Skip to content

Advertisement

  • Research
  • Open Access

Some computational formulas related the Riemann zeta-function tails

Journal of Inequalities and Applications20162016:132

https://doi.org/10.1186/s13660-016-1068-2

Received: 14 March 2016

Accepted: 11 April 2016

Published: 4 May 2016

Abstract

In this paper we present two computational formulae for one kind of reciprocal sums related to the Riemann zeta-function at integer points \(s=4,5\), which answers an open problem proposed by Lin (J. Inequal. Appl. 2016:32, 2016).

Keywords

Riemann zeta-functioninequalitycontinued fractiontelescoping methodmultiple-correction method

MSC

11B8311M0611J70

1 Introduction and main results

Let \((a_{k} )_{k\ge1}\) be a strictly increasing positive sequence such that
$$ \sum_{k=1}^{\infty} \frac{1}{a_{k}}< +\infty. $$
(1.1)
Many authors study the computational formula for infinite sums of reciprocal \(a_{k}\),
$$ \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{a_{k}} \Biggr)^{-1} \Biggr\rfloor ,\quad n\in\mathbb{N}, $$
(1.2)
where \(\lfloor x \rfloor\) denotes the integer part of x.
For example, let \((F_{k})\) be the famous Fibonacci sequence: \(F_{k+1}=F_{k}+F_{k-1}\) with the initial values \(F_{0}=0\) and \(F_{1}=1\). Ohtsuka and Nakamura [2] showed that
$$\begin{aligned}& \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{F_{k}} \Biggr)^{-1} \Biggr\rfloor = \textstyle\begin{cases} F_{n-2} &\mbox{if } n\ge2 \mbox{ is even}, \\ F_{n-2}-1 &\mbox{if } n\ge1 \mbox{ is odd}, \end{cases}\displaystyle \\& \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{F_{k}^{2}} \Biggr)^{-1} \Biggr\rfloor = \textstyle\begin{cases} F_{n-1}F_{n}-1 &\mbox{if } n\ge2 \mbox{ is even}, \\ F_{n}F_{n-1} &\mbox{if } n\ge1 \mbox{ is odd}. \end{cases}\displaystyle \end{aligned}$$
Xu and Wang [3] obtained a complex computational formula for \(a_{k}=F_{k}^{3}\).
Zhang and Wang [4] studied this problem for the Pell numbers \(P_{k}\) and showed that
$$\Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{P_{k}} \Biggr)^{-1} \Biggr\rfloor = \textstyle\begin{cases} P_{n-1}+P_{n-2} &\mbox{if } n\ge2 \mbox{ is even}, \\ P_{n-1}+P_{n-2}-1 &\mbox{if } n\ge1 \mbox{ is odd}, \end{cases} $$
where the Pell numbers \(P_{k}\) are defined by \(P_{0}=0\), \(P_{1}=1\), and the recurrence relation \(P_{k+1}=2P_{k}+P_{k-1}\).

For some other results related to recursive sequences, recursive polynomials, and their promotion forms, see [513] and references therein.

Very recently, Lin [1] investigated the related problem for the sequence \(a_{k}=k^{s}\) with integer \(s\ge2\) and showed the following two interesting identities:
$$\begin{aligned} & \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{k^{2}} \Biggr)^{-1} \Biggr\rfloor =n-1, \end{aligned}$$
(1.3)
$$\begin{aligned} & \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{k^{3}} \Biggr)^{-1} \Biggr\rfloor =2n(n-1). \end{aligned}$$
(1.4)
This is an important problem, which has a close relationship with the Riemann zeta-function \(\zeta(s)\). Lin noted that there does not exist an integer-coefficient polynomial \(q(x)\) of degree 3 such that the following identity holds:
$$ \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{k^{4}} \Biggr)^{-1} \Biggr\rfloor =q(n). $$
(1.5)

In [1], Lin declared that giving a precise calculation formula for \((\sum_{k=n}^{\infty}\frac{1}{k^{s}} )^{-1}\) with \(s=4\) is a very complicated problem. In this paper, we tackle this open problem.

Theorem 1

For all integer \(n\ge2\), we have the identity
$$ \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{k^{4}} \Biggr)^{-1} \Biggr\rfloor =-1+4n-5n^{2}+3n^{3} + \biggl\lfloor \frac{(2n+1)(n-1)}{4} \biggr\rfloor . $$
(1.6)

Furthermore, for \(a_{k}=k^{5}\), we also have an analogous computational formula.

Theorem 2

For all integer \(n\ge4\), we have
$$ \Biggl\lfloor \Biggl(\sum_{k=n}^{\infty} \frac{1}{k^{5}} \Biggr)^{-1} \Biggr\rfloor =-5 n + 9 n^{2} - 8 n^{3} + 4 n^{4} + \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor . $$
(1.7)

2 Proof of Theorem 1

Assume that
$$g(n)-g(n+1)< \frac{1}{n^{4}}< f(n)-f(n+1), $$
and \(f(\infty)=g(\infty)=0\). Summing the inequalities from n to ∞, we have
$$ g(n)< \sum_{k=n}^{\infty} \frac{1}{k^{4}}< f(n). $$
(2.1)
These inequalities allow us to study the computational formulas of Theorem 1. The problem of finding the functions \(f(n)\), \(g(n)\) (or \(F(n)\), \(G(n)\) in Section 3) is transformed into solving the finite continued fraction approximation solution of difference equation for ‘large’ n:
$$ y(n)-y(n+1)=\frac{1}{n^{4}}. $$
(2.2)
We will apply the multiple-correction method (see [1416]) and solve it as follows.

Step 1

(The initial correction)

Choosing \(\eta_{0}(n)=\frac{b}{n^{3} + a_{2} n^{2} + a_{1} n + a_{0}}\) and developing the expression \(\eta_{0}(n)-\eta_{0}(n+1)-\frac{1}{n^{4}}\) into power series expansion in \(1/n\), we easily obtain
$$\begin{aligned} &\eta_{0}(n)-\eta_{0}(n+1)- \frac{1}{n^{4}} \\ &\quad=(3b-1)\frac{1}{n^{4}}+(-6 b - 4 a_{2} b) \frac{1}{n^{5}}+\bigl(10 b - 5 a_{1} b + 10 a_{2} b + 5 a_{2}^{2} b\bigr)\frac{1}{n^{6}} \\ &\qquad{}+\bigl(-15 b - 6 a_{0} b + 15 a_{1} b - 20 a_{2} b + 12 a_{1} a_{2} b - 15 a_{2}^{2} b - 6 a_{2}^{3} b\bigr) \frac{1}{n^{7}}+O \biggl(\frac{1}{n^{8}} \biggr). \end{aligned}$$
(2.3)
If \(b=\frac{1}{3}\), \(a_{2}=-\frac{3}{2}\), \(a_{1}=\frac{5}{4}\), \(a_{0}=-\frac{3}{8}\), then we can get the approximation solution
$$g(n)=\frac{b}{n^{3} + a_{2} n^{2} + a_{1} n + a_{0}} $$
of difference equation (2.2), which is the best possible rational approximation solution of such structure as n tends to infinity.

Step 2

(The first correction)

Choose \(\eta_{1}(n)=\frac{b}{n^{3} + a_{2} n^{2} + a_{1} n + a_{0}+\frac{u}{x+v}}\) and developing the expression \(\eta_{1}(n)-\eta_{1}(n+1)-\frac{1}{n^{4}}\) into power series expansion in \(1/n\), we easily obtain
$$ \eta_{1}(n)-\eta_{1}(n+1)- \frac{1}{n^{4}}= \biggl(-\frac{7}{16}-\frac{7}{3} u \biggr) \frac{1}{n^{8}}+\frac{4}{3} (u+2uv)\frac{1}{n^{9}}+O \biggl( \frac{1}{n^{10}} \biggr). $$
(2.4)
If \(u=-\frac{3}{16}\), \(v=-\frac{1}{2}\), then we can get the approximation solution
$$f(n)=\frac{b}{n^{3} + a_{2} n^{2} + a_{1} n + a_{0}+\frac{u}{x+v}} $$
of difference equation (2.2), which has a better approximation rate than \(g(n)\) for ‘large’ n.

So we can get following inequalities necessary in the proofs of our theorems.

Lemma 1

Let
$$ f(n)=\frac{ \frac{8}{3} }{ -1 + 10 n - 12 n^{2} + 8 n^{3}-\frac{3}{2n-1}}. $$
(2.5)
Then, for \(n\ge2\),
$$ f(n)-f(n+1)>\frac{1}{n^{4}}. $$
(2.6)

Proof

We easily check that
$$f(n)-f(n+1)-\frac{1}{n^{4}}=\frac{1}{n^{4} (1 + n) (1 + n + n^{2}) (-1 + 2 n - 2 n^{2} + n^{3}) }. $$
Note that \(-1 + 2 n - 2 n^{2} + n^{3}=(-2 + n)(2 + n^{2})+ 3\), so the above polynomial is positive for \(n\ge2\). Then, for \(n\in\mathbb{N}\),
$$f(n)-f(n+1)-\frac{1}{n^{4}}>0. $$
 □

Lemma 2

Let
$$ g(n)=\frac{8}{3 (-3 + 10 n - 12 n^{2} + 8 n^{3})}. $$
(2.7)
Then, for \(n\in\mathbb{N}\),
$$ g(n)-g(n+1)< \frac{1}{n^{4}}. $$
(2.8)

Proof

We have
$$g(n)-g(n+1)-\frac{1}{n^{4}}=\frac{- 28 n^{2}+9}{n^{4} (-3 + 10 n - 12 n^{2} + 8 n^{3}) (3 + 10 n + 12 n^{2} + 8 n^{3})}, $$
where \(h(n):=-3 + 10 n - 12 n^{2} + 8 n^{3}=(-2 + n)(18 + 4 n + 8 n^{2})+ 33>0\) for \(n\ge2\), and \(h(1)=3>0\). So \(h(n)>0\) for \(n\in\mathbb{N}\). This completes the proof of Lemma 2. □

Proof of Theorem 1

Summing the inequalities of the form
$$g(n)-g(n+1)< \frac{1}{n^{4}}< f(n)-f(n+1) $$
from n to ∞ and noting that \(f(\infty)=g(\infty)=0\), we have
$$ g(n)< \sum_{k=n}^{\infty}\frac{1}{k^{4}}< f(n), \quad n\ge2. $$
(2.9)
Then, for \(n\ge2\),
$$ \frac{3}{8} \biggl(\frac{-3}{2n-1}-3 + 10 n - 12 n^{2} + 8 n^{3} \biggr)< \Biggl(\sum _{k=n}^{\infty}\frac{1}{k^{4}} \Biggr)^{-1} < \frac{3}{8}\bigl(-3 + 10 n - 12 n^{2} + 8 n^{3}\bigr). $$
(2.10)
Note that
$$\frac{3}{8}\bigl(-3 + 10 n - 12 n^{2} + 8 n^{3} \bigr)=3n^{3}-5n^{2}+4n-1 + \biggl(\frac {n(2n-1)}{4}- \frac{1}{8} \biggr) $$
and
$$\frac{3}{8} \biggl(\frac{-3}{2n-1} -3+ 10 n - 12 n^{2} + 8 n^{3} \biggr)=3n^{3}-5n^{2}+4n-1 + \biggl( \frac{n(2n-1)}{4}-\frac{9}{8(2n-1)}-\frac{1}{8} \biggr). $$
For \(n\ge5\), we have \(\frac{9}{8(2n-1)}\le\frac{1}{8}\). Then
$$\begin{aligned} \biggl\lfloor \frac{n(2n-1)}{4}-\frac{1}{8} \biggr\rfloor &\ge \biggl\lfloor \frac{n(2n-1)}{4}-\frac{9}{8(2n-1)}-\frac{1}{8} \biggr\rfloor \ge \biggl\lfloor \frac{n(2n-1)}{4}-\frac{1}{4} \biggr\rfloor \\ &= \biggl\lfloor \frac{(2n+1)(n-1)}{4} \biggr\rfloor \end{aligned}$$
and, for \(n\in\mathbb{N}\),
$$\biggl\lfloor \frac{n(2n-1)}{4}-\frac{1}{8} \biggr\rfloor = \biggl\lfloor \frac{(2n+1)(n-1)}{4} \biggr\rfloor . $$
It follows that, for \(n\ge5\),
$$ \biggl\lfloor \frac{n(2n-1)}{4}-\frac{1}{8} \biggr\rfloor = \biggl\lfloor \frac{n(2n-1)}{4}-\frac{9}{8(2n-1)}-\frac{1}{8} \biggr\rfloor = \biggl\lfloor \frac{(2n+1)(n-1)}{4} \biggr\rfloor . $$
(2.11)
Finally, we note that the above identities hold for \(n=2,3,4\). Combining (2.10) and (2.11), we prove Theorem 1. □

3 Proof of Theorem 2

Similarly to Section 2, by the multiple-correction method we can solve the finite continued fraction approximation solution \(F(n)\), \(G(n)\) of the differential equation
$$ y(n)-y(n+1)=\frac{1}{n^{5}}. $$
(3.1)
So we have the following inequalities.

Lemma 3

Let
$$ F(n)=\frac{9}{-2 - 48 n + 84 n^{2} - 72 n^{3} + 36 n^{4}}. $$
(3.2)
Then, for \(n\ge2\),
$$ F(n)-F(n+1)>\frac{1}{n^{5}}. $$
(3.3)

Proof

$$\begin{aligned} &F(n)-F(n+1)-\frac{1}{n^{5}} \\ &\quad=\frac{-1 + 660 n^{2}}{n^{5} (-1 - 24 n + 42 n^{2} - 36 n^{3} + 18 n^{4}) (-1 +24 n + 42 n^{2} + 36 n^{3} + 18 n^{4})}. \end{aligned}$$
(3.4)
Note that
$$ -1 - 24 n + 42 n^{2} - 36 n^{3} + 18 n^{4}=(n-2) \bigl(60 + 42 n + 18 n^{3}\bigr)+119. $$
(3.5)
Then, for \(n\ge2\), we have
$$F(n)-F(n+1)-\frac{1}{n^{5}}>0. $$
 □

Lemma 4

Let
$$ G(n)=\frac{9}{-1/4 - 48 n + 84 n^{2} - 72 n^{3} + 36 n^{4}}. $$
(3.6)
Then, for \(n\ge5\),
$$ G(n)-G(n+1)< \frac{1}{n^{5}}. $$
(3.7)

Proof

Similarly to the proof of Lemma 3, we have
$$\begin{aligned} &G(n)-G(n+1)-\frac{1}{n^{5}} \\ &\quad=-\frac{1-37\text{,}536 n^{2} +2\text{,}016 n^{4}}{n^{5} (-1 - 192 n + 336 n^{2} - 288 n^{3} + 144 n^{4}) (-1 + 192 n +336 n^{2} + 288 n^{3} + 144 n^{4})}. \end{aligned}$$
Note that
$$1-37\text{,}536 n^{2} +2\text{,}016 n^{4}=(-5 + n) \bigl(64\text{,}320 + 12\text{,}864 n + 10\text{,}080 n^{2} + 2\text{,}016 n^{3}\bigr)+ 321\text{,}601 $$
and
$$-1 - 192 n + 336 n^{2} - 288 n^{3} + 144 n^{4}=(-4 + n) \bigl(5\text{,}760 + 1\text{,}488 n + 288 n^{2} + 144 n^{3}\bigr)+ 23\text{,}039. $$
Then, for \(n\ge5\), the inequality \(G(n)-G(n+1)-\frac{1}{n^{5}}<0\) holds. □

Proof of Theorem 2

We assume that \(n\ge5\) in the following proof. Summing the inequalities of the form
$$G(n)-G(n+1)< \frac{1}{n^{4}}< F(n)-F(n+1) $$
from n to ∞ and noting that \(F(\infty)=G(\infty)=0\), we have
$$\begin{aligned} &{-} 5 n + 9 n^{2} - 8 n^{3} + 4n^{4}+ \biggl(-\frac{2}{9} -\frac{n}{3} + \frac {n^{2}}{3} \biggr) \\ &\quad< \Biggl(\sum_{k=n}^{\infty}\frac{1}{k^{5}} \Biggr)^{-1} \\ &\quad< - 5 n + 9 n^{2} - 8 n^{3} + 4n^{4}+ \biggl(- \frac{1}{36} -\frac{n}{3} + \frac {n^{2}}{3} \biggr). \end{aligned}$$
(3.8)
Next, for \(n\ge3\), we will prove the following identities:
$$ \biggl\lfloor -\frac{2}{9} -\frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor -\frac{1}{36} - \frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor . $$
(3.9)
Since
$$\frac{(n+1)(n-2)}{3}< -\frac{2}{9} -\frac{n}{3} + \frac{n^{2}}{3}< -\frac {1}{36} -\frac{n}{3} + \frac{n^{2}}{3}, $$
it suffices to prove that
$$ \biggl\lfloor -\frac{1}{36} -\frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor . $$
(3.10)
We will consider three cases.
Case 1. If \(n=3m\), \(m\in\mathbb{N}\), then we have
$$\biggl\lfloor -\frac{1}{36} -\frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor =-1-m+3m^{2}. $$
Case 2. If \(n=3m+1\), \(m\in\mathbb{N}\), then we have
$$\biggl\lfloor -\frac{1}{36} -\frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor =-1+m + 3 m^{2}. $$
Case 3. If \(n=3m+2\), \(m\in\mathbb{N}\), then we have
$$\biggl\lfloor -\frac{1}{36} -\frac{n}{3} + \frac{n^{2}}{3} \biggr\rfloor = \biggl\lfloor \frac{(n+1)(n-2)}{3} \biggr\rfloor =3m + 3 m^{2}. $$
This proves that (3.10) holds. Finally, combining (3.8) and (3.9), we prove Theorem 2. □

Declarations

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors’ Affiliations

(1)
Department of Mathematics and Physics, Beijing Institute of Petrochemical Technology, Beijing, P.R. China

References

  1. Lin, X: Some identities related to Riemann zeta-function. J. Inequal. Appl. 2016, 32 (2016) MathSciNetView ArticleMATHGoogle Scholar
  2. Ohtsuka, H, Nakamura, S: On the sum of reciprocal Fibonacci numbers. Fibonacci Q. 46/47, 153-159 (2008/2009) Google Scholar
  3. Xu, Z, Wang, T: The infinite sum of the cubes of reciprocal Fibonacci numbers. Adv. Differ. Equ. 2013, 184 (2013) View ArticleGoogle Scholar
  4. Zhang, W, Wang, T: The infinite sum of reciprocal Pell numbers. Appl. Math. Comput. 218, 6164-6167 (2012) MathSciNetMATHGoogle Scholar
  5. Zhang, H, Wu, Z: On the reciprocal sums of the generalized Fibonacci sequences. Adv. Differ. Equ. 2013, 377 (2013) MathSciNetView ArticleGoogle Scholar
  6. Kilic, E, Arikan, T: More on the infinite sum of reciprocal usual Fibonacci, Pell and higher order recurrences. Appl. Math. Comput. 219, 7783-7788 (2013) MathSciNetMATHGoogle Scholar
  7. Komatsu, T, Laohakosol, V: On the sum of reciprocals of numbers satisfying a recurrence relation of orders. J. Integer Seq. 13, Article ID 10.5.8 (2010) MathSciNetMATHGoogle Scholar
  8. Komatsu, T: On the nearest integer of the sum of reciprocal Fibonacci numbers. In: Proceedings of the Fourteenth International Conference on Fibonacci Numbers and Their Applications. Aportaciones Matematicas Investigacion, vol. 20, pp. 171-184 (2011) Google Scholar
  9. Mansour, T, Shattuck, M: Restricted partitions and q-Pell numbers. Cent. Eur. J. Math. 9, 346-355 (2011) MathSciNetView ArticleMATHGoogle Scholar
  10. Wu, Z, Zhang, H: On the reciprocal sums of higher-order sequences. Adv. Differ. Equ. 2013, 189 (2013) MathSciNetView ArticleGoogle Scholar
  11. Wu, Z, Zhang, W: The sums of the reciprocals of Fibonacci polynomials and Lucas polynomials. J. Inequal. Appl. 2012, 134 (2012) MathSciNetView ArticleMATHGoogle Scholar
  12. Wu, Z, Zhang, W: Several identities involving the Fibonacci polynomials and Lucas polynomials. J. Inequal. Appl. 2013, 205 (2013) MathSciNetView ArticleMATHGoogle Scholar
  13. Anantakitpaisal, P, Kuhapatanakul, K: Reciprocal sums of the tribonacci numbers. J. Integer Seq. 19, Article ID 16.2.1 (2016) MATHGoogle Scholar
  14. Cao, X, Xu, H, You, X: Multiple-correction and faster approximation. J. Number Theory 149, 327-350 (2015) MathSciNetView ArticleMATHGoogle Scholar
  15. Cao, X, You, X: Multiple-correction and continued fraction approximation (II). Appl. Math. Comput. 261, 192-205 (2015) MathSciNetGoogle Scholar
  16. Cao, X, Mortici, C: Multiple-correction and summation of the rational series. Available online at arXiv:1511.00198v1

Copyright

© Xu 2016

Advertisement