Generalized product formulas for Whittaker’s functions and a novel class of index transforms

Semyon YAKUBOVICH · 2025

Read Generalized product formulas for Whittaker’s functions and a novel class of index transforms by Semyon YAKUBOVICH (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-2506-05013.txt.

Introduction

Semyon Yakubovich, writing alone, develops product formulas for Whittaker functions in this 2025 math paper. Whittaker functions are special functions that appear in many areas of analysis and physics. The work builds generalized formulas for products of these functions with different indices. From those formulas it constructs a new class of index transforms and finds the matching inversion formulas. Special cases cover transforms built from products of modified Bessel functions and Whittaker functions. The proofs draw on the Kontorovich-Lebedev and Olevskii transforms with complex index.

Catalog record

CREATORSemyon YAKUBOVICH
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE85 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-05013

Browse the full library · About the Tower of Babel

Full text

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

Generalized product formulas for Whittaker’s functions and a novel class of index transforms Semyon YAKUBOVICH Abstract Generalized product formulas and index transforms, involving products of Whittaker’s functions of different indices are established and investigated. The corresponding inversion formulas are found. Particular cases cover index transforms with products of the modified Bessel and Whittaker’s functions. For our goals the Kontorovich-Lebedev and Olevskii transforms of a complex index with nonzero real part are involved. Keywords : Whittaker function, modified Bessel function, Kontorovich-Lebedev transform, Olevskii transform, Laplace transform, Riemann-Liouville fractional integrals, Mellin transform AMS subject classification : 45A05, 44A15, 33C15, 33C10 1 Product formulas for Whittaker’s functions The main goal of this paper is to investigate a class of index transformations with respect to the second pure imaginary index i\tau,\ \tau\in\mathbb{R} of the product of Whittaker’s functions W_{\alpha,i\tau}(x)W_{\beta,i\tau}(x) [2, Vol. III] of the positive argument x>0 and real indices \alpha,\beta\in\mathbb{R} . Historically, index transforms, involving Whittaker’s functions (index Whittaker transforms) appear for the first time in 1964 when Wimp [5] discovered the following reciprocal pair of integral transformations F(\tau)=\int_{0}^{\infty}W_{\mu,i\tau}(x)f(x){dx\over x^{2}},\ \tau>0, [formula] f(x)={1\over\pi^{2}}\int_{0}^{\infty}\tau\sinh(2\pi\tau)\left|\Gamma\left({1\over 2}-\mu+i\tau\right)\right|^{2}W_{\mu,i\tau}(x)F(\tau)d\tau,\ x>0, [formula] where \mu\in\mathbb{R} and \Gamma(z) is the Euler gamma function [2, Vol. III]. As we see in (1.1), (1.2), the corresponding integrals depend upon a second parameter (index) of the Whittaker function and involve it as the variable of integration. This function generalizes the modified Bessel or Macdonald function K_{\nu}(x) [2, Vol. II] via the equality W_{0,i\tau}(2x)=\sqrt{{2x\over\pi}}\ K_{i\tau}(x), [formula] and the pair (1.1), (1.2) reduces to Kontorovich-Lebedev transforms [6]. Generally, these functions are related through the following integral (cf. Entry 2.16.8.4 in [2, Vol. II]) \Gamma\left({1\over 2}-\mu+i\tau\right)\Gamma\left({1\over 2}-\mu-i\tau\right)W_{\mu,i\tau}(x) =2(4x)^{\mu}e^{-x/2}\int_{0}^{\infty}t^{-2\mu}e^{-t^{2}/(4x)}K_{2i\tau}(t)dt,\ x>0,\ {\rm Re}(\mu)<{1\over 2},\ \tau\in\mathbb{R}. [formula] The Whittaker function W_{\alpha,\nu}(x) is, by definition, the solution of Whittaker’s differential equation {d^{2}u\over dx^{2}}+\left({\alpha\over x}-{1\over 4}+{1/4-\nu^{2}\over x^{2}}\right)u=0. [formula] Concerning the product of the Whittaker functions, recently we established the product formula for the case \alpha=\beta,\ \mu=\nu and different arguments (see [3, formula (3.19)) W_{\alpha,\nu}(x)W_{\alpha,\nu}(y)=\int_{0}^{\infty}W_{\alpha,\nu}(\xi)k_{\alpha}(x,y,\xi){d\xi\over\xi^{2}},\quad x,y>0,\ \alpha,\nu\in\mathbb{C}, [formula] where k_{\alpha}(x,y,\xi)=2^{-1-\alpha}\pi^{-1/2}(xy\xi)^{1/2}\exp\left({x\over 2}+{y\over 2}+{\xi\over 2}-{(xy+x\xi+y\xi)^{2}\over 8xy\xi}\right) \times D_{2\alpha}\left({xy+x\xi+y\xi\over(2xy\xi)^{1/2}}\right) and D_{\mu}(z) is the parabolic cylinder function [1, Vol. II]. It gives a departure point for the corresponding convolution theory for the index Whittaker transforms. However, for our further purposes in the sequel we will generalize the product formula (3.26) in [3] in terms of the Gauss hypergeometric function [2, Vol. III] W_{\alpha,i\tau}(x)W_{\alpha,i\tau}(y)={(xy)^{\alpha}\ e^{-(x+y)/2}\over\Gamma(1-2\alpha)}\int_{0}^{\infty}e^{-t}t^{-2\alpha} \times{}_{2}F_{1}\left({1\over 2}-\alpha-i\tau,\ {1\over 2}-\alpha+i\tau;1-2\alpha;-\left({1\over x}+{1\over y}\right)t-{t^{2}\over xy}\right)dt, [formula] where x,y>0,\ \tau\in\mathbb{R},\ \alpha<1/2 . To do this, we will employ the expression of the Whittaker function in terms of the Tricomi function or the Kummer confluent hypergeometric function of the second kind [2, Vol. III] \Psi(a,b;x)=e^{x/2}x^{-b/2}\ W_{b/2-a,(b-1)/2}(x). Indeed, combining with Entry 2.21.2.17 in [2, Vol. III], we derive immediately the following product formula W_{\alpha,i\tau}(x)W_{\beta,i\tau}(y)=(xy)^{1/2+i\tau}\ e^{-(x+y)/2}\ \Psi\left({1\over 2}-\alpha+i\tau,1+2i\tau;x\right) \times\ \Psi\left({1\over 2}-\beta+i\tau,1+2i\tau;y\right)={(xy)^{1/2+i\tau}\ e^{-(x+y)/2}\over\Gamma(1-\alpha-\beta)}\int_{0}^{\infty}e^{-t}t^{-\alpha-\beta} \times(t+x)^{\alpha-1/2-i\tau}(t+y)^{\beta-1/2-i\tau} \times{}_{2}F_{1}\left({1\over 2}-\alpha+i\tau,\ {1\over 2}-\beta+i\tau;1-\alpha-\beta;1-{xy\over(t+x)(t+y)}\right)dt,\quad\alpha+\beta<1. [formula] Meanwhile, Entry 7.3.1.3 in [2, Vol. III] for the Gauss hypergeometric function suggests the following equalities from (1.8) W_{\alpha,i\tau}(x)W_{\beta,i\tau}(y)={(xy)^{\alpha}\ e^{-(x+y)/2}\over\Gamma(1-\alpha-\beta)}\int_{0}^{\infty}e^{-t}t^{-\alpha-\beta}(t+y)^{\beta-\alpha} \times{}_{2}F_{1}\left({1\over 2}-\alpha+i\tau,\ {1\over 2}-\alpha-i\tau;1-\alpha-\beta;-t\left({1\over x}+{1\over y}\right)-{t^{2}\over xy}\right)dt ={(xy)^{\beta}\ e^{-(x+y)/2}\over\Gamma(1-\alpha-\beta)}\int_{0}^{\infty}e^{-t}t^{-\alpha-\beta}(t+x)^{\alpha-\beta} \times{}_{2}F_{1}\left({1\over 2}-\beta+i\tau,\ {1\over 2}-\beta-i\tau;1-\alpha-\beta;-t\left({1\over x}+{1\over y}\right)-{t^{2}\over xy}\right)dt. [formula] In particular, for x=y after a simple substitution it gives the following Laplace integrals for the Gauss hypergeometric functions W_{\alpha,i\tau}(x)W_{\beta,i\tau}(x)={x\ e^{-x}\over\Gamma(1-\alpha-\beta)}\int_{0}^{\infty}e^{-xt}t^{-\alpha-\beta}(t+1)^{\beta-\alpha} \times{}_{2}F_{1}\left({1\over 2}-\alpha+i\tau,\ {1\over 2}-\alpha-i\tau;1-\alpha-\beta;-2t-t^{2}\right)dt ={x\ e^{-x}\over\Gamma(1-\alpha-\beta)}\int_{0}^{\infty}e^{-xt}t^{-\alpha-\beta}(t+1)^{\alpha-\beta} \times{}_{2}F_{1}\left({1\over 2}-\beta+i\tau,\ {1\over 2}-\beta-i\tau;1-\alpha-\beta;-2t-t^{2}\right)dt. [formula] In order to extend the product formula for Whittaker functions with real parameters \alpha,\beta and two complex indices \mu-i\tau,\ \nu+i\tau,\ \mu,\nu,\tau\in\mathbb{R} we employ Entry 2.16.8.4 in [2, Vol. II] to write \Gamma\left({1\over 2}-\alpha+\mu-i\tau\right)\Gamma\left({1\over 2}-\alpha-\mu+i\tau\right)\Gamma\left({1\over 2}-\beta+\nu+i\tau\right)\Gamma\left({1\over 2}-\beta-\nu-i\tau\right) \times x^{\alpha+\beta}e^{1/x}W_{\alpha,\mu-i\tau}\left({1\over x}\right)W_{\beta,\nu+i\tau}\left({1\over x}\right) =4\int_{0}^{\infty}\int_{0}^{\infty}e^{-x(s^{2}+t^{2})}s^{-2\alpha}t^{-2\beta}K_{2(\mu-i\tau)}\left(s\right)K_{2(\nu+i\tau)}\left(t\right)dsdt [formula] under the convergence conditions x>0,\ \tau\in\mathbb{R},,\ \alpha,\beta<1/2,\ 1-2\alpha>2|\mu|,\ 1-2\beta>2|\nu| . Meanwhile, the product of the Macdonald functions can be expressed via relation 2.16.13.4 in [2, Vol. II], and we have 2K_{2(\mu-i\tau)}\left(s\right)K_{2(\nu+i\tau)}\left(t\right)=\int_{-\infty}^{\infty}e^{-2i\tau y+(\mu-\nu)y}\left({s+te^{y}\over se^{y}+t}\right)^{\mu+\nu} \times K_{2(\mu+\nu)}\left(\sqrt{s^{2}+t^{2}+2st\cosh y}\right)dy,\quad s,t>0. [formula] Plugging the right-hand side of the equality (1.12) into (1.11), we interchange the order of integration in the obtained iterated integral via the dominated convergence theorem because \int_{0}^{\infty}\int_{0}^{\infty}e^{-x(s^{2}+t^{2})}s^{-2\alpha}t^{-2\beta}\int_{-\infty}^{\infty}e^{(\mu-\nu)y}\left({s+te^{y}\over se^{y}+t}\right)^{\mu+\nu} \times K_{2(\mu+\nu)}\left(\sqrt{s^{2}+t^{2}+2st\cosh y}\right)dydsdt =2\int_{0}^{\infty}\int_{0}^{\infty}e^{-x(s^{2}+t^{2})}s^{-2\alpha}t^{-2\beta}K_{2\mu}\left(s\right)K_{2\nu}\left(t\right)dsdt<\infty,\ x>0. Therefore, using polar coordinates, equality (1.11) becomes \Gamma\left({1\over 2}-\alpha+\mu-i\tau\right)\Gamma\left({1\over 2}-\alpha-\mu+i\tau\right)\Gamma\left({1\over 2}-\beta+\nu+i\tau\right)\Gamma\left({1\over 2}-\beta-\nu-i\tau\right) \times x^{\alpha+\beta}e^{1/x}W_{\alpha,\mu-i\tau}\left({1\over x}\right)W_{\beta,\nu+i\tau}\left({1\over x}\right)=2\int_{-\infty}^{\infty}e^{-2i\tau y+(\mu-\nu)y} \times\int_{0}^{\pi/2}\int_{0}^{\infty}e^{-xr^{2}}r^{1-2(\alpha+\beta)}\left({\cos\theta+e^{y}\sin\theta\over e^{y}\cos\theta+\sin\theta}\right)^{\mu+\nu} \times K_{2(\mu+\nu)}\left(r\sqrt{1+2\cos\theta\sin\theta\cosh y}\right){drd\theta dy\over\cos^{2\alpha}\theta\ \sin^{2\beta}\theta}. [formula] The inner integral with respect to r is calculated as in (1.11). Hence, changing 1/x on x , we derive from (1.13) the following product formula {\Gamma\left(1/2-\alpha+\mu-i\tau\right)\Gamma\left(1/2-\alpha-\mu+i\tau\right)\Gamma\left(1/2-\beta+\nu+i\tau\right)\Gamma\left(1/2-\beta-\nu-i\tau\right)\over\Gamma\left(1-\alpha-\beta+\mu+\nu\right)\Gamma\left(1-\alpha-\beta-\mu-\nu\right)} \times e^{x}\ W_{\alpha,\mu-i\tau}\left(x\right)W_{\beta,\nu+i\tau}\left(x\right)=\sqrt{x}\int_{-\infty}^{\infty}e^{-2i\tau y+(\mu-\nu)y} \times\int_{0}^{\pi/2}{e^{x(1+2\cos\theta\sin\theta\cosh y)/8}\over\sqrt{1+2\cos\theta\sin\theta\cosh y}}\ \left({\cos\theta+e^{y}\sin\theta\over e^{y}\cos\theta+\sin\theta}\right)^{\mu+\nu} \times W_{\alpha+\beta-1/2,\mu+\nu}\left({x\over 4}\ \left(1+2\cos\theta\sin\theta\cosh y\right)\right){d\theta dy\over\cos^{2\alpha}\theta\ \sin^{2\beta}\theta}. [formula] After simple changes of variables it can be written in a more symmetric form {\Gamma\left(1/2-\alpha+\mu-i\tau\right)\Gamma\left(1/2-\alpha-\mu+i\tau\right)\Gamma\left(1/2-\beta+\nu+i\tau\right)\Gamma\left(1/2-\beta-\nu-i\tau\right)\over\Gamma\left(1-\alpha-\beta+\mu+\nu\right)\Gamma\left(1-\alpha-\beta-\mu-\nu\right)} \times e^{x}\ W_{\alpha,\mu-i\tau}\left(x\right)W_{\beta,\nu+i\tau}\left(x\right)=\sqrt{x}\int_{-\infty}^{\infty}\int_{0}^{\pi/2}e^{-2i\tau y+(\mu-\nu)y} \times{\left(1+\tan^{2}\theta\right)^{\alpha+1/4}\left(1+\cot^{2}\theta\right)^{\beta+1/4}\over\sqrt{\tan\theta+\cot\theta+e^{y}+e^{-y}}}\ \left({1+e^{y}\tan\theta\over e^{y}+\tan\theta}\right)^{\mu+\nu} \times\exp\left({x\left(\tan\theta+\cot\theta+e^{y}+e^{-y}\right)\over 8(\tan\theta+\cot\theta)}\ \right) \times W_{\alpha+\beta-1/2,\mu+\nu}\left({x\left(\tan\theta+\cot\theta+e^{y}+e^{-y}\right)\over 4(\tan\theta+\cot\theta)}\ \right)d\theta dy =\ \sqrt{x}\int_{-\infty}^{\infty}\int_{-\infty}^{\infty}e^{-2i\tau y}\ {(1+e^{2u})^{\alpha+\beta-1/2}\over e^{(2\beta-1/2)u+(\nu-\mu)y}}\left({1+e^{y+u}\over e^{y}+e^{u}}\right)^{\mu+\nu} \times\exp\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\right)W_{\alpha+\beta-1/2,\mu+\nu}\left(x\left(1+{\cosh y\over\cosh u}\right)\ \right){dudy\over\sqrt{\cosh u+\cosh y}}, i.e., finally, {\Gamma\left(1/2-\alpha+\mu-i\tau\right)\Gamma\left(1/2-\alpha-\mu+i\tau\right)\Gamma\left(1/2-\beta+\nu+i\tau\right)\Gamma\left(1/2-\beta-\nu-i\tau\right)\over\Gamma\left(1-\alpha-\beta+\mu+\nu\right)\Gamma\left(1-\alpha-\beta-\mu-\nu\right)} \times e^{x}\ W_{\alpha,\mu-i\tau}\left(x\right)W_{\beta,\nu+i\tau}\left(x\right) =\sqrt{x}\int_{-\infty}^{\infty}\int_{-\infty}^{\infty}e^{-2i\tau y}\ {(1+e^{2u})^{\alpha+\beta-1/2}\over e^{(2\beta-1/2)u+(\nu-\mu)y}}\left({1+e^{y+u}\over e^{y}+e^{u}}\right)^{\mu+\nu} \times\exp\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\right)W_{\alpha+\beta-1/2,\mu+\nu}\left(x\left(1+{\cosh y\over\cosh u}\right)\ \right){dudy\over\sqrt{\cosh u+\cosh y}} [formula] under conditions x>0,\ \tau\in\mathbb{R},\ \alpha,\beta<1/2,1-2\alpha>2|\mu|,\ 1-2\beta>2|\nu| . In particular, letting \alpha,\beta>0,\alpha+\beta=1/2,\ \mu=\nu=0 , we recall (1.3) the get the following representation for the product of Whittaker’s functions of the pure imaginary index |\Gamma\left(1/2-\alpha+i\tau\right)\Gamma\left(1/2-\beta+i\tau\right)|^{2}\ W_{\alpha,i\tau}\left(x\right)W_{\beta,i\tau}\left(x\right)=4\sqrt{\pi}\ xe^{-x} \times\int_{0}^{\infty}\int_{0}^{\infty}\cos(2\tau y)\cosh\left(\left(2\beta-{1\over 2}\right)u\right)\ \exp\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\right) \times K_{0}\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\ \right){dudy\over\sqrt{\cosh u}},\quad x,\alpha,\beta>0,\ \alpha+\beta={1\over 2}. [formula] Finally, for \alpha=\beta=1/4 a double integral for the square of the Whittaker function W_{1/4,i\tau}^{2}\left(x\right) W^{2}_{1/4,i\tau}\left(x\right)={4\sqrt{\pi}\ xe^{-x}\over|\Gamma\left(1/4+i\tau\right)|^{4}}\int_{0}^{\infty}\int_{0}^{\infty}\cos(2\tau y)\ \exp\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\right) \times K_{0}\left({x\over 2}\left(1+{\cosh y\over\cosh u}\right)\ \right){dudy\over\sqrt{\cosh u}},\quad x>0. [formula] Another approach to establish product formulas is to employ the Laplace integral for the Whittaker function [3, formula (3.11)] W_{\alpha,\mu-i\tau}(x)={e^{-x/2}\ x^{\mu-i\tau+1/2}\over\Gamma(1/2-\alpha+\mu-i\tau)}\int_{0}^{\infty}e^{-xs}s^{\mu-i\tau-\alpha-1/2}(1+s)^{\mu-i\tau+\alpha-1/2}ds [formula] under the convergence conditions x>0,\ \mu+1/2>\alpha . Hence we see that the product W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x) can be treated as the Laplace transform of the corresponding convolution. Indeed, we have W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{\mu+\nu+1}\over\Gamma(1/2-\alpha+\mu-i\tau)\Gamma(1/2-\beta+\nu+i\tau)} \times\int_{0}^{\infty}e^{-xs}\int_{0}^{s}t^{\mu-i\tau-\alpha-1/2}(1+t)^{\mu-i\tau+\alpha-1/2}(s-t)^{\nu+i\tau-\beta-1/2}(1+s-t)^{\nu+i\tau+\beta-1/2}dtds. [formula] The inner integral in (1.19) we write in the form \int_{0}^{s}t^{\mu-i\tau-\alpha-1/2}(1+t)^{\mu-i\tau+\alpha-1/2}(s-t)^{\nu+i\tau-\beta-1/2}(1+s-t)^{\nu+i\tau+\beta-1/2}dt =s^{\mu+\nu-\alpha-\beta}\int_{0}^{1}t^{\mu-i\tau-\alpha-1/2}(1+st)^{\mu-i\tau+\alpha-1/2}(1-t)^{\nu+i\tau-\beta-1/2}(1+s(1-t))^{\nu+i\tau+\beta-1/2}dt =s^{\mu+\nu-\alpha-\beta}(1+s)^{\nu+\mu+\beta+\alpha-1}\int_{0}^{1}t^{\mu-i\tau-\alpha-1/2}\left(1-{s(1-t)\over 1+s}\right)^{\mu-i\tau+\alpha-1/2}(1-t)^{\nu+i\tau-\beta-1/2} \times\left(1-{st\over 1+s}\right)^{\nu+i\tau+\beta-1/2}dt=s^{\mu+\nu-\alpha-\beta}(1+s)^{\nu+\mu+\beta+\alpha-1} \times\sum_{n=0}^{\infty}{(1/2-\mu+i\tau-\alpha)_{n}\ s^{n}\over n!\ (1+s)^{n}}\int_{0}^{1}t^{\mu-i\tau-\alpha-1/2}(1-t)^{\nu+i\tau+n-\beta-1/2} \times\left(1-{st\over 1+s}\right)^{\nu+i\tau+\beta-1/2}dt, where (a)_{n} is the Pochhammer symbol and the interchange of the order of integration and summation is due to the dominated convergence theorem. The latter integral is the Euler formula for the Gauss hypergeometric function (see Entry 2.2.6.1 in [2, Vol. I]), and we derive, combining with (1.19) the representation W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{\mu+\nu+1}\over\Gamma(\mu+\nu-\alpha-\beta+1)} \times\int_{0}^{\infty}e^{-xs}s^{\mu+\nu-\alpha-\beta}(1+s)^{\nu+\mu+\beta+\alpha-1} \times\sum_{n=0}^{\infty}{(1/2-\mu+i\tau-\alpha)_{n}\ (1/2+\nu+i\tau-\beta)_{n}\over n!\ (\mu+\nu-\alpha-\beta+1)_{n}}\left({s\over 1+s}\right)^{n} \times\ {}_{2}F_{1}\left({1\over 2}-\nu-i\tau-\beta,\ {1\over 2}+\mu-i\tau-\alpha;\ \mu+\nu-\alpha-\beta+1+n;\ {s\over 1+s}\right)ds. [formula] The latter series is calculated via relation 6.7.1.2 in [2, Vol. III], and we establish the product formula in terms of the Appell F_{3} -function [1], [2] W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{\mu+\nu+1}\over\Gamma(\mu+\nu-\alpha-\beta+1)} \times\int_{0}^{\infty}e^{-xs}s^{\mu+\nu-\alpha-\beta}(1+s)^{\nu+\mu+\beta+\alpha-1}\ F_{3}\left({1\over 2}-\nu-i\tau-\beta,\ {1\over 2}-\mu+i\tau-\alpha,\right. \left.{1\over 2}+\nu+i\tau-\beta,\ {1\over 2}+\mu-i\tau-\alpha;\ \mu+\nu-\alpha-\beta+1;\ {s\over 1+s},{s\over 1+s}\right)ds [formula] under conditions x>0,\ \tau\in\mathbb{R},\ \mu+1/2>\alpha,\ \nu+1/2>\beta. Meanwhile, using the auto-transformation formula for the Gauss hypergeometric function (cf. Entry 7.3.1.4 in [2, Vol. III]), we have {}_{2}F_{1}\left({1\over 2}-\nu-i\tau-\beta,\ {1\over 2}+\mu-i\tau-\alpha;\ \mu+\nu-\alpha-\beta+1+n;\ {s\over 1+s}\right)=(s+1)^{-2\nu-2i\tau-n} \times{}_{2}F_{1}\left({1\over 2}+\nu+i\tau-\beta+n,\ {1\over 2}+2\nu+i\tau-\alpha+n;\ \mu+\nu-\alpha-\beta+1+n;\ {s\over 1+s}\right). Thus equality (1.20) becomes W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{\mu+\nu+1}\over\Gamma(\mu+\nu-\alpha-\beta+1)} \times\int_{0}^{\infty}e^{-xs}s^{\mu+\nu-\alpha-\beta}(1+s)^{\mu-\nu-2i\tau+\beta+\alpha-1} \times\sum_{n=0}^{\infty}{(1/2-\mu+i\tau-\alpha)_{n}\ (1/2+\nu+i\tau-\beta)_{n}\over n!\ (\mu+\nu-\alpha-\beta+1)_{n}}\left({s\over(1+s)^{2}}\right)^{n} \times\ {}_{2}F_{1}\left({1\over 2}+2\nu+i\tau-\alpha+n,\ {1\over 2}+\nu+i\tau-\beta+n;\ \mu+\nu-\alpha-\beta+1+n;\ {s\over 1+s}\right)ds. [formula] The series in (1.22) is calculated in [2, Vol.III], Entry 6.7.1.11 in terms of the Appell F_{1} -function, and we derive the product formula W_{\alpha,\mu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{\mu+\nu+1}\over\Gamma(\mu+\nu-\alpha-\beta+1)} \times\int_{0}^{\infty}e^{-xs}s^{\mu+\nu-\alpha-\beta}(1+s)^{\mu-\nu-2i\tau+\beta+\alpha-1}\ F_{1}\left({1\over 2}+\nu+i\tau-\beta,\ {1\over 2}-\mu+i\tau-\alpha,\right. \left.\ 2\nu+\mu;\ \mu+\nu-\alpha-\beta+1;\ {s\over 1+s},\ {s^{2}\over(1+s)^{2}}\right)ds. [formula] It is worth to mention an interesting particular case 2\nu=-\mu , since relation 6.7.1.10 in [2, Vol. III] suggests a more simple form of the previous product formula. Precisely, we obtain W_{\alpha,-2\nu-i\tau}(x)W_{\beta,\nu+i\tau}(x)={e^{-x}\ x^{1-\nu}\over\Gamma(1-\nu-\alpha-\beta)} \times\int_{0}^{\infty}e^{-xs}s^{-\nu-\alpha-\beta}(1+s)^{-3\nu-2i\tau+\beta+\alpha-1} \times\ {}_{2}F_{1}\left({1\over 2}+\nu+i\tau-\beta,\ {1\over 2}+2\nu+i\tau-\alpha;\ 1-\nu-\alpha-\beta;\ {s^{2}+2s\over(1+s)^{2}}\right)ds. [formula] Using the same as in (1.9) transformation formula for the Gauss

Read the complete text