2

On Wikipedia I came across the following equation for the central binomial coefficients: $$ \binom{2n}{n}=\frac{4^n}{\sqrt{\pi n}}\left(1-\frac{c_n}{n}\right) $$ for some $1/9<c_n<1/8$.

Does anyone know of a better reference for this fact than wikipedia or planet math? Also, does the equality continue to hold for positive real numbers $x$ instead of the integer $n$ if we replace the factorials involved in the definition of the binomial coefficient by Gamma functions?

Eckhard
  • 7,705

3 Answers3

1

It appears to be true for $x > .8305123339$ approximately: $c_x \to 0$ as $x \to 0+$.

Robert Israel
  • 448,999
0

One can find a number of estimates of the central binomial coefficient in the following papers and closely-related references therein.

@article {MR4127765, AUTHOR = {Popov, A. Yu.}, TITLE = {Two-sided estimates of the central binomial coefficient}, JOURNAL = {Chelyab. Fiz.-Mat. Zh.}, FJOURNAL = {Chelyabinski\u{\i} Fiziko-Matematicheski\u{\i} Zhurnal}, VOLUME = {5}, YEAR = {2020}, NUMBER = {1}, PAGES = {56-69}, ISSN = {2500-0101}, MRCLASS = {05A10 (05A15)}, MRNUMBER = {4127765}, DOI = {10.24411/2500-0101-2020-15105}, URL = {https://doi.org/10.24411/2500-0101-2020-15105}, }

@article {MR4127766, AUTHOR = {Tikhonov, I. V. and Sherstyukov, V. B. and Tsvetkovich, D. G.}, TITLE = {Comparative analysis of two-sided estimates of the central binomial coefficient}, JOURNAL = {Chelyab. Fiz.-Mat. Zh.}, FJOURNAL = {Chelyabinski\u{\i} Fiziko-Matematicheski\u{\i} Zhurnal}, VOLUME = {5}, YEAR = {2020}, NUMBER = {1}, PAGES = {70-95}, ISSN = {2500-0101}, MRCLASS = {05A10 (33B15 40A15)}, MRNUMBER = {4127766}, DOI = {10.24411/2500-0101-2020-15106}, URL = {https://doi.org/10.24411/2500-0101-2020-15106}, }

@article {MR1702663, AUTHOR = {Sasv'{a}ri, Zolt'{a}n}, TITLE = {Inequalities for binomial coefficients}, JOURNAL = {J. Math. Anal. Appl.}, FJOURNAL = {Journal of Mathematical Analysis and Applications}, VOLUME = {236}, YEAR = {1999}, NUMBER = {1}, PAGES = {223--226}, ISSN = {0022-247X}, MRCLASS = {26D15 (05A10)}, MRNUMBER = {1702663}, DOI = {10.1006/jmaa.1999.6420}, URL = {https://doi.org/10.1006/jmaa.1999.6420}, }

qifeng618
  • 1,691
0

Let \begin{align*} \mathbb{Z}&=\{0,\pm1,\pm2,\dotsc\}, & \mathbb{N}&=\{1,2,\dotsc\},\\ \mathbb{N}_0&=\{0,1,2,\dotsc\}, & \mathbb{N}_-&=\{-1,-2,\dotsc\}. \end{align*} The extended binomial coefficient $\binom{z}{w}$ for $z,w\in\mathbb{C}$ is defined by \begin{equation}\label{Gen-Coeff-Binom} \binom{z}{w}= \begin{cases} \dfrac{\Gamma(z+1)}{\Gamma(w+1)\Gamma(z-w+1)}, & z\not\in\mathbb{N}_-,\quad w,z-w\not\in\mathbb{N}_-;\\ 0, & z\not\in\mathbb{N}_-,\quad w\in\mathbb{N}_- \text{ or } z-w\in\mathbb{N}_-;\\ \dfrac{\langle z\rangle_w}{w!},& z\in\mathbb{N}_-, \quad w\in\mathbb{N}_0;\\ \dfrac{\langle z\rangle_{z-w}}{(z-w)!}, & z,w\in\mathbb{N}_-, \quad z-w\in\mathbb{N}_0;\\ 0, & z,w\in\mathbb{N}_-, \quad z-w\in\mathbb{N}_-;\\ \infty, & z\in\mathbb{N}_-, \quad w\not\in\mathbb{Z}, \end{cases} \end{equation} where \begin{equation*}%\label{falling-Factorial} \langle\alpha\rangle_n=\prod_{k=0}^{n-1}(\alpha-k) = \begin{cases} \alpha(\alpha-1)\dotsm(\alpha-n+1), & n\in\mathbb{N}\\ 1, & n=0 \end{cases} \end{equation*} is called the falling factorial.

Equation (10) on page 116 in the paper [1] below reads that the double inequality \begin{equation}\label{Merkle-gamma-ineq}\tag{1} 6^x<\frac{\Gamma(2(1+x))}{\Gamma^2(1+x)}<(1+x)3^x \end{equation} holds for $x\in(0,1)$ and its reversed inequality is valid for $x>1$. We can reformulate the double inequality \eqref{Merkle-gamma-ineq} as follows: the double inequality \begin{equation}\label{Gen-Central-Binom-Coeff-bounds-Eq}\tag{2} \frac{6^x}{2x+1}>\binom{2x}{x}>\frac{(x+1)3^x}{2x+1} \end{equation} is valid for $x>1$ and its reversed version holds for $0<x<1$.

By the way, I have verified the following double inequalities:

  • For $(0,+\infty)$, the double inequality \begin{equation}\label{centr-binom-(2x+1)ineq}\tag{3} \frac{2^{2x}}{2x+1}<\binom{2x}{x}<\frac{e^{2x}}{2x+1} \end{equation} is sharp in the sense that the bases $2$ and $e$ cannot be replaced by larger and smaller constants, respectively. For $x\in\bigl(-\frac{1}{2},0\bigr)$, the right hand side inequality in \eqref{centr-binom-(2x+1)ineq} is still valid, but the left hand side inequality in \eqref{centr-binom-(2x+1)ineq} is reversed.
  • For $x\in(0,+\infty)$, the double inequality \begin{equation}\label{binom(2x+1)(x)-ineq}\tag{4} e^x<\binom{2x+1}{x}<4^x \end{equation} is sharp in the sense that the bases $e$ and $4$ cannot be replaced by larger and smaller constants, respectively. For $x\in\bigl(-\frac{3}{2},0\bigr)$, the left hand side inequality in \eqref{binom(2x+1)(x)-ineq} is still valid, but the right hand side inequality in \eqref{binom(2x+1)(x)-ineq} is reversed.

I am looking for a suitable outlet for publishing, among other closely related things, the double inequalities \eqref{centr-binom-(2x+1)ineq} and \eqref{binom(2x+1)(x)-ineq}.

References

  1. Milan Merkle, On log-convexity of a ratio of gamma functions, Univ. Beograd. Publ. Elektrotehn. Fak. Ser. Mat. 8 (1997), 114--119; available online at https://www.jstor.org/stable/43666390.
  2. Feng Qi, Chao-Ping Chen, and Dongkyu Lim, Several identities containing central binomial coefficients and derived from series expansions of powers of the arcsine function, Results in Nonlinear Analysis 4 (2021), no. 1, 57--64; available online at https://doi.org/10.53006/rna.867047.
  3. Yue-Wu Li and Feng Qi, The sum of an alternating series involving central binomial numbers and its three proofs, Journal of the Korea Society of Mathematical Education Series B: The Pure and Applied Mathematics 28 (2021), no. 4, in press.
qifeng618
  • 1,691