A two-parameter deformation of the gamma function, associated functions, and some related inequalities

Anton Asare-Tuah, Emmanuel Djabang, Eyram K. A Schwinger, Benoît Florent Sehba, Ralph A. Twum · 2025

Read A two-parameter deformation of the gamma function, associated functions, and some related inequalities by Anton Asare-Tuah, Emmanuel Djabang, Eyram K. A Schwinger, Benoît Florent Sehba, Ralph A. Twum (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-2507-08888.txt.

Introduction

Anton Asare-Tuah and co-authors introduce a new math function in this 2025 paper. The function is a changed form of the classic gamma function, altered by two settings that can be tuned. The gamma function extends the idea of factorials beyond whole numbers and shows up across math and physics.

The authors study how this new function behaves as its settings change. They show it covers some known gamma-like functions as special cases. They also prove bounds and other results for linked functions, such as the beta function and forms built from the gamma function's rate of change. The work adds a flexible tool for later analytic work.

Catalog record

CREATORAnton Asare-Tuah, Emmanuel Djabang, Eyram K. A Schwinger, Benoît Florent Sehba, Ralph A. Twum
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-08888

Browse the full library · About the Tower of Babel

Full text

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

A two-parameter deformation of the gamma function, associated functions, and some related inequalities Anton Asare-Tuah, Emmanuel Djabang, Eyram K. A Schwinger, Benoît Florent Sehba, Ralph A. Twum Abstract . In this paper, we introduce a new two-parameter deformation of the Gamma function that generalizes some existing Gamma-type functions in the literature. We study properties of this function and associated Beta, Psi and Zeta functions. We also prove some inequalities for the corresponding Beta function and Polygamma functions. 2010 Mathematics Subject Classification Primary 26A48, 33B15; Secondary 26D0733C47. Dedicated to Dr. Douglas Adu-Gyamfi in honor of his 80th birthday. Hodie te celebramus, pastor senex. 1. Introduction The aim of this paper is to introduce a new two-parameter gamma function, which is another deformation of the usual gamma function and related functions, and to prove some inequalities. There are some deformations of the usual Gamma function in the literature that have been intensively studied (see [ 5 , 8 , 9 , 11 , 12 , 13 , 16 , 17 , 18 , 19 , 20 ] ). Our proposed deformation generalizes two existing deformations in the literature. Let us recall these two deformations here: • The k -Gamma function of [ 5 ] is defined by \Gamma_{k}(x)=\lim_{n\to\infty}\frac{n!k^{n}(nk)^{\frac{x}{k}-1}}{(x)_{n,k}},\quad k>0,x\in\mathbb{C}\setminus k\mathbb{Z}^{-}. • The \nu -analogue of the Gamma function was introduced and studied in [ 9 ] and is defined by \Gamma^{\nu}(x)=\lim_{n\to\infty}\frac{n!\nu^{n}\left(\frac{n}{\nu}\right)^{\frac{x}{\nu}-1}}{(x)_{n,\nu}},\quad k>0,x\in\mathbb{C}\setminus\nu\mathbb{Z}^{-}. Here, (x)_{n,k}=x(x+k)\cdots(x+(n-1)k) is the so-called Pochhammer k -symbol. Deformations of the classical Gamma function have applications in quantum mechanics and physics, combinatorics, special functions, and orthogonal polynomials (see, for example, [ 4 , 5 , 6 , 7 , 8 ] and the references therein). In this note, we introduce the (k,\nu) -Gamma function as follows: (1) \Gamma_{k,\nu}(x)=\lim_{n\to\infty}\frac{n!(k\nu)^{n}\left(\frac{nk}{\nu}\right)^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}},\quad k>0,\nu>0,x\in\mathbb{C}\setminus k\nu\mathbb{Z}^{-}. One clearly has that \Gamma_{1,\nu}=\Gamma^{\nu}\quad\text{and}\quad\Gamma_{k,1}=\Gamma_{k}. In this work, the (k,\nu) Gamma function and the associated Beta function, Psi function, and Zeta function are studied. We provide various properties and study some inequalities. In the next section, we bring some known results that are useful for our journey. In Section 3, we give an equivalent definition of \Gamma_{k,\nu} and several properties. In Section 4, the associated Beta function that we call (k,\nu) Beta function is introduced and some of its properties are studied. In particular, we obtain some inequalities for this function and estimates involving the (k,\nu) Gamma function. In Section 5, we introduce the (k,\nu) Psi function, provide some properties, some associated partial differential equations, and prove some inequalities for the associated polygamma functions. We end this fifth section by introducing the (k,\nu) Zeta function. In Section 6, we compare the bounds obtained for the ratios of the Beta functions in Section 4. The notation \Gamma , B , \Psi ( k=\nu=1 ), will always represent the usual Gamma, Beta and Digamma functions, respectively. For more on these classical functions, we refer the reader to [ 1 ] . 2. preliminaries In this short section, we invoke some known results needed in our presentation. We start with the following. Lemma 2.1 . Let x>0 . Then the following holds. (2) \int_{0}^{1}t^{x-1}(1-t)^{-x}dt=\int_{0}^{\infty}\frac{t^{x-1}}{1+t}dt=\frac{\pi}{\sin(\pi x)}. Proof. The first equality follows from a basic substitution, and the second can be proved using the residues theorem. ∎ The following Chebychev integral inequality will be needed (see [ 10 ] ). Lemma 2.2 . Let f,g,h:I\subseteq\mathbb{R}\to\mathbb{R} be so that h(x)\geq 0 for all x\in I , and the functions h,hf,hg and hfg are integrable on I . If f and g are synchronous (asynchronous) on I , i.e. (3) \left(f(x)-f(y)\right)\left(g(x)-g(y)\right)\geq(\leq)\,0\quad\text{for all}\quad x,y\in I, then (4) \int_{I}h(x)dx\int_{I}h(x)f(x)g(x)dx\geq(\leq)\,\int_{I}h(x)f(x)dx\int_{I}h(x)g(x)dx. The following result can be found in [ 5 ] . It is originally from P. Etingof [ 14 ] . Proposition 2.3 . Let f,g:[a,b]\to\mathbb{R} be smooth functions on (a,b) where a,b\in[0,\infty) . Assume that f reaches a global minimum at a unique point c\in(a,b) and f^{\prime\prime}(c)>0 . Then one has that \int_{a}^{b}g(x)e^{-\frac{f(x)}{h}}dx=\sqrt{h}e^{-\frac{f(c)}{h}}\sqrt{2\pi}\frac{g(c)}{\sqrt{f^{\prime\prime}(c)}}+O(h). 3. The (k,\nu) Gamma function We discuss in this section, equivalent characterizations of the (k,\nu) Gamma function and some of its properties. We recall that for x\in\mathbb{C} , a\in\mathbb{R} and n\in\mathbb{N} , the Pochhammer a-symbol is defined as (5) (x)_{n,a}=x(x+a)\cdots(x+(n-1)a). We also recall our definition of \Gamma_{k,\nu} . Definition 3.1 . For k>0 and \nu>0 , the (k,\nu) Gamma function \Gamma_{k,\nu} is defined as \Gamma_{k,\nu}(x)=\lim_{n\to\infty}\frac{n!(k\nu)^{n}\left(\frac{nk}{\nu}\right)^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}},\quad x\in\mathbb{C}\setminus k\nu\mathbb{Z}^{-}. We have the following basic facts. Proposition 3.2 . Let k,l,\nu,\mu>0, . For x\in\mathbb{C}\setminus(k\nu\mathbb{Z}^{-}\cap l\mu\mathbb{Z}^{-}) , and n\in\mathbb{N} , the following hold. (a) (x)_{n,l\mu}=\left(\frac{l\mu}{k\nu}\right)^{n}\left(\frac{k\nu x}{l\mu}\right)_{n,k\nu} . (b) \Gamma_{l,\mu}(x)=\left(\frac{l\nu}{k\mu}\right)^{\frac{x}{l\mu}-1}\Gamma_{k,\nu}\left(\frac{k\nu x}{l\mu}\right) . Let us prove the following equivalent definition of \Gamma_{k,\nu} . Proposition 3.3 . Let k,\nu>0 . For any x\in\mathbb{C} with \Re(x)>0 , we have (6) \Gamma_{k,\nu}(x)=\int_{0}^{\infty}e^{-t}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}dt. Proof. We recall that e^{-t}=\lim_{n\to\infty}\left(1-\frac{t}{n}\right)^{n}. Using for example the Monotone Convergence Theorem, we obtain \displaystyle\int_{0}^{\infty}e^{-t}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}dt \displaystyle= \displaystyle\lim_{n\to\infty}\int_{0}^{n}\left(1-\frac{t}{n}\right)^{n}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}dt \displaystyle= \displaystyle\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}-1}\lim_{n\to\infty}\int_{0}^{n}\left(1-\frac{t}{n}\right)^{n}t^{\frac{x}{k\nu}-1}dt. Put A_{n,n-j}(x):=\int_{0}^{n}\left(1-\frac{t}{n}\right)^{n-j}t^{\frac{x}{k\nu}-1}dt,j=0,1,\cdots,n. Then using integration by parts, we obtain \displaystyle A_{n,n-j}(x) \displaystyle= \displaystyle\frac{n-j}{n}\frac{k\nu}{x}\int_{0}^{n}\left(1-\frac{t}{n}\right)^{n-(j+1)}t^{\frac{x}{k\nu}}dt \displaystyle= \displaystyle\frac{n-j}{n}\frac{k\nu}{x}A_{n,n-(j+1)}(x+k\nu). We deduce that (7) \displaystyle A_{n,n}(x)=\frac{n!(k\nu)^{n}}{n^{n}x(x+k\nu)\cdots\left(x+(n-1)k\nu\right)}A_{n,0}(x+nk\nu). But A_{n,0}(x+nk\nu)=\int_{0}^{n}t^{\frac{x}{k\nu}+n-1}dt=\frac{k\nu}{x+nk\nu}n^{\frac{x}{k\nu}+n}. Hence A_{n,n}(x)=\frac{n!(k\nu)^{n}n^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}\left(1+\frac{x}{nk\nu}\right)}. Thus \displaystyle\int_{0}^{\infty}e^{-t}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}dt \displaystyle= \displaystyle\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}-1}\lim_{n\to\infty}A_{n,n}(x) \displaystyle= \displaystyle\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}-1}\lim_{n\to\infty}\frac{n!(k\nu)^{n}n^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}\left(1+\frac{x}{nk\nu}\right)} \displaystyle= \displaystyle\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}-1}\lim_{n\to\infty}\frac{n!(k\nu)^{n}n^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}} \displaystyle= \displaystyle\displaystyle\lim_{n\to\infty}\frac{n!(k\nu)^{n}\left(\frac{nk}{\nu}\right)^{\frac{x}{k\nu}-1}}{(x)_{n,k\nu}} \displaystyle=: \displaystyle\Gamma_{k,\nu}(x). ∎ We have the following identities. Proposition 3.4 . The following hold. (1) \Gamma_{k,\nu}(x+k\nu)=x\nu^{-2}\Gamma_{k,\nu}(x) . (2) \Gamma_{k,\nu}(kx)=k^{\frac{x}{\nu}-1}\Gamma_{1,\nu}(x) . (3) \Gamma_{k,\nu}(\nu x)=\nu^{1-\frac{x}{k}}\Gamma_{k,1}(x) . (4) (x)_{n,k\nu}=\nu^{2n}\frac{\Gamma_{k,\nu}(x+nk\nu)}{\Gamma_{k,\nu}(x)} . (5) \Gamma_{k,\nu}(x)\Gamma_{k,\nu}(k\nu-x)=\frac{\nu}{k}\frac{\pi}{\sin\left(\frac{\pi x}{k\nu}\right)} . (6) \Gamma_{k,\nu}(2x)=\frac{2^{\frac{2x}{k\nu}-1}}{\sqrt{\pi}}\left(\frac{\nu}{k}\right)^{\frac{x}{k\nu}-1}\Gamma_{k,\nu}(x)\Gamma_{k,\nu}(x+\frac{k\nu}{2}) . Proof. Assertions (1), (2) and (3) follow from the definition of \Gamma_{k,\nu} . Assertion (4) follows from assertion (1). We will deduce assertion (6) from a property of the (k,\nu) Beta function later. Let us prove assertion (5). We have \displaystyle\Gamma_{k,\nu}(x)\Gamma_{k,\nu}(k\nu-x) \displaystyle= \displaystyle\int_{0}^{\infty}\int_{0}^{\infty}e^{-(t+u)}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}\left(\frac{k}{\nu}u\right)^{-\frac{x}{k\nu}}dtdu \displaystyle= \displaystyle\int_{0}^{\infty}\int_{t}^{\infty}e^{-v}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}\left(\frac{k}{\nu}(v-t)\right)^{-\frac{x}{k\nu}}dtdv \displaystyle= \displaystyle\int_{0}^{\infty}\left(\int_{0}^{v}\left(\frac{k}{\nu}t\right)^{\frac{x}{k\nu}-1}\left(\frac{k}{\nu}(v-t)\right)^{-\frac{x}{k\nu}}dt\right)e^{-v}dv \displaystyle= \displaystyle\frac{\nu}{k}\int_{0}^{\infty}\left(\int_{0}^{1}\lambda^{\frac{x}{k\nu}-1}\left(1-\lambda\right)^{-\frac{x}{k\nu}}d\lambda\right)e^{-v}dv \displaystyle= \displaystyle\frac{\nu}{k}\frac{\pi}{\sin\left(\frac{\pi x}{k\nu}\right)} where at the last line, we have used Lemma 2.1 . The proof is complete. ∎ Remark 3.5 . We obviously have \Gamma_{k,\nu}(x)=\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}-1}\Gamma\left(\frac{x}{k\nu}\right) . This relation can be combined with the properties of the usual gamma function to derive some properties of \Gamma_{k,\nu} . We do not use this approach here as we wish the paper to be essentially self-contained. We have the following analogue of Stirling’s formula. Theorem 3.6 . Let \Re(x)>0 . Then the following identity holds. (8) \Gamma_{k,\nu}(x+k\nu)(x)=\sqrt{2\pi}\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}}\left(\frac{x}{\nu^{2}}\right)^{\frac{x}{k\nu}}\left(\frac{x}{k\nu}\right)^{1/2}e^{-\frac{x}{k\nu}}+O(\frac{1}{x}). Proof. We note that \Gamma_{k,\nu}(x+k\nu)=\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}}\int_{0}^{\infty}t^{\frac{x}{k\nu}}e^{-t}dt. Now, \displaystyle\int_{0}^{\infty}t^{\frac{x}{k\nu}}e^{-t}dt \displaystyle= \displaystyle\int_{0}^{\infty}e^{\frac{x}{k\nu}\ln(t)-t}dt \displaystyle= \displaystyle\int_{0}^{\infty}e^{-x\left(\frac{t}{x}-\frac{1}{k\nu}\ln(t)\right)}dt. Put f(t)=\frac{t}{x}-\frac{1}{k\nu}\ln(t) and g(t)=1 . Then f^{\prime}(t)=\frac{1}{x}-\frac{1}{k\nu t} and vanishes only at t=\frac{x}{k\nu} . One easily checks that f has a global minimum at c=\frac{x}{k\nu} . Also, we have that f^{\prime\prime}(c)=\frac{1}{k\nu t^{2}}>0. Take h=\frac{1}{x} and note that f(c)=\frac{1}{k\nu}-\frac{1}{k\nu}\ln\left(\frac{x}{k\nu}\right) and f^{\prime\prime}(c)=\frac{k\nu}{x^{2}} . It follows from Proposition 2.3 that \displaystyle\int_{0}^{\infty}e^{-x\left(\frac{t}{x}-\frac{1}{k\nu}\ln(t)\right)}dt \displaystyle= \displaystyle\frac{1}{\sqrt{x}}e^{-x\left(\frac{1}{k\nu}-\frac{1}{k\nu}\ln\left(\frac{x}{k\nu}\right)\right)}\sqrt{2\pi}\frac{x}{\sqrt{k\nu}}+O(\frac{1}{x}) \displaystyle= \displaystyle\sqrt{2\pi}\left(\frac{x}{\nu^{2}}\right)^{\frac{x}{k\nu}}\left(\frac{x}{k\nu}\right)^{1/2}e^{-\frac{x}{k\nu}}+O(\frac{1}{x}). The proof is complete. ∎ Combining the above with the first identity in Proposition 3.4 , we obtain the following. Corollary 3.7 . Let \Re(x)>0 . Then the following identity holds. (9) \Gamma_{k,\nu}(x)=\sqrt{2\pi}\left(\frac{k}{\nu}\right)^{\frac{x}{k\nu}}\left(\frac{x}{\nu^{2}}\right)^{\frac{x}{k\nu}-1}\left(\frac{x}{k\nu}\right)^{1/2}e^{-\frac{x}{k\nu}}+O(\frac{1}{x}). We also have the following formula for the reciprocal of the (k,\nu) Gamma function. Proposition 3.8 . Let k,\nu>0 . We have (10) \frac{1}{\Gamma_{k,\nu}(x)}=\nu^{\frac{x}{k\nu}-1}k^{-\frac{x}{k\nu}}\frac{x}{\nu}e^{\frac{\gamma x}{k\nu}}\prod_{j=1}^{\infty}\left(1+\frac{x}{jk\nu}\right)e^{-\frac{x}{jk\nu}}. Proof. We have from the definition of \Gamma_{k,\nu}(x) that \displaystyle\frac{1}{\Gamma_{k,\nu}(x)} \displaystyle= \displaystyle\lim_{n\to\infty}\left(\frac{\nu}{nk}\right)^{\frac{x}{k\nu}-1}\frac{x(x+k\nu)\cdots\left(x+(n-1)k\nu\right)}{(k\nu)(2k\nu)\cdots\left((n-1)k\nu\right)(nk\nu)} \displaystyle= \displaystyle\lim_{n\to\infty}\left(\frac{\nu}{nk}\right)^{\frac{x}{k\nu}-1}\frac{x}{nk\nu}\prod_{j=1}^{n-1}\left(1+\frac{x}{jk\nu}\right) \displaystyle= \displaystyle\lim_{n\to\infty}\left(\frac{\nu}{k}\right)^{\frac{x}{k\nu}-1}\frac{x}{k\nu}\left(1-\frac{1}{n}\right)^{\frac{x}{k\nu}}\left(\frac{1}{n-1}\right)^{\frac{x}{k\nu}}\prod_{j=1}^{n-1}\left(1+\frac{x}{jk\nu}\right) \displaystyle= \displaystyle\lim_{n\to\infty}\left(\frac{\nu}{k}\right)^{\frac{x}{k\nu}-1}\frac{x}{k\nu}\exp\left({\frac{x}{k\nu}\left(\displaystyle\sum_{j=1}^{n-1}\frac{1}{j}-\ln(n-1)\right)}\right)\prod_{j=1}^{n-1}\left(1+\frac{x}{jk\nu}\right)e^{-\frac{x}{jk\nu}} \displaystyle= \displaystyle\lim_{n\to\infty}\left(\frac{\nu}{k}\right)^{\frac{x}{k\nu}-1}\frac{x}{k\nu}e^{\frac{\gamma x}{k\nu}}\prod_{j=1}^{n-1}\left(1+\frac{x}{jk\nu}\right)e^{-\frac{x}{jk\nu}} \displaystyle= \displaystyle\nu^{\frac{x}{k\nu}-1}k^{-\frac{x}{k\nu}}\frac{x}{\nu}e^{\frac{\gamma x}{k\nu}}\prod_{j=1}^{\infty}\left(1+\frac{x}{jk\nu}\right)e^{-\frac{x}{jk\nu}}. The proof is complete. ∎ We note that the function (k,\nu) Gamma is logarithmically convex. Indeed, applying Hölder’s inequality, one easily obtains that for 0<p,q<1 with p+q=1 , and x,y>0 , (11) \Gamma_{k,\nu}(px+qy)\leq\left(\Gamma_{k,\nu}(x)\right)^{p}\left(\Gamma_{k,\nu}(y)\right)^{q}. Let us prove the following extension of Bohr-Mollerup’s theorem (for the the functions \Gamma(x) and \Gamma_{k,1}(x) , see [ 3 , 5 ] ). Theorem 3.9 . Let f:(0,\infty)\to(0,\infty) , and let k,\nu>0 .

Read the complete text