geq 1),\,\,x^{[0]}=1 , (see [12]). Let Y be a random variable whose moment generating function exists in some neighborhood of the origin (see ( 8 )). Recently, Adell and Lekuona introduced the probabilistic Stirling numbers of the second kind associated with Y , denoted by S_{2}^{Y}(n,k) (see [1,3]). Furthermore, Adell and Bényi defined the probabilistic Stirling numbers of the first kind associated with Y , denoted by s_{Y}(n,k) , using the cumulant generating function (see [2]). However, S_{2}^{Y}(n,k) and s_{Y}(n,k) do not satisfy orthogonality and inverse relations. In addition, s_{Y}(n,k) does not reduce to the classical Stirling numbers of the first kind, S_{1}(n,k) , when Y=1 . We address these shortcomings by redefining the probabilistic Stirling numbers of the first kind, denoting them by S_{1}^{Y}(n,k) . Our redefined S_{1}^{Y}(n,k) and S_{2}^{Y}(n,k) do satisfy orthogonality and inverse relations (see Proposition 1.1), and S_{1}^{Y}(n,k) reduces to S_{1}(n,k) , when Y=1 . We also consider the probabilistic degenerate Stirling numbers of the second kind associated with Y , denoted by S_{2,\lambda}^{Y}(n,k) (see ( 14 ) and [21]), and the probabilistic degenerate Stirling numbers of the first kind associated with Y , denoted by S_{1,\lambda}^{Y}(n,k) (see ( 15 )). S_{2,\lambda}^{Y}(n,k) and S_{1,\lambda}^{Y}(n,k) satisfy orthogonality and inverse relations (see Proposition 1.2) and S_{1,\lambda}^{Y}(n,k) reduces to the standard degenerate Stirling numbers of the first kind, S_{1,\lambda}(n,k) , when Y=1 . However, in [18], S_{1,\lambda}^{Y}(n,k) is defined differently in that work using the degenerate cumulant generating function. Consequently, the S_{1,\lambda}^{Y}(n,k) defined there and S_{2,\lambda}^{Y}(n,k) do not together satisfy orthogonality and inverse relations. In Section 2, we explicitly compute S_{2}^{Y}(n,k),S_{1}^{Y}(n,k),S_{2,\lambda}^{Y}(n,k) , and S_{1,\lambda}^{Y}(n,k) for several discrete and continuous random variables Y (see Theorems 2.1-2.8). Section 3 illustrates the orthogonality relations for the geometric random variable with parameter 0<p<1 . Furthermore, we address the problem of expressing an arbitrary polynomial in terms of the probabilistic Euler polynomials \mathcal{E}_{n}^{Y}(x) and the probabilistic degenerate Euler polynomials \mathcal{E}_{n,\lambda}^{Y}(x) , demonstrating the necessity of inversion relations. Probabilistic extensions of numerous special polynomials and numbers have been studied recently (see, for example, [6,17,19,28,29] and the references therein). The Stirling numbers of the second kind, S_{2}(n,k) , are defined by (1) \displaystyle x^{n}=\sum_{k=0}^{n}S_{2}(n,k)(x)_{k}, \displaystyle\frac{1}{k!}(e^{t}-1)^{k}=\sum_{n=k}^{\infty}S_{2}(n,k)\frac{t^{n}}{n!}, \displaystyle S_{2}(n,k)=\frac{1}{k!}\sum_{j=0}^{k}\binom{k}{j}(-1)^{k-j}j^{n}, where the falling factorials (x)_{n} are given by (x)_{0}=1,\quad(x)_{n}=x(x-1)\cdots(x-n+1),\quad(n\geq 1). The Stirling numbers of the first kind, S_{1}(n,k) , are defined by (2) \displaystyle(x)_{n}=\sum_{k=0}^{n}S_{1}(n,k)x^{k}, \displaystyle\frac{1}{k!}\big(\log(1+t)\big)^{k}=\sum_{n=k}^{\infty}S_{1}(n,k)\frac{t^{n}}{n!}. Let \lambda be any nonzero real number. The degenerate Stirling numbers of the second kind, S_{2,\lambda}(n,k) , are defined by (3) \displaystyle(x)_{n,\lambda}=\sum_{k=0}^{n}S_{2,\lambda}(n,k)(x)_{k}, \displaystyle\frac{1}{k!}(e_{\lambda}(t)-1)^{k}=\sum_{n=k}^{\infty}S_{2,\lambda}(n,k)\frac{t^{n}}{n!}, \displaystyle S_{2,\lambda}(n,k)=\frac{1}{k!}\sum_{j=0}^{k}\binom{k}{j}(-1)^{k-j}(j)_{n,\lambda}, where (x)_{n,\lambda} are the degenerate (or generalized) falling factorials given by (x)_{0,\lambda}=1,\ (x)_{n,\lambda}=x(x-\lambda)\cdots(x-(n-1)\lambda),\ (n\geq 1), and e_{\lambda}^{x}(t) are the degenerate exponentials defined by (4) e_{\lambda}^{x}(t)=(1+\lambda t)^{\frac{x}{\lambda}}=\sum_{n=0}^{\infty}(x)_{n,\lambda}\frac{t^{n}}{n!},\quad e_{\lambda}(t)=e_{\lambda}^{1}(t),\quad(\mathrm{see}\ [9,17,18,21]). Note here that \lim_{\lambda\rightarrow 0}e_{\lambda}^{x}(t)=e^{xt} . The degenerate Stirling numbers of the first kind, S_{1,\lambda}(n,k) , are given by (5) \displaystyle(x)_{n}=\sum_{k=0}^{n}S_{1,\lambda}(n,k)(x)_{k,\lambda}, \displaystyle\frac{1}{k!}\big(\log_{\lambda}(1+t)\big)^{k}=\sum_{n=k}^{\infty}S_{1,\lambda}(n,k)\frac{t^{n}}{n!}, where \log_{\lambda}(t) are the degenerate logarithm defined by (6) \log_{\lambda}(t)=\frac{1}{\lambda}\big(t^{\lambda}-1\big). Note here that the degenerate exponential e_{\lambda}(t) in ( 4 ) and the degenerate logarithm \log_{\lambda}(t) in ( 6 ) satisfy (7) e_{\lambda}\big(\log_{\lambda}(t)\big)=\log_{\lambda}\big(e_{\lambda}(t)\big)=t. Among other things, the Stirling numbers of both kinds and the degenerate Stirling numbers of both kinds satisfy orthogonality and inversion relations (see Propositions 1.1 and 1.2 with Y=1 ). Let Y be a random variable whose moment generating function exists in a neighborhood of the origin: (8) E[e^{Yt}]=\sum_{n=0}^{\infty}E[Y^{n}]\frac{t^{n}}{n!}\,\,\,\mathrm{exists,\,\,for}\,\,|x|<r, for some positive real number r . Let (Y_{j})_{j\geq 1} be a sequence of mutually independent copies of the random variable Y , and let (9) S_{k}=Y_{1}+Y_{2}+\cdots+Y_{k},\quad(k\geq 1),\quad S_{0}=0.

Read the complete text