The Apéry numbers, the Almkvist-Zudilin numbers and new series for \(1/\pi\) (Q843055): Difference between revisions

Let \(\alpha_n\) denote the Apéry number \[ \alpha_n := \sum_{j=0}^n \binom{n}{j}^2\binom{n+j}{j}^2. \] In (apparently) unpublished work, T. Sato gave the following series for \(1/\pi\): \[ \frac{1}{\pi} = (72\sqrt{15} - 160\sqrt{3})\sum_{n \geq 0} \alpha_n \left(\frac{1}{2} - \frac{3\sqrt{5}}{20} + n \right) \left(\frac{1-\sqrt{5}}{2}\right)^{12n}. \] Later \textit{H. H. Chan, S. H. Chan} and \textit{Z. Liu} [Adv. Math. 186, No. 2, 396--410 (2004; Zbl 1122.11087)] gave a similar formula in terms of the ``Domb numbers'' \[ \tilde{\beta}_n := \sum_{j =0}^n \binom{n}{j}^2\binom{2j}{j}\binom{2(n-j)}{n-j}, \] namely \[ \frac{8}{\sqrt{3}\pi} = \sum_{n \geq 0}\tilde{\beta}_n(1+5n)\left(\frac{1}{64}\right)^n. \] In this paper the authors prove the two formulas \[ \frac{3\sqrt{3}}{2\pi} = \sum_{n \geq 0} \gamma_n(4n+1)\left(\frac{1}{81}\right)^n \] and \[ \frac{3\sqrt{3}}{\pi} = \sum_{n \geq 0} \gamma_n(4n+1)\left(\frac{-1}{27}\right)^n, \] where \(\gamma_n\) is the ``Almkvist-Zudilin number'', \[ \gamma_n := \sum_{j=0}^n (-1)^{n-j}\frac{3^{n-3j}(3j)!}{(j!)^3}\binom{n}{3j}\binom{n+j}{j}. \] There are similarities between the present methods and those of previous studies, but here the authors argue in greater generality which allows them to empirically discover many more formulas like those above. Twenty-one such formulas are presented (generally without proof), nine involving \(\tilde{\beta}_n\) and six involving each of \(\alpha_n\) and \(\gamma_n\).