Abstract
In this paper, we are going to perform the shuffle products of \(Z_-({m}) = \sum _{a+b=m} (-1)^{b} \zeta (\{1\}^{a},b+2)\) and \(Z_+^\star (n) = \sum _{c+d=n} \zeta ^{\star }(\{1\}^{c},d+2)\) with \(m+n = p\). The resulted shuffle relation is a weighted sum formula stated below:
where the summation \(\sum ^*\) and the weights \(W_{\varvec{\alpha }}(a,b,c)\) are given appropriately. We also give some weighted alternating Euler sums formulas.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
For an r-tuple \(\varvec{\alpha } = (\alpha _{1}, \alpha _{2}, \ldots , \alpha _{r})\) of positive integers with \(\alpha _{r} \ge 2\), a multiple zeta value \(\zeta (\varvec{\alpha })\) and a multiple zeta-star value \(\zeta ^\star (\varvec{\alpha })\) are defined to be [9, 10]
and
We denote the parameters \(w(\varvec{\alpha })=\mid \varvec{\alpha }\mid = \alpha _{1} + \alpha _{2} + \cdots + \alpha _{r}\), \(d(\varvec{\alpha })=r\), and \(h(\varvec{\alpha })=\#\{i\mid \alpha _i>1, 1\le i\le r\}\), called, respectively, the weight, the depth, and the height of \(\varvec{\alpha }\) (or of \(\zeta (\varvec{\alpha })\), or of \(\zeta ^\star (\varvec{\alpha })\)). In fact, these sums were originally studied by Euler. Beginning in the early 1990s, the works of Hoffman [14] and Zagier [25] began to stimulate research again.
For any multiple zeta value \(\zeta (\varvec{\alpha })\), we put a bar on top of \(\alpha _j\) (\(j=1,2,\ldots ,r\)) if there is a sign \((-1)^{k_j}\) appearing in the numerator of the summand [22, 23]. We call it an alternating multiple zeta value. For example, an alternating multiple zeta value
where \(\{a\}^{k}\) is k repetitions of a. Its weight is \(m+3\), the height is one, and the depth is \(m+2\).
Due to Kontsevich, multiple zeta values can be represented by iterated integrals over a simplex of dimension weight:
with \(E_{\mid \varvec{\alpha }\mid }: 0< t_{1}< t_{2}< \cdots< t_{\mid \varvec{\alpha }\mid } < 1\) and
Once multiple zeta values are expressed as iterated integrals, the shuffle product of two multiple zeta values can be defined as:
where \(\phi \) ranges over all permutations on the set \(\{ 1,2,\ldots ,m+n \}\) that preserve the orderings of \(\Omega _{1} \Omega _{2} \cdots \Omega _{m}\) and \(\Omega _{m+1} \Omega _{m+2} \cdots \Omega _{m+n}\). This is equivalent to saying that for all \(1 \le i < j \le m\) or \(m+1 \le i < j \le m+n\), we have \(\phi ^{-1}(i) < \phi ^{-1}(j)\). It is clear that the shuffle product of two multiple zeta values of weight m and n, respectively, will produce \(\left( {\begin{array}{c}m+n\\ n\end{array}}\right) \) multiple zeta values of weight \(m+n\). Transforming the Kontsevich iterated integrals into log-type iterated integrals can make great progress in the research of multiple zeta values, alternating multiple zeta values, and so on, refer to [4, 6, 11, 12, 23].
In this paper, we will investigate the following integral:
We derive our main theorem by finding its different representations.
Main Theorem For integer \(p \ge 0\), we have
where \(W_{\varvec{\alpha }}(a,b,c) = 2^{\sigma (a+b+1) -\sigma (a)-(b+1)} (1-2^{1-\alpha _{a+b+1}})\), with \(\sigma (r) = \sum _{j=0}^{r} \alpha _{j}\), and
For brevity, we hereafter use the lowercase English letters and lowercase Greek letters in the summation, with or without subscripts, to denote the nonnegative and positive integers, unless otherwise specified.
We present our formulas for \(p=0\) and \(p=1\) as follows:
In fact, our main theorem is from the integral
As a replacement of shuffle product, we decompose the domain \(\{(t_1,t_2,u_1,u_2)\in \mathbb R^4\mid 0<t_1<t_2<1,0<u_1<u_2<1\}\) into 6 disjoint simplices of dimension 4:
So
with
The main task now is to evaluate \(\mathbb J(j)\) one by one in terms of multiple zeta values. The weighted sum in our main theorem is \(\sum _{j=3}^6\mathbb J(j)\). The crucial step is to prove
and the resulted shuffle relation is
We organize this paper as follows: We give some preliminaries and auxiliary tools in Sect. 2. In Sect. 3, we break \(J_p\) into six parts and evaluate \(\mathbb J(1)\) and \(\mathbb J(2)\) in terms of multiple zeta values. In the next section, we express \(\mathbb J(j)\) for \(3\le j\le 6\) as the part of the weighted sum of our main theorem. We prove the crucial step Eq. (1.1) in Sect. 5. In Sect. 6, we give another application of the identity Eq. (1.2) to another sum formula which is also proved by a duality theorem due to Ohno [19]. In Sect. 7, we relate \(\mathbb J(3)\) with the weighted sum
In the final section, we will give some formulas related to weighted alternating Euler sums, e.g.,
2 Some Preliminaries and Auxiliary Tools
Here, we list two general integral representations that we will use frequently in the similar tricks.
Proposition 2.1
[11] For nonnegative integers \({r}, b_{1}, b_{2}, \ldots , b_{r}, b_{r+1}\), we have
where \(E_{r+2}\) is a simplex defined as: \(0< t_{1}< t_{2}< \cdots< t_{r+2}{<1}\).
Proposition 2.2
[11, Proposition 2.1] For integers \(p,q,r,\ell \ge 0\),
The well-known sum formula due to Granville [13] asserted that
If we let \(p = \ell = 0\) in the formula of Proposition 2.2, then we have
Therefore, the integral
Our main results are obtained by the convolution of the following sums:
A special case of results proved by Le and Murakami [18] is stated as follows:
Both \(Z_{-}(n)\), and \(Z^\star _{+}(n)\) are sums of multiple zeta(-star) values of height one and can be expressed as double integrals (see [6])
Through the duality theorem \(\zeta (\{1\}^{a},b+2) = \zeta (\{1\}^{b},a+2)\), we see that \(Z_-(2m+1) = 0\). Therefore, we combine these with Eq. (2.3), and then, we have
Proposition 2.3
[6, Proposition 3.1] For any nonnegative integer m, we have \(Z_-(2m+1)=0\) and
The sum of multiple zeta-star values \(Z^\star _+(n)\) appeared as the principal term of the evaluation of \(\zeta ^{\star }(3,\{2\}^{n})\) (see [5]):
Arakawa and Kaneko [1] defined the function
where \({\text {Li}}_k(s)\) denotes the k-th polylogarithm \({\text {Li}}_k(s)=\sum ^\infty _{n=1}\frac{s^n}{n^k}\). It is exactly the multiple zeta-star values of height one
Many properties of the generalized Arakawa–Kaneko zeta functions have been discovered recently (ref. [3, 8, 15,16,17, 24]).
Indeed, \(Z^\star _+(n)\) has the generating function
The dual of the above is
Using the binomial theorem to expand \((1-u_1)^{-x-1}\), we can evaluate the integral as:
Fortunately, we have the identity
by investigating possible poles of both sides of the meromorphic functions. This leads to the evaluation of \(Z^\star _+(n)\):
where \(w>1\) is an integer. This formula was first proved by Ohno [20, Theorem 8] in 2005.
3 The Main Integral \(J_p\) and its \(\mathbb J(1), \mathbb J(2)\) Parts
We begin with the integral
where \(E_{2} \times E_{2} = \{ (t_{1}, t_{2}, u_{1}, u_{2}) \in \mathbb {R}^{4} \mid 0< t_{1}< t_{2}< 1, 0< u_{1}< u_{2} < 1 \}\). As
so we have the evaluation
where \(Z_-(m)\) and \(Z_+^\star (n)\) are defined in Eq (2.2). By Proposition 2.3\(Z_-(2m+1)=0\), thus
We got an expression of the finite convolution of \(Z_-(m)\) and \(Z_+^\star (n)\) in [6, Corollary 5.4]:
In this paper, we consider another expression of \(J_p\). When we decompose \(E_{2} \times E_{2}\) into 6 simplices of dimension 4:
and let
Then, \(J_p\) is broken into 6 parts:
In the following, we will evaluate \(\mathbb J(j)\), for \(1\le j\le 6\), in terms of multiple zeta values.
Proposition 3.1
Given any nonnegative integer p, we have
Proof
We first expand the integrand of \(\mathbb J(1)\) as:
Since \(\mathbb J(1)\) is an integral on \(D_1: 0<t_1<t_2<u_1<u_2<1\), we replace the factor
by its equal
Then, we expand the integrand of \(\mathbb J(1)\) as:
The factor
forms a sum and the other factor
forms another sum, so
Therefore, we complete the proof. \(\square \)
Proposition 3.2
For any nonnegative integer p, we have
Proof
On \(D_{2}: 0< u_{1}< u_{2}< t_{1}< t_{2} < 1\), the factor
is replaced by
and the integrand
is expanded into
So that
\(\square \)
It is noting that we have another expression of \(\mathbb J(2)\) in [7, Theorem 3]:
4 The Integrals \(\mathbb J(3), \mathbb J(4), \mathbb J(5)\), and \(\mathbb J(6)\) Parts
In this section, we will give the evaluations of \(\mathbb J(j)\) for \(3\le j\le 6\). These expressions give us the part of the weighted sum of our main theorem. Our method of proving the following proposition is inspired by the approach used in [21, Theorem 5.3].
Proposition 4.1
Let p be a nonnegative integer. We have
with \(\sigma (r) = \alpha _{0} + \alpha _{1} + \cdots + \alpha _{r}\) and
Proof
The integral \(\mathbb J(3)\) is on \(D_3: 0<t_1<u_1<t_2<u_2<1\). Therefore, we replace the integrand of \(\mathbb J(3)\) by
Then, we expand it as:
This yields
We change the variables \(\varvec{\alpha ,\beta ,\gamma }\) to a nonnegative vector variable \(\varvec{e}=(e_0,e_1,\ldots ,e_{m+1})\). Then,
Since
we can change the nonnegative vector variable \(\varvec{e}\) to a new positive variable \(\varvec{\alpha }\) and \(\mathbb J(3)\) becomes
We rewrite \(\mathbb J(3)\) as:
with \(\sigma (r) = \alpha _{0} + \alpha _{1} + \cdots + \alpha _{r}\) and
\(\square \)
In the same manner as the previous proposition, we obtain the following:
Proposition 4.2
Notation as introduced above, we have
where \(W_{\varvec{\alpha }}(a,b,c)\) is defined in Eq. (4.1) and
Proposition 4.3
Notation as introduced above, then we have
where
Proposition 4.4
Notation as introduced above, then we have
with \(\sigma (r) = \alpha _{0} + \alpha _{1} + \cdots + \alpha _{r}\) and
So up to now, our shuffle relation appeared to be the form:
We write it in a compact form:
5 Another Way to Evaluate \(\mathbb J(1)\) and \(\mathbb J(2)\) Parts
The modified Bell polynomials are defined by [2, 8]
So that \(P_{m}(x_{1},x_{2},\ldots ,x_{m})\) has the expression
In particular, for \(m = 0,1,2,3\),
It is well known that
In order to give another expression of \(\mathbb J(1)\), we need the following lemma:
Lemma 5.1
[4, Proposition 4.4] For a pair of positive integers \(k_{1}\), \(k_{2}\) with \(k_{1} \le k_{2}\), let
Then, for any nonnegative integer n,
with \(h_{n} = \sum _{j=k_{1}}^{k_{2}} \frac{1}{j^{n}}\).
Proposition 5.2
Let p be a nonnegative integer and \(\mathbb J(1)\) be equal to
Then,
Proof
Let S(m, n) be the general term inside the summation of \(\mathbb J(1)\). Then,
is the generating function for the double sequence \(\{S(m,n)\}\) and
Beginning with
and then integrating with respect to \(t_{1}\), \(t_{2}\) and \(u_{1}\), we obtain that
The value of integral is
So
after differentiations with respect to x, y for m, n times, and we use Lemma 5.1,
The general term of S(m, n) is
and hence
We use a result in [7, Theorem 2]:
Therefore we have
\(\square \)
To transform
into multiple zeta values related to \(\mathbb J(1)\), we need the following reflection formula.
Proposition 5.3
[7, Proposition 4] For an r-tuple \(\varvec{\alpha }=(\alpha _1,\alpha _2,\ldots ,\alpha _r)\) of positive integers with \(\alpha _1\ge 2\), \(\alpha _r\ge 2\), we have
Proposition 5.4
Let p be a nonnegative integer and
Then, we have
Proof
By the reflection formula (see Proposition 5.3), we know that \(\zeta (c+2,\{1\}^{m},d+2)\) is equal to
so that
The first sum can be rewritten as:
which is precisely equal to \(J_{p}\) expressed as:
\(\square \)
According to previous propositions concerning \(\mathbb J(1)\) and \(\mathbb J(2)\), we have the following
Corollary 5.5
Notation as above, then we have
Now we combine Eq. (4.2) and we conclude our main theorem.
6 An Application of \(\mathbb J(3)+\mathbb J(4)+\mathbb J(5)+\mathbb J(6)\)
Let us begin with another integral
where p is a nonnegative integer. Since
the integral \(\mathbb I(p)\) can be written as:
On the other hand, we decomposed the integral \(\mathbb I(p)\) into six parts in a similar way to \(\mathbb J_p\):
where
The first integral \(\mathbb I(p;1)\) is on \(D_1: 0<t_1<t_2<u_1<u_2<1\), and we have
The integral \(\mathbb I(p;2)\) is
The next four parts have corresponding equality with the parts \(\mathbb J(i)\), for \(3\le i\le 6\), as
Therefore, we get
This gives us the following theorem.
Theorem 6.1
Let p be an nonnegative integer. Then,
Here, we also give another proof using a duality theorem due to Ohno [19]. Given any nonnegative integer a, we know that the dual of \(\zeta (\{1\}^a,2,2)\) is \(\zeta (2,a+2)\) and the dual of \(\zeta (2,\{1\}^a,2)\) is \(\zeta (a+2,2)\), respectively. Then, for a nonnegative integer m, we have
We substitute the above formulas in \(\mathbb I(p;1)+\mathbb I(p;2)\), and then, the desired result will be obtained.
7 Another Expression of \(\mathbb J(3)\)
The sum of multiple zeta values
came from a joint of three sums
and it has the integral representation
which is equal to
Let T(m, n) be the general term in the above sum. Here, we are going to find the value of T(m, n) through its generating function.
Proposition 7.1
For integers \(m,n \ge 0\), let
Then,
Proof
The generating function for the double sequence T(m, n) is
Like the case of \(\mathbb J(1)\), G(x, y) can be evaluated as:
Hence,
It is equal to
as asserted. \(\square \)
Corollary 7.2
For any nonnegative integer p, we have
Remark 7.3
We have [7, Theorem 1] a duality theorem:
However, it is still unknown for the weighted sum
However, our weighted sum formula did provide an approximate evaluation of weighted sum of zeta-star values of such kind.
In the next section, we will try give some weighted alternating Euler sums.
8 Weighted Alternating Euler Sum Formulas
The weighted alternating Euler sum formula
was obtained in [6, Eq. (5.3)]. Now we produce a more general formula which it covers the above.
Theorem 8.1
For any nonnegative integers p, q and a real number \(\lambda \), we have
Proof
It is known that [6, Eq. (2.3)]:
We evaluate the following sum:
as an integral form
This integral can be decomposed into two parts:
It can be seen that
Therefore, we have
We differential q times with \(\lambda \) in the above equation, and we obtain the final identity. \(\square \)
We list some examples. Let \(\lambda =0\) in Eq. (8.1). The following identity is obtained.
Then,
It is note that the first identity in the above formulas can be found in [22, Eq. (2.15)].
Let \(\lambda =-1\) in Eq. (8.1). Then,
The \(q=0\) and \(q=1\) cases are
Let \(\lambda =1\) in Eq. (8.1), we have
We set \(q=0\) in the above identity; we will get [6, Eq. (2.14)]
The well-known double-shuffle relation [22, Eq. (2.2)]
where \(r\ge 1\), \(s\ge 1\). We use the above relation and the sum formula [22, Eq. (2.14)]
we derive the following
This identity was obtained in [6, Eq. (5.3)]. We indicate this formula in the first paragraph of this section.
Let \(\lambda =-2\) in Eq. (8.1), we have
We set \(q=0\) in the above identity; we have
Using the double-shuffle relation Eq (8.3), we have
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
References
Arakawa, T., Kaneko, M.: Multiple zeta values, poly-Bernoulli numbers, and related zeta functions. Nagoya Math. J. 153, 189–209 (1999). https://doi.org/10.1017/S0027763000006954
Chen, K.-W.: Generalized harmonic numbers and Euler sums. Int. J. Number Theor. 13(2), 513–528 (2017). https://doi.org/10.1142/S1793042116500883
Chen, K.-W.: Generalized Arakawa-Kaneko zeta functions. Integr. Transforms Spec. Funct. 30(4), 282–300 (2019). https://doi.org/10.1080/10652469.2018.1562450
Chen, K.-W., Chung, C.-L., Eie, M.: Sum formulas and duality theorems of multiple zeta values. J. Number Theor. 158, 33–53 (2016). https://doi.org/10.1016/j.jnt.2015.06.014
Chen, K.-W., Eie, M.: Some special Euler sums and \(\zeta ^{\star }(r+2,\{2\}^{n})\), (2018), http://arxiv.org/abs/1810.11795v1 [math.NT]
Chen, K.-W., Eie, M.: On the convolutions of sums of multiple zeta(-star) values of height one, Ramanujan J. 59(4), 1197–1223 (2022). https://doi.org/10.1007/s11139-022-00628-7
Chen, K.-W., Eie, M.: On three general forms of multiple zeta(-star) values, (2022), http://arxiv.org/abs/2202.03839v1 [math.NT]
Coppo, M.-A., Candelpergher, B.: The Arakawa-Kaneko zeta function. Ramanujan J. 22(2), 153–162 (2010). https://doi.org/10.1007/s11139-009-9205-x
Eie, M.: Topics in number theory, monographs in number theory 2. World Scientific, Singapore (2009)
Eie, M.: The theory of multiple zeta values with applications in combinatorics, monographs in number theory 7. World Scientific, Singapore (2013)
Eie, M., Wei, C.-S.: A short proof for the sum formula and its generalization. Arch. Math. 91, 330–338 (2008). https://doi.org/10.1007/s00013-008-2547-6
Eie, M., Yang, F.-Y.: Weighted sum formulas from shuffle products of multiples of Riemann zeta values. J. Number Theor. 147, 749–765 (2015). https://doi.org/10.1016/j.jnt.2014.08.017
Granville, A.: A decomposition of Riemann’s zeta-function, in: Analytic Number Theory (Kyoto, 1996), 95–102, London Math. Soc. Lecture Notes Ser. 247, Cambridge Univ. Press, Cambridge, 1997. https://doi.org/10.1017/CBO9780511666179.009
Hoffman, M.E.: Multiple harmonic series. Pac. J. Math. 152, 275–290 (1992). https://doi.org/10.2140/pjm.1992.152.275
Kaneko, M., Tsumura, H.: Multi-poly-Bernoulli numbers and related zeta functions. Nagoya Math. J. 232, 19–54 (2018). https://doi.org/10.1017/nmj.2017.16
Kargın, L.: Poly-p-Bernoulli polynomials and generalized Arakawa-Kaneko zeta function. Lith. Math. J. 60, 29–50 (2020). https://doi.org/10.1007/s10986-019-09448-7
Kawasaki, N., Ohno, Y.: Combinatorial proofs of identities for special values of Arakawa-Kaneko multiple zeta functions. Kyushu J. Math. 72, 215–222 (2018). https://doi.org/10.2206/kyushujm.72.215
Le, T.Q.T., Murakami, J.: Kontsevich’s integrals for the Homfly polynomial and relations between values of multiple zeta functions. Topol. Appl. 62(2), 193–206 (1995). https://doi.org/10.1016/0166-8641(94)00054-7
Ohno, Y.: A generalization of the duality and sum formulas on the multiple zeta values. J. Number Theor. 74(1), 39–43 (1999). https://doi.org/10.1006/jnth.1998.2314
Ohno, Y.: Sum relations for multiple zeta values. In: Aoki, T., Kanemitsu, S., Nakahara, M., Ohno, Y. (eds.) Zeta functions, topology and quantum physics, developments in mathematics. Springer, Boston (2005)
Ong, Y.L., Eie, M., Liaw, W.-C.: On generalizations of weighted sum formulas of multiple zeta values. Int. J. Number Theor. 9(5), 1185–1198 (2013). https://doi.org/10.1142/S179304211350019X
Teo, L.-P.: Alternating double Euler sums, hypergeometric identities and a theorem of Zagier. J. Math. Anal. Appl. 462, 777–800 (2018). https://doi.org/10.1016/j.jmaaa.2018.02.037
Xu, C.: Integrals of logarithmic functions and alternating multiple zeta values. Math. Slovaca 69(2), 339–356 (2019). https://doi.org/10.1515/ms-2017-0227
Xu, C.: Duality formulas for Arakawa-Kaneko zeta values and related variants. Bull. Malaysia Math. Sci. Soc. 44, 3001–3018 (2021). https://doi.org/10.1007/s40840-021-01099-7
Zagier, D.: Values of zeta functions and their applications, in: First European Congress of Mathematics, vol. II, Birkhauser, Boston, 497–512 (1994). https://doi.org/10.1007/978-3-0348-9112-7_23
Funding
The first author (corresponding author) was funded by the Ministry of Science and Technology, Taiwan, R.O.C., under Grant MOST 110-2115-M-845-001.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
No potential conflict of interest was reported by the authors.
Additional information
Communicated by Rosihan M. Ali.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
M. Eie: These authors contributed equally to this work.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Chen, KW., Eie, M. Weighted Sum Formulas from Shuffle Products of Multiple Zeta-Star Values. Bull. Malays. Math. Sci. Soc. 46, 50 (2023). https://doi.org/10.1007/s40840-022-01446-2
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40840-022-01446-2