q q -analogues of π \pi -formulas due to Ramanujan and Guillera
2025
Read q q -analogues of π \pi -formulas due to Ramanujan and Guillera (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-2509-05897.txt.
Introduction
John M. Campbell constructs new q-analogues of famous series for the reciprocal of pi in a 2025 paper. A q-analogue rewrites a classic formula with a parameter q so the original returns as q approaches one. Ramanujan's 17 formulas for one over pi are landmarks of this area, and Guillera added series for one over pi squared.
Campbell's tool is a method he calls EKHAD-normalization, built from q-polynomial coefficients inside difference equations produced by a q-version of Zeilberger's algorithm. With it he proves four new q-analogues covering three of Ramanujan's formulas, plus q-analogues of Guillera's first two series. A survey of earlier q-analogues supports his claim that the method is truly new, and he shows how it can extend Guillera's pairs further through hypergeometric expansions. The proofs are short to state but rest on deep facts about series.
Catalog record
| YEAR | 2025 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-05897
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
q q -analogues of π \pi -formulas due to Ramanujan and Guillera Abstract The first known q q -analogues for any of the 17 17 formulas for 1 π \frac{1}{\pi} due to Ramanujan were introduced in 2018 by Guo and Liu (J. Difference Equ. Appl. 29:505–513, 2018), via the q q -Wilf–Zeilberger method. Through a “normalization” method, which we refer to as EKHAD -norma-lization, based on the q q -polynomial coefficients involved in first-order difference equations obtained from the q q -version of Zeilberger’s algorithm, we introduce q q -WZ pairs that extend WZ pairs introduced by Guillera (Adv. in Appl. Math. 29:599–603, 2002) (Ramanujan J. 11:41–48, 2006). We apply our EKHAD -normalization method to prove four new q q -analogues for three of Ramanujan’s formulas for 1 π \frac{1}{\pi} along with q q -analogues of Guillera’s first two series for 1 π 2 \frac{1}{\pi^{2}} . Our normalization method does not seem to have been previously considered in any equivalent way in relation to q q -series, and this is substantiated through our survey on previously known q q -analogues of Ramanujan-type series for 1 π \frac{1}{\pi} and of Guillera’s series for 1 π 2 \frac{1}{\pi^{2}} . We conclude by showing how our method can be adapted to further extend Guillera’s WZ pairs by introducing hypergeometric expansions for 1 π 2 \frac{1}{\pi^{2}} . John M. Campbell Dedicated to my hero Jesús Guillera and his mathematical legacy Keywords : Wilf–Zeilberger method; Wilf–Zeilberger pair; difference equation; Ramanujan-type series, q -Pochhammer symbol, q -analogue. 2020 Mathematics Subject Classification : 33F10, 05A10. 1 Introduction The Wilf–Zeilberger (WZ) method [ 22 ] is of great importance in combinatorics, special functions theory, and computer science. One of the most notable developments in the history of WZ theory is given by the discoveries due to Guillera, with a 2002 contribution of Guillera [ 15 ] having led to a breakthrough in the areas of mathematics related to Ramanujan’s series for \frac{1}{\pi} [ 5 , pp. 352–354] [ 24 ] . Guillera applied the WZ method, through the use of Zeilberger’s EKHAD package for the Maple computer algebra system, to introduce and prove the first Ramanujan-type series for \frac{1}{\pi^{2}} . This motivates the development of techniques to build upon Guillera’s applications of the WZ method and the EKHAD package. This provides the main purpose of our paper. A discrete function A(n,k) is said to be hypergeometric if \frac{A(n+1,k)}{A(n,k)} and \frac{A(n,k+1)}{A(n,k)} are rational functions. A pair (F,G) of bivariate, hypergeometric functions F=F(n,k) and G=G(n,k) is called a WZ pair if the discrete difference equation F(n+1,k)-F(n,k)=G(n,k+1)-G(n,k) (1) holds. For a WZ pair (F,G) , there exists a rational function R(n,k) such that G(n,k)=R(n,k)F(n,k) [ 27 ] , and R(n,k) is referred to as the certificate associated with (F,G) . Setting H(n,k):=F(n+1,n+k)+G(n,n+k), (2) the relation \sum_{n=0}^{\infty}G(n,0)=\sum_{n=0}^{\infty}H(n,0) (3) holds under the assumption that both sums in ( 3 ) converge and under the assumption that \lim_{n\to\infty}\sum_{k=0}^{n-1}F(n,k) vanishes [ 28 ] . The relation in ( 3 ) provides a key in Guillera’s derivations of remarkable formulas such as \frac{128}{\pi^{2}}=\sum_{n=0}^{\infty}\frac{(-1)^{n}\binom{2n}{n}^{5}}{2^{20n}}\big(820n^{2}+180n+13\big). (4) The WZ pair (F,G) employed by Guillera to prove ( 4 ) was determined through the EKHAD package and is such that F(n,k)=512\frac{(-1)^{k}}{2^{16n}2^{4k}}\frac{n^{3}}{4n-2k-1}\frac{\binom{2n}{n}^{4}\binom{2k}{k}^{3}\binom{4n-2k}{2n-k}}{\binom{2n}{k}\binom{n+k}{n}^{2}} (5) and G(n,k)=\frac{(4n-2k-1)\left(84kn+10k+120n^{2}+34n+3\right)}{512n^{3}}F(n,k). The problem of determining q -analogues of the WZ pairs applied by Guil-lera [ 15 ] has led us to formulate an “ EKHAD -type” method, which we refer to as EKHAD -normalization, toward extending known WZ pairs. We apply our method to obtain four new q -analogues of three of Ramanujan’s formulas for \frac{1}{\pi} and new q -analogues for the first two known Ramanujan-type formulas for \frac{1}{\pi^{2}} . Devising q -analogues for \pi -formulas is considered to be challenging [ 26 ] , which motivates the q -analogues introduced in this paper. One of these q -analogues is highlighted as a motivating result in Section 1.2 . Prior to this, we review some required preliminaries, as in Section 1.1 . In Section 2.1 , we provide a survey on past formulas that provide q -analogues of Ramanujan-type formulas for \frac{1}{\pi} and Guillera’s formulas for \frac{1}{\pi^{2}} . This survey is followed by an outline, in Section 2.2 , as to our EKHAD -normalization method. This, in turn, is followed by a review, in Section 2.3 , of past methods employed in the derivation of the q -analogues covered in Section 2.1 , and this is followed by a review of the method of WZ seeds recently due to Au [ 1 ] . Our main results are given in Sections 3 – 5 . We conclude in Section 6 by showing how our EKHAD -normalization method can be adapted to obtain new hypergeometric series for \frac{1}{\pi^{2}} in the spirit of Guillera’s discoveries. 1.1 Preliminaries The \Gamma -function is defined for arguments with a positive real part so that \Gamma(x)=\int_{0}^{\infty}t^{x-1}e^{-t}\,dt and its analytic continuation is obtained through the relation \Gamma(x)=\frac{\Gamma(x+1)}{x} . The Pochhammer symbol may then be defined so that (x)_{n}=\frac{\Gamma(x+n)}{\Gamma(x)} , and we write \left[\begin{matrix}\alpha,\beta,\ldots,\gamma\vskip 2.84526pt\\ A,B,\ldots,C\end{matrix}\right]_{n}=\frac{(\alpha)_{n}(\beta)_{n}\cdots(\gamma)_{n}}{(A)_{n}(B)_{n}\cdots(C)_{n}}. (6) A generalized hypergeometric series may then be defined so that {}_{p}F_{q}\!\!\left[\begin{matrix}a_{1},a_{2},\ldots,a_{p}\vskip 2.84526pt\\ b_{1},b_{2},\ldots,b_{q}\end{matrix}\ \Bigg|\ x\right]=\sum_{n=0}^{\infty}\frac{\left(a_{1}\right)_{n}\left(a_{2}\right)_{n}\cdots\left(a_{p}\right)_{n}}{\left(b_{1}\right)_{n}\left(b_{2}\right)_{n}\cdots\left(b_{q}\right)_{n}}\frac{x^{n}}{n!}, (7) and we define the convergence rate of the {}_{p}F_{q} -series in ( 7 ) as x . A Ramanujan-type series is a series satisfying \sum_{n=0}^{\infty}\left[\begin{matrix}\frac{1}{s},\frac{1}{2},1-\frac{1}{s}\vskip 2.84526pt\\ 1,1,1\end{matrix}\right]_{n}z^{n}(a+bn)=\frac{1}{\pi}, for s\in\{2,3,4,6\} and such that the parameters a , b , and z are real algebraic numbers. The q -Pochhammer symbol may be defined so that (a;q)_{n}=\prod_{k=0}^{n-1}(1-aq^{k}). (8) We then let (a;q)_{\infty}=\prod_{k=0}^{\infty}\left(1-aq^{k}\right), and this allows us to extend the definition in ( 8 ) so as to allow for non-integer values for n , according to the relation (a;q)_{n}=\frac{\left(a;q\right)_{\infty}}{\left(aq^{n};q\right)_{\infty}}, and we write \left[\begin{matrix}\alpha,\beta,\ldots,\gamma\vskip 2.84526pt\\ A,B,\ldots,C\end{matrix}\,\Bigg|\,q\right]_{n}=\frac{\left(\alpha;q\right)_{n}\left(\beta;q\right)_{n}\cdots\left(\gamma;q\right)_{n}}{\left(A;q\right)_{n}\left(B;q\right)_{n}\cdots\left(C;q\right)_{n}}. (9) A fundamental property concerning the q -Pochhammer symbol is such that \lim_{q\to 1}\frac{\left(q^{x};q\right)_{n}}{(1-q)^{n}}=(x)_{n}. (10) We let the q -bracket symbol be such that [n]_{q}=[n]=\frac{1-q^{n}}{1-q} . Also, the q -Gamma function may be defined so that \Gamma_{q}(x)=(1-q)^{1-x}\frac{(q;q)_{\infty}}{(q^{x};q)_{\infty}}, with \lim_{q\to 1^{-}}\Gamma_{q}(x)=\Gamma(x), (11) and, by analogy with ( 6 ) and ( 9 ), we write \Gamma_{q}\left[\begin{matrix}\alpha,\beta,\ldots,\gamma\vskip 2.84526pt\\ A,B,\ldots,C\end{matrix}\right]=\frac{\Gamma_{q}(\alpha)\Gamma_{q}(\beta)\cdots\Gamma_{q}(\gamma)}{\Gamma_{q}(A)\Gamma_{q}(B)\cdots\Gamma_{q}(C)}. Unilateral basic hypergeometric series may be defined so that {}_{j}\phi_{k}\!\!\left[\begin{matrix}a_{1},a_{2},\ldots,a_{j}\vskip 2.84526pt\\ b_{1},b_{2},\ldots,b_{k}\end{matrix}\ \Bigg|\ q;z\right]\\ =\sum_{n=0}^{\infty}\frac{z^{n}}{(q;q)_{n}}\left[\begin{matrix}a_{1},a_{2},\ldots,a_{j}\vskip 2.84526pt\\ b_{1},b_{2},\ldots,b_{k}\end{matrix}\,\Bigg|\,q\right]_{n}\left((-1)^{n}q^{\binom{n}{2}}\right)^{1+k-j}. 1.2 A motivating result One of the main results introduced in this paper is the q -identity \sum_{n=0}^{\infty}(-1)^{n}q^{n^{2}}\frac{(1-q^{-1+2n})(2q^{1+4n}-q^{1+2n}-1)}{1+q^{1+2n}}\\ \times\left[\begin{matrix}\frac{1}{q},q,q,q,q\vskip 2.84526pt\\ q^{2},q^{2},q^{2},q^{1-2n},q^{2+2n}\end{matrix}\,\Bigg|\,q^{2}\right]_{n}=\frac{(1-q)^{2}(1+q)}{q}\left[\begin{matrix}q,q\vskip 2.84526pt\\ q^{2},q^{3}\end{matrix}\,\Bigg|\,q^{2}\right]_{\frac{1}{2}} that we have obtained through our EKHAD -normalization method described in Section 2.2 below. By dividing both sides of the above equality by (1-q)^{2} and by then letting q approach 1 , and by then applying the limiting relation in ( 10 ), we obtain the series -\frac{1}{2}\sum_{n=0}^{\infty}(-1)^{n}(2n-1)(6n+1)\left[\begin{matrix}-\frac{1}{2},\frac{1}{2},\frac{1}{2},\frac{1}{2},\frac{1}{2}\vskip 2.84526pt\\ 1,1,1,\frac{1}{2}-n,1+n\end{matrix}\right]_{n}, and this may be rewritten as \frac{1}{2}\sum_{n=0}^{\infty}\left(\frac{1}{4}\right)^{n}\left[\begin{matrix}\frac{1}{2},\frac{1}{2},\frac{1}{2}\vskip 2.84526pt\\ 1,1,1\end{matrix}\right]_{n}(6n+1), (12) and, as shown by Ramanujan [ 5 , pp. 352–354] [ 24 ] in an equivalent way, the hypergeometric series in ( 12 ) reduces to \frac{2}{\pi} . We obtain the same evaluation by dividing the right-hand side of the above q -identity by (1-q)^{2} and by again letting q approach 1 , according to the limiting property of the q -Gamma function shown in ( 11 ). We prove the above q -series identity in Section 5 , and our main results are summarized in Table 1 . Constant Convergence rate Guillera’s WZ pair New q -analogue \frac{1}{\pi} -\frac{1}{4} [ 15 ]