Previous Article in Journal
An Innovative Grey Approach for Group Multi-Criteria Decision Analysis Based on the Median of Ratings by Using Python
Previous Article in Special Issue
B-Fredholm Spectra of Drazin Invertible Operators and Applications

Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

# Explicit Formulas for Some Infinite 3F2(1)-Series

by
Kwang-Wu Chen
Department of Mathematics, University of Taipei, No. 1, Ai-Guo West Road, Taipei 10048, Taiwan
Axioms 2021, 10(2), 125; https://doi.org/10.3390/axioms10020125
Submission received: 23 April 2021 / Revised: 13 June 2021 / Accepted: 14 June 2021 / Published: 19 June 2021

## Abstract

:
We establish two recurrence relations for some Clausen’s hypergeometric functions with unit argument. We solve them to give the explicit formulas. Additionally, we use the moments of Ramanujan’s generalized elliptic integrals to obtain these recurrence relations.
MSC:
33C20; 33C75

## 1. Introduction

The (generalized) hypergeometric function [1] is defined to be the complex analytic function
where
$( α ) n = Γ ( α + n ) Γ ( α ) = α ( α + 1 ) ⋯ ( α + n − 1 )$
denotes the The Pochhammer symbol, $Γ ( s )$ is the Euler’s $Γ$-function, p is a non-negative integer, the complex numbers $a i$, $b j$ are called, respectively, the numerator and denominator parameters, and z is called the variable. The denominator parameters are not allowed to be zero or negative integers ($b j ∉ Z ≤ 0$). If the numerator parameters $a i ∈ Z ≤ 0$, then the series $p + 1 F p$ reduces to a finite sum. The series $p + 1 F p$ converges when $| z | < 1$ for all choices of $a i$, $b j$. If $z = 1$, the series converges for
$ℜ ( b 1 + b 2 + ⋯ + b p − a 1 − a 2 − ⋯ − a p + 1 ) > 0 .$
The case $p + 1 F p$ when $p = 1$ is called the Gauss hypergeometric function. The following well-known and celebrated summation formula for $2 F 1 ( 1 )$ is due to Gauss:
Another interesting formula for $3 F 2 ( 1 )$ is due to Ramanujan:
where n is a positive integer and which is obtained by replacing n by $n − 1$ in Entry 29(b) in ([2], p. 39). This formula was stated without proof by Ramanujan in his first letter to Hardy. There are numerous hypergeometric series identities in mathematical literature (see [3,4]). The evaluation of the hypergeometric sum $3 F 2 ( 1 )$ (the Clausenian hypergeometric function with unit argument) is of ongoing interest, since it appears ubiquitously in many physics and statistics problems [5,6,7]. The q-extension of the 3F2(1)-series is also very interesting and has been studied by many researchers; for example, see [8] and the references therein.
Recently, Asakura et al. [9] proved that
where $B ( a , b ) = Γ ( a ) Γ ( b ) / Γ ( a + b )$ is the beta function and the right hand side denotes the $Q ¯$-linear subspace of $C$ generated by 1, $2 π i$, and $log ( α )$’s, $α ∈ Q ¯ ×$ under some conditions on $a , b , q ∈ Q ∖ Z$.
To obtain an explicit description of Equation (3) has not been completed except some cases. Asakura, Yabu [10] evaluated the cases with $a = 1 / 6$, $b = 5 / 6$ for the examples of their works. For example,
and
where
They list all the explicit values of the cases $q = 1 2$, $i 3$, $j 4$, $k 5$, where $i ∈ { 1 , 2 }$, $j ∈ { 1 , 2 , 3 }$, and $k ∈ { 1 , 2 , 3 , 4 }$ by applying their method to the elliptic fibration $y 2 = 2 x 3 − 3 x 2 + t ℓ$ where $ℓ = 2 , 3 , 4 , 5$, respectively. Motivated by their works, it is interesting to give an explicit formula for the corresponding general form. In this paper, we aim to give an explicit formula for
where n is an arbitrary integer and $q = 1 2$, $i 3$, $j 4$, where $i ∈ { 1 , 2 }$, and $j ∈ { 1 , 2 , 3 }$. For the sake of brevity and our convenience, we sometimes will denote as $F ( x )$. For example, for any non-negative integer n, we have the following explicit formulas:
where $A , B$ is stated in Equation (6).
For $0 ≤ s < 1 2$ and $0 ≤ k ≤ 1$, let
be Ramanujan’s generalized elliptic integral of the first kind of order s. The moment $K n , s$ is given by
$K n , s = ∫ 0 1 k n K s ( k ) d k ,$
where n is a real number. Borwein et al. ([11], Theorem 2) proved that for $0 ≤ s < 1 2$,
Thus, our hypergeometric series $F ( x )$ can be got by setting $s = 1 / 3$, and we have
In the last section, we will use the moments of Ramanujan’s generalized elliptic integral to give another method of obtaining the explicit evaluations.
The organization of this paper is as follows. In Section 2, we give some preliminaries. We provide two recurrence relations for the hypergeometric series $F ( x )$. Then, we solve these recurrence relations to obtain explicit evaluations of the hypergeometric series $F ( n ± q )$ for $n ∈ Z$ and $q = 1 , 1 / 2 , 1 / 3 , 2 / 3 , 1 / 4 , 3 / 4$ in Section 3. In Section 4, we list the explicit forms of $F ( n ± q )$ for $n = ± 1 , ± 2 , ± 3$. In the final section, we use the moments of Ramanujan’s generalized elliptic integral to give another method of obtaining the same evaluations.

## 2. Preliminaries

We list an explicit formula in ([12], Equation 3.13-(41)) which we need to use later.
where $a ≠ 1$, $b ≠ 1$, and $ℜ ( c − a − b + 1 ) > 0$. Thus, let $a = 1 6 , b = 5 6 , c = 2$ we have
We first prove a useful lemma.
Lemma 1.
Let x be a complex number with $x ∉ Z ≤ 0$. Then
Proof.
We rewrite this hypergeometric series $F ( x )$ as
We use the partial fraction decomposition of
$( n + 1 6 ) ( n + 5 6 ) x ( n + 1 ) 2 ( n + 1 + x ) = 5 36 ( n + 1 ) 2 − 36 x + 5 36 x ( n + 1 ) + ( 6 x + 1 ) ( 6 x + 5 ) 36 x ( n + 1 + x )$
to the above identity, we have
The first two hypergeometric series in the right-hand side of the above equation can be evaluated by Equation (7) and the Gauss formula Equation (1):
Substituting these values into the last equation of our $3 F 2 ( 1 )$, we can get the required formula. □
We reverse Equation (8) and get another recurrence relation for our $3 F 2 ( 1 )$.
Lemma 2.
Let x be a complex number with $x ∉ Z ≤ 0$ and $x ≠ 1 6 , 5 6$. Then
We have listed the explicit formulas of
which Asakura, Yabu have obtained in [10], in Equations (4) and (5). Here we list the remaining explicit formulas for $q = 1 / 4 , 3 / 4 , 1 / 5 , 2 / 5 , 3 / 5 , 4 / 5$, which is given in [10].
where
Let $ζ = e 2 π i / 5$, $ζ 20 = e 2 π i / 20$, $α = 1 / 24 10 > 0$,
and
$A j = Γ ( j 5 + 1 6 ) Γ ( j 5 + 5 6 ) Γ ( j 5 ) 2 ,$
where $log ( x )$ takes the principal values,
$log ( x ) = log | x | + arg ( x ) i ( − π < arg ( x ) ≤ π ) .$
Then Asakura, Yabu [10] gave
This formula for $q = k / 5$ is complicated. It can be seen that, although the results in Theorems 1 and 2 can be used to obtain their general formulas, the formulas will be more cumbersome and complicated, so we will not deal with the formula and its general form for $q = k / 5$ in this paper.
We give examples applying Equation (8) and note that
Thus,
where $A , B$ are defined in Equation (6) and $C , D$ are defined in Equation (10).

## 3. Explicit Formulas

We will solve the recurrence relations in Lemmas 1 and 2 as explicit formulas in this section.
Theorem 1.
Let m be a non-negative integer, p be a non-zero complex number such that $m + p + 1$ is not a non-positive integer. Then
Proof.
Consider the hypergeometric series $F ( x )$ with $x > 1$. So we can decomposite $x = m + p$, where $m ∈ N$. Applying the recurrence relation in Lemma 2 a positive integer times, we get
where
$T ( ℓ ) = ∏ j = 1 ℓ ( x + 1 − j ) ( x − j ) ( x − j + 5 6 ) ( x − j + 1 6 ) .$
Using the mathematical induction on the integer , it is easy to prove that the above formula is correct. We use the Pochhammer symbols to rewrite the function T and note that $x = m + p$, we have
$T ( m − k ) = ( p + 1 ) m ( p ) m ( p + 5 6 ) m ( p + 1 6 ) m ( p + 5 6 ) k ( p + 1 6 ) k ( p + 1 ) k ( p ) k .$
Therefore,
Our result is followed by the fact $p ( p + 1 ) m = ( p ) m ( p + m )$. □
Followed by using the similar method to the recurrence relation in Lemma 1, we have the explicit formula for $F ( p − m )$:
Furthermore, we use the Pochhammer symbols at the negative integer index $( a ) − k$, which is defined by
$( a ) − k = Γ ( a − k ) Γ ( a ) = 1 ( a − k ) ( a − k + 1 ) ⋯ ( a − 1 ) = 1 ( a − k ) k .$
Then, the explicit formula $F ( p − m )$ is symmetry to the formula $F ( p + m )$.
Theorem 2.
Let m be a non-negative integer, p be a non-zero complex number such that $p + 1 − m$ is not a non-positive integer. Then
where $( a ) − k$ is defined by
$( a ) − k = Γ ( a − k ) Γ ( a ) = 1 ( a − k ) ( a − k + 1 ) ⋯ ( a − 1 ) = 1 ( a − k ) k .$
In the end of this section we give the explicit formulas $F ( q + n )$ with $q = 1 / 2$, $1 / 3$, $2 / 3$, $1 / 4$, $3 / 4$, and $n ∈ Z$.
Proposition 1.
For any non-negative integer n, we have
where $A , B$ are defined in Equation (6) and $C , D$ are defined in Equation (10).

## 4. Examples

In this section, we list some concrete examples of our results by using Equations (17)–(26) with $n = 1 , 2$. First we indicate that our formula also cover the most well-known formula, for $n ∈ N$,
with the parameters $p = 1$ and $m = n − 1 ∈ N 0$ in Theorem 1. This identity can be derived by using the more general formula about in ([13], Equation (1.7)), or ([3], Equation (16)), where $n , p$ are positive integers. It is note that we use the notations $A , B$ defined in Equation (6) and $C , D$ defined in Equation (10) in the following examples.
Example 1.
The Cases $p = 1 / 2 , − 1 / 2$
Example 2.
The Case $p = 1 / 3 , 2 / 3$
Example 3.
The Case $p = − 1 / 3 , − 2 / 3$
Example 4.
The Cases $p = 1 / 4$ and $p = 3 / 4$
Example 5.
The Cases $p = − 1 / 4$ and $p = − 3 / 4$

## 5. Moments of Ramanujan’s Generalized Elliptic Integrals

For $0 ≤ s < 1 2$ and $0 ≤ k ≤ 1$, let
be Ramanujan’s generalized elliptic integral of the first kind of order s. The moment $K n , s$ is given by
$K n , s = ∫ 0 1 k n K s ( k ) d k ,$
where n is a real number. Borwein et al. ([11], Theorem 2) proved that for $0 ≤ s < 1 2$,
Thus, our hypergeometric series $F ( x )$ can be got by setting $s = 1 / 3$, and we have
The following formula is in ([11], Equation (29)).
where
and $Ψ ( s ) = Γ ′ ( s ) Γ ( s )$ is the digamma function. We substitute $a = 1 / 6$ in the above identity and use the fact
$γ ( 1 6 ) = π + 3 log ( 2 + 3 ) ,$
we have
Therefore, we give another evaluation of Equation (4), which recently was obtained in ([10], Equation (4.1)). In fact, this number is related to the generalized Catalan constant $G 1 / 3$ which was defined in [11].
Borwen et al. ([11], Equation (76)) found a result that followed by Carlson’s Theorem:
$( ( 2 r + 1 ) 2 − 4 s 2 ) K 2 r + 1 , s − ( 2 r ) 2 K 2 r − 1 , s = cos π s .$
Using Equation (48) we transform the above identity into the following
$F ( r ) = ( 6 r + 1 ) ( 6 r + 5 ) 36 r ( r + 1 ) F ( r + 1 ) − 1 2 r π .$
This is exactly the same recurrence relation Equation (8) in Lemma 1. This provides a new approach to our results. Moreover, if we use the formula for odd moments of $K s$ in ([11], Theorem 3), we could get Equation (27):

## Funding

This research was funded by the Ministry of Science and Technology, Taiwan, R.O.C., grant number MOST 109-2115-M-845-001.

Not applicable.

Not applicable.

## Conflicts of Interest

The author declares no conflict of interest.

## References

1. Bailey, W.N. Generalized Hypergeometric Series; Cambridge Tracts in Mathematics and Mathematical Physics, No. 32; Stechert-Hafner Inc.: New York, NY, USA, 1964. [Google Scholar]
2. Berndt, B. Ramanujan’s Notebooks; Springer: New York, NY, USA, 1989; Volume 2. [Google Scholar]
3. Masjed-Jamei, M.; Koepf, W. A new identity for generalized hypergeometric functions and applications. Axioms 2019, 8, 12. [Google Scholar] [CrossRef] [Green Version]
4. Li, N.N.; Chu, W. Non-terminating 3F2-series with unit argument. Integral Transform. Spec. Funct. 2018, 29, 450–469. [Google Scholar] [CrossRef]
5. Milgram, M. On hypergeometric 3F2(1)—A review. arXiv 2010, arXiv:1011.4546v1. [Google Scholar]
6. Shpot, M.A.; Srivastava, H.M. The Clausenian hypergeometric function 3F2 with unit argument and negative integral parameter differences. Appl. Math. Comput. 2015, 259, 819–827. [Google Scholar] [CrossRef] [Green Version]
7. Srivastava, H.M.; Qureshi, M.I.; Jabee, S. Some general series identities and summation theorems for Clausen’s hypergeometric function with negative integer numerator and denominator parameters. J. Nonlinear Convex Anal. 2020, 21, 805–819. [Google Scholar]
8. Srivastava, H.M.; Damjanović, S.L. An elementary proof of a certain transformation for an n-Balanced hypergeometric 3Φ2 Series. Proc. Jpn. Acad. Ser. A Math. Sci. 1988, 64, 198–200. [Google Scholar] [CrossRef]
9. Asakura, M.; Otsubo, N.; Terasoma, T. An algebro-geometric study of special values of hypergeometric functions 3F2. Nagoya Math. J. 2019, 236, 47–62. [Google Scholar] [CrossRef]
10. Asakura, M.; Yabu, T. Explicit logarithmic formulas of special values of hypergeometric functions 3F2. Commun. Contemp. Math. 2020, 22, 1950040. [Google Scholar] [CrossRef] [Green Version]
11. Borwein, D.; Borwein, J.M.; Glasser, M.L.; Wan, J.G. Moments of Ramanujan’s generalized elliptic integrals and extensions of Catalan’s constant. J. Math. Anal. Appl. 2011, 384, 478–496. [Google Scholar] [CrossRef] [Green Version]
12. Luke, Y.L. The Special Functions and Their Approximations; Academic Press, Inc.: San Diego, CA, USA, 1969; Volume 1. [Google Scholar]
13. Miller, A.R.; Paris, R.B. Clausen’s series 3F2(1) with integral parameter differences and transformations of the hypergeometric function 2F2(x). Integral Transform. Spec. Funct. 2012, 23, 21–33. [Google Scholar] [CrossRef]
 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Share and Cite

MDPI and ACS Style

Chen, K.-W. Explicit Formulas for Some Infinite 3F2(1)-Series. Axioms 2021, 10, 125. https://doi.org/10.3390/axioms10020125

AMA Style

Chen K-W. Explicit Formulas for Some Infinite 3F2(1)-Series. Axioms. 2021; 10(2):125. https://doi.org/10.3390/axioms10020125

Chicago/Turabian Style

Chen, Kwang-Wu. 2021. "Explicit Formulas for Some Infinite 3F2(1)-Series" Axioms 10, no. 2: 125. https://doi.org/10.3390/axioms10020125

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.