Next Article in Journal
Estimates for Certain Rough Multiple Singular Integrals on Triebel–Lizorkin Space
Next Article in Special Issue
On a Class of Generalized Multivariate Hermite–Humbert Polynomials via Generalized Fibonacci Polynomials
Previous Article in Journal
Existence of Three Positive Solutions for Boundary Value Problem of Fourth Order with Sign-Changing Green’s Function
Previous Article in Special Issue
On Fractional Operators Involving the Incomplete Mittag-Leffler Matrix Function and Its Applications
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Linear Arrangement of Euler Sums with Multiple Argument

School of Science, RMIT University, Melbourne, VIC 3000, Australia
Symmetry 2024, 16(10), 1322; https://doi.org/10.3390/sym16101322
Submission received: 2 September 2024 / Revised: 2 October 2024 / Accepted: 5 October 2024 / Published: 8 October 2024
(This article belongs to the Special Issue Research in Special Functions)

Abstract

:
We investigate the linear arrangement of Euler harmonic sums that may be expressed in closed form in terms of special functions such as the classical Riemann zeta function and the Dirichlet eta function. Particular emphasis is given to Euler harmonic sums with even weight. New examples highlighting the theorems will be presented.

1. Introduction and Background

Euler initiated the study of linear harmonic number sums of the form, using the notation of Flajolet and Salvy [1]
S p , t + + ( q ) : = n = 1 H q n ( p ) n t , S p , t + ( q ) : = n = 1 ( 1 ) n + 1 H q n ( p ) n t ,
for the specific case q = 1 and with the integer p + t designated as the weight. The harmonic numbers H n p of integer order p are defined as
H n p : = r = 1 n 1 r p , p C and n N ,
where C and N are the sets of complex numbers and positive integers, respectively. The harmonic numbers of order one, H q n , are given by, for q R + the set of positive real numbers,
H q n = γ + ψ q n + 1 = γ + ψ q n + 1 q n n N and H 0 : = 0 .
The term γ represents the familiar Euler–Mascheroni constant (see, e.g., ([2], Section 1.2)), and ψ t denotes the digamma (or psi) function defined by
ψ t : = d d t log Γ t = Γ t Γ t t C · , · , 3 , 2 , 1 , 0 , and ψ t + 1 = ψ t + 1 t ,
where Γ t is the familiar Gamma function (see, e.g., ([2], Section 1.1)). A generalization of the digamma function is the polygamma function ψ k ( z ) defined by
ψ k ( z ) : = d k d z k { ψ ( z ) } = 1 k + 1 k ! r = 0 1 r + z k + 1 = 1 k + 1 k ! ζ ( k + 1 , z ) , k N ; z C · , · , 3 , 2 , 1 , 0 .
In his works on series, Euler solved the famous Basel problem,
n = 1 1 n 2 = ζ 2 = π 2 6 .
Furthermore, Euler proved from the properties of the Bernoulli numbers, B 0 = 1 , B 1 = 1 2 , B 2 = 1 6 , B 2 r + 1 = 0 , for r 1 , given by the generating function
w exp w 1 = j 0 w j j ! B j , for w < 2 π
and the classical Euler–Bernoulli relation (see [2], p. 166), that
ζ 2 r = 1 r + 1 B 2 r 2 2 r 1 π 2 r 2 r ! .
The Riemann zeta function, ζ t , is defined as
ζ t : = j 1 1 j t = 1 1 2 t j 1 1 2 j 1 t ; ( t ) > 1 1 1 2 1 t j 1 1 j 1 j t , ( t ) > 0 , t 1 ,
The Dirichlet eta function, η ( t ) is defined by
η ( t ) : = n 1 ( 1 ) n 1 n t = 1 2 1 t ζ ( t ) ( ( t ) > 0 ) ,
where ( t ) defines the real part of t. The objective of this paper is to obtain linear arrangements of (1) in closed form in terms of special functions, such as
1 t t q t + 1 S 1 , t + 2 + + ( q ) S 1 , t + 2 + + ( q 2 ) + 1 t q t 1 t S 1 , t + 2 + ( 1 q ) + 1 t q t S 2 , t + 1 + + ( q ) 1 2 S 2 , t + 1 + + ( q 2 ) + 1 t q 1 t S 2 , t + 1 + ( 1 q )
= 1 q ζ 2 η t + 1 + 1 t q t η 2 ζ t + 1 + 1 t j = 2 t 1 j q t + 2 j t + 1 j η j ζ ( t + 3 j ) ,
and, in a specific case, a linear arrangement of harmonic sums with even weight
2 S 2 , 12 + 4095 4096 S 2 , 12 + + S 12 , 2 + + 12 S 1 , 13 + = 75 , 989 , 887 1 , 720 , 320 ζ ( 14 ) 189 32 ζ ( 7 ) 2 375 32 ζ ( 5 ) ζ ( 9 ) 5883 512 ζ ( 11 ) ζ ( 3 ) .
To the author’s knowledge, there are few, if any, expressions of the type (4) in the published literature. The idea of expressing infinite series, of Euler type, in closed form is often difficult but often important. Any expression represented in closed form has obvious and desirable properties, as may be ascertained from a published work of Borwein and Crandall [3]. It is generally accepted that S p , q + + may not have a closed-form representation other than the pair 2 , 4 , 4 , 2 and p = q . If S p , q + + can be evaluated, then S q , p + + exists by virtue of the shuffle relation
S p , q + + + S q , p + + = ζ ( p ) ζ ( q ) + ζ ( p + q ) .
The symmetrical nature of the shuffle relation and the identities established in this paper lend themselves to further exploration. In a forthcoming paper, we shall exploit the identity (4) in another symmetrical arrangement of harmonic numbers of a higher order. For the case q = 1 , we recall the following known results
S 1 , t + + = 1 + t 2 ζ t + 1 1 2 j = 1 t 2 ζ j + 1 ζ t j ,
due to Euler [4]. For odd weight p + t , Borwein et al. [5] gave the identity:
S p , t + + = 1 1 p 2 ζ p ζ t + 1 1 p p + t t p + t 1 p ζ p + t 2
+ 1 p j = 1 p 2 p + t 2 j 1 t 1 ζ 2 j ζ p + t 2 j
+ 1 p j = 1 t 2 p + t 2 j 1 p 1 ζ 2 j ζ p + t 2 j .
For the alternating case, for odd weight 1 + t , Sitaramachandrarao [6] published
2 S 1 , t + = t + 1 η t + 1 ζ t + 1 2 j = 1 t 2 1 ζ 2 j ζ t + 1 2 j .
Flajolet and Salvy [1] gave the expression for odd weight p + t
2 S p , t + = 1 1 p ζ p η t + η p + t
+ 2 j + 2 k = p t + j 1 t 1 1 j + 1 η t + j η 2 k + 2 1 p i + 2 k = t p + i 1 p 1 ζ p + i η 2 k ,
where η · is the Dirichlet eta function. Recently, Alzer and Choi [7] obtained a good result for p N 1
2 S p , 1 + = p ζ p + 1 ζ t + 1 2 j = 1 p η j η p + 1 j .
The polylogarithm function Li p ( z ) of order p , and for each integer p 1 , is defined by (see, e.g., [2], p. 198)
Li p ( z ) = m = 1 z m m p | z | 1 ; p N 2
and
Li 1 ( z ) = log 1 z , | z | 1 .
There are many interesting and significant results associated with Euler harmonic sum identities, some of which may be seen in the works of [5,8,9,10,11,12,13,14]. Li and Chu [15] have published some important results in relation to generating functions for binomial series involving harmonic-like numbers, and Chen [16] has investigated some general Tornheim-type series. The majority of the published works that deal with Euler harmonic sums of the type (11) deal with the case q = 1 .
This paper will be organized in the following way. The main results will be established, proving that
1 t q t 1 + t S 1 , t + 2 + + ( q ) + S 2 , t + 1 + + ( 1 q ) + 1 t q t 1 S 2 , t + 1 + + ( q )
has a closed form. A special case in the form of a Corollary is detailed. The second theorem deals with expressing the linear arrangement
1 t t q t + 1 S 1 , t + 2 + + ( q ) S 1 , t + 2 + + ( q 2 ) + S 1 , t + 2 + ( 1 q ) + 1 t q t S 2 , t + 1 + + ( q ) 1 2 S 2 , t + 1 + + ( q 2 ) + 1 q S 2 , t + 1 + ( 1 q )
in closed form with its resultant corollary. Some examples will be given, highlighting the linear arrangement of Euler sums with even weight in closed form.

2. The Main Results

Consider the following theorem.
Theorem 1.
Let t N , t 2 and let q R + 0 ; the following identity is valid:
1 t q t 1 + t S 1 , t + 2 + + ( q ) + S 2 , t + 1 + + ( 1 q ) + 1 t q t 1 S 2 , t + 1 + + ( q )
= 1 q ζ ( 3 ) ζ ( t ) + 1 t t q 1 t ζ ( 2 ) ζ ( t + 1 )
+ j = 3 t 1 j 1 t q t + 1 j t + 1 j ζ ( j ) ζ t + 3 j + 1 q j 1 ζ ( j + 1 ) ζ t + 2 j ,
where S 1 , t + 2 + + ( q ) , S 2 , t + 1 + + ( q ) are the linear Euler harmonic sum (1) of weight t + 3 , and ζ t is the classical Riemann zeta function.
Proof. 
Consider the standard integral representation for the harmonic number
q n H n q : = 0 1 x n q 1 ln 1 x d x = 0 1 ln 1 x x x n q d x ,
Then,
q q n H n q = 0 1 ln 1 x x q x n q d x
q 2 n 2 H n q + q n H n q 2 q ζ 2 = 0 1 ln 1 x ln x x x n q d x .
Summing over the integers n for t N , t 2 ,
n 1 1 n t q 2 n 2 H n q + q n H n q 2 q n ζ 2 = 0 1 ln x ln 1 x x n 1 x n q n t d x ,
and applying definition (10), we obtain
n 1 1 n t q 2 n 2 H n q + q n H n q 2 q n ζ 2 = 0 1 ln x ln 1 x x Li t ( x 1 q ) d x .
Using the transformation y q = x , we obtain
S 1 , t + 2 + + ( 1 q ) + 1 q S 2 , t + 1 + + ( 1 q ) 1 q ζ 2 ζ t + 1 = 0 1 ln x ln 1 y q y Li t ( y ) d y ,
A Taylor series expansion of the ln 1 y q term gives us the representation
S 1 , t + 2 + + ( 1 q ) + 1 q S 2 , t + 1 + + ( 1 q ) 1 q ζ 2 ζ t + 1 = n 1 1 n 0 1 y q n 1 ln x Li t ( y ) d y
= n 1 t 1 t + 1 H n q n n q t + 1 + n 1 1 t + 1 ζ ( 2 ) H n q 2 n n q t + 1 t j = 2 t 1 j t + 1 j ζ ( j ) n 1 1 n n q t + 2 j
= t 1 t + 1 q t + 1 S 1 , t + 2 + + ( q ) + 1 t q t ζ 2 ζ t + 1 + 1 t q t S 2 , t + 1 + + ( q ) + 1 t j = 2 t 1 j q t + 2 j t + 1 j ζ ( j ) ζ ( t + 3 j )
= t 1 t + 1 q t + 1 S 1 , t + 2 + + ( q ) + 1 t q t S 2 , t + 1 + + ( q ) + 1 t j = 3 t 1 j q t + 2 j t + 1 j ζ ( j ) ζ ( t + 3 j ) .
By rearranging, we obtain
q S 1 , t + 2 + + ( 1 q ) + S 2 , t + 1 + + ( 1 q ) + t 1 t q t S 1 , t + 2 + + ( q ) + 1 t q t 1 S 2 , t + 1 + + ( q ) = 1 + 1 t t q t 1 ζ 2 ζ t + 1 + 1 t j = 3 t 1 j q t + 1 j t + 1 j ζ ( j ) ζ ( t + 3 j ) ,
From the following relation (see [11]), we substitute the identity
S 1 , t + 1 + + ( 1 q ) = 1 t + 1 q t S 1 , t + 1 + + ( q ) + j = 1 t 1 1 j + 1 q j ζ ( j + 1 ) ζ t + 1 j
into the previous equation; hence,
1 t q t 1 + t S 1 , t + 2 + + ( q ) + S 2 , t + 1 + + ( 1 q ) + 1 t q t 1 S 2 , t + 1 + + ( q ) = 1 + 1 t t q 1 t ζ ( 2 ) ζ ( t + 1 )
+ 1 t j = 3 t 1 j q t + 1 j t + 1 j ζ ( j ) ζ t + 3 j + j = 1 t 1 j q j 1 ζ ( j + 1 ) ζ t + 2 j .
Here, we isolate the j = 1 , 2 terms of the second sum and the proof of the theorem is complete. □
As a direct consequence of Theorem 1, we can highlight some special cases, as in the following corollary.
Corollary 1.
From Theorem 1, we can highlight some special cases. Consider the case of  q = 1 ; then,
1 t 1 + t S 1 , t + 2 + + + S 2 , t + 1 + + + 1 t S 2 , t + 1 + + = 1 t t ζ ( 2 ) ζ ( t + 1 ) + ζ ( 3 ) ζ ( t )
+ j = 3 t 1 j 1 t t + 1 j ζ ( j ) ζ t + 3 j + ζ ( j + 1 ) ζ t + 2 j .
If t is even, let  t = 2 t *  (and rename  t * = t ); hence,
2 S 2 , 2 t + 1 + + = 2 t ζ ( 2 ) ζ ( 2 t + 1 ) + ζ ( 3 ) ζ ( 2 t ) 1 + 2 t S 1 , 2 t + 2 + + + j = 3 2 t 1 j 2 t + 1 j ζ ( j ) ζ 2 t + 3 j + ζ ( j + 1 ) ζ 2 t + 2 j .
If t is odd, let t = 2 t * 1 (and rename t * = t ); hence,
2 t S 1 , 2 t + 1 + + = 2 t 1 ζ ( 2 ) ζ ( 2 t ) + ζ ( 3 ) ζ ( 2 t 1 ) + j = 3 2 t 1 j ζ ( j + 1 ) ζ 2 t + 1 j 2 t j ζ ( j ) ζ 2 t + 2 j .
From (11), for the case q = 2 , let t = 2 t * 1 (and rename t * = t ), where t N , we obtain
2 t 2 2 t 1 S 1 , 2 t + 1 + + ( 2 ) + S 2 , 2 t + + ( 1 2 ) 1 2 2 t 2 S 2 , 2 t + + ( 2 )
= 1 2 ζ ( 3 ) ζ ( 2 t 1 ) 2 t 1 2 2 t 2 ζ ( 2 ) ζ ( 2 t )
+ j = 3 2 t 1 1 j 1 2 j 1 ζ ( j + 1 ) ζ 2 t + 1 j 2 t j 2 2 t j ζ ( j ) ζ 2 t + 2 j .
By applying the decomposition relation
2 t S 1 , t + 1 + + ( q ) = S 1 , t + 1 + + q 2 S 1 , t + 1 + q 2 ,
we can rewrite (13) as
4 t S 1 , 2 t + 1 + + S 2 , 2 t + + ( 1 2 ) + 2 S 2 , 2 t + 2 S 2 , 2 t + +
= 4 t S 1 , 2 t + 1 + + + 1 2 ζ ( 3 ) ζ ( 2 t 1 ) 2 t 1 2 2 t 2 ζ ( 2 ) ζ ( 2 t )
+ j = 3 2 t 1 1 j 1 2 j 1 ζ ( j + 1 ) ζ 2 t + 1 j 2 t j 2 2 t j ζ ( j ) ζ 2 t + 2 j ,
where S 1 , 2 t + 1 + + is the Euler identity (5). In ref. [17], the authors proved that
S 2 , 2 t + + ( 1 2 ) = 2 1 2 t S 2 , 2 t + + + 2 S 2 , 2 t + 2 S 2 t , 2 + 2 η 2 η 2 t + 2 η 2 t + 2
By simplifying (15), we have that
4 t S 1 , 2 t + 1 + + 2 1 2 t 2 S 2 , 2 t + + + 4 S 2 , 2 t + 2 S 2 t , 2 + = 2 η 2 η 2 t 2 η 2 t + 2
+ 4 t S 1 , 2 t + 1 + + + 1 2 ζ ( 3 ) ζ ( 2 t 1 ) 2 t 1 2 2 t 2 ζ ( 2 ) ζ ( 2 t )
+ j = 3 2 t 1 1 j 1 2 j 1 ζ ( j + 1 ) ζ 2 t + 1 j 2 t j 2 2 t j ζ ( j ) ζ 2 t + 2 j .
The individual Euler sums in (15) and (16) are not known explicitly in closed form for t 3 . Therefore, (15) and (16) provide a linear arrangement of Euler sums with even weight that can be expressed in closed form.
For the case t = 4 , we have
16 S 1 , 9 + + S 2 , 8 + + ( 1 2 ) + 2 S 2 , 8 + 2 S 2 , 8 + + = 341 8 ζ ( 10 ) 15 2 ζ ( 5 ) 2 237 16 ζ ( 7 ) ζ ( 3 ) ,
and
16 S 1 , 9 + 255 128 S 2 , 8 + + + 4 S 2 , 8 + 2 S 8 , 2 + = 108201 2560 ζ ( 10 ) 15 2 ζ ( 5 ) 2 237 16 ζ ( 7 ) ζ ( 3 ) .
For the case where t is even, from (12), let t = 2 t * (and rename t * = t ), where t N ,
1 2 2 t 1 + 2 t S 1 , 2 t + 2 + + ( 2 ) + S 2 , 2 t + 1 + + ( 1 2 ) + 1 2 2 t 1 S 2 , 2 t + 1 + + ( 2 )
= 1 2 ζ ( 3 ) ζ ( 2 t ) + t 2 2 2 t ζ ( 2 ) ζ ( 2 t + 1 )
+ j = 3 2 t 1 j 1 2 2 t + 1 j 2 t + 1 j ζ ( j ) ζ 2 t + 3 j + 1 2 j 1 ζ ( j + 1 ) ζ 2 t + 2 j .
From ref. [17] (p. 414), we have that
S 2 , 2 t 1 + + ( 1 2 ) = 2 2 t S 2 , 2 t + 1 + + 22 S 2 , 2 t + 2 S 2 t + 1 , 2 + 2 η 2 η 2 t + 1 + 2 η 2 t + 3
and, therefore, an explicit representation of the linear arrangement of
1 2 2 t 1 + 2 t S 1 , 2 t + 2 + + ( 2 ) + 1 2 2 t 1 S 2 , 2 t + 1 + + ( 2 ) = 1 2 ζ ( 3 ) ζ ( 2 t ) + t 2 2 2 t ζ ( 2 ) ζ ( 2 t + 1 )
S 2 , 2 t + 1 + + ( 1 2 ) + j = 3 2 t 1 j 1 2 2 t + 1 j 2 t + 1 j ζ ( j ) ζ 2 t + 3 j + 1 2 j 1 ζ ( j + 1 ) ζ 2 t + 2 j .
In the next theorem, we establish a linear arrangement of Euler sums involving the alternating sum S 2 , t + 1 + ( 1 q ) , which was not evident in Theorem 1.
Theorem 2.
Let  t N , t 2  and let   q R + 0  the following identity is valid:
1 t t q t + 1 S 1 , t + 2 + + ( q ) S 1 , t + 2 + + ( q 2 ) + S 1 , t + 2 + ( 1 q ) + 1 t q t S 2 , t + 1 + + ( q ) 1 2 S 2 , t + 1 + + ( q 2 ) + 1 q S 2 , t + 1 + ( 1 q )
= 1 q ζ 2 η t + 1 + 1 t q t η 2 ζ t + 1 + 1 t j = 2 t 1 j q t + 2 j t + 1 j η j ζ ( t + 3 j ) ,
where  S 2 , t + 1 + + ( q 2 )  is the linear Euler harmonic sum (1) of weight  t + 3 , and  η t  is the classical Riemann eta function.
Proof. 
As in Theorem 1, the standard integral representation for the harmonic number
q n H n q : = 0 1 x n q 1 ln 1 x d x = 0 1 ln 1 x x x n q d x ,
Then,
q q n H n q = 0 1 ln 1 x x q x n q d x
q 2 n 2 H n q + q n H n q 2 q ζ 2 = 0 1 ln 1 x ln x x x n q d x .
By summing over the alternating integers n for t N 2 ,
n 1 1 n n t q 2 n 2 H n q + q n H n q 2 q n ζ 2 = 0 1 ln x ln 1 x x n 1 1 n x n q n t d x = 0 1 ln x ln 1 x x Li t ( x 1 q ) d x .
By using the transformation y q = x , we obtain
n 1 1 n n t q 2 n 2 H n q + q n H n q 2 q n ζ 2 = 0 1 ln x ln 1 y q y Li t ( y ) d y
A Taylor series expansion of the ln 1 y q term gives us the representation
n 1 1 n n t 1 n 2 H n q + 1 q n H n q 2 1 q n ζ 2 = n 1 1 n 0 1 y q n 1 ln x Li t ( y ) d y ,
Hence, we obtain
1 q ζ 2 η t + 1 1 q S 1 , t + 2 + ( 1 q ) 1 q S 2 , t + 1 + ( 1 q )
= 1 t t 2 q t + 1 n 1 ψ q n 2 + 1 ψ q n 2 + 1 2 n t + 2 + 1 t 4 q t n 1 ψ q n 2 + 1 ψ q n 2 + 1 2 n t + 1
+ 1 t + 1 j = 1 t 1 j t + 1 j ζ t + 3 j η j q t + 2 j .
From the polygamma harmonic number identity (see [2])
1 m m ! ψ m α + 1 = H α m + 1 ζ m + 1
we obtain
1 t t q t + 1 S 1 , t + 2 + + ( q ) S 1 , t + 2 + + ( q 2 ) + S 1 , t + 2 + ( 1 q ) + 1 t q t S 2 , t + 1 + + ( q ) 1 2 S 2 , t + 1 + + ( q 2 ) + 1 q S 2 , t + 1 + ( 1 q )
= 1 q ζ 2 η t + 1 + 1 t t q t + 1 ζ t + 2 ln 2 + 1 t q t η 2 ζ t + 1 + 1 t j = 1 t 1 j q t + 2 j t + 1 j η j ζ ( t + 3 j ) .
In the finite sum, we isolate the j = 1 term, where η 1 = ln 2 , and this completes the proof of Theorem 2. □
As a direct consequence of Theorem 2, we can highlight some special cases, as in the following Corollary.
Corollary 2.
From Theorem 2, we can highlight some special cases. For the case of  q = 1 ,
1 t t S 1 , t + 2 + + S 1 , t + 2 + + ( 1 2 ) + S 1 , t + 2 + + 1 t S 2 , t + 1 + + 1 2 S 2 , t + 1 + + ( 1 2 ) + S 2 , t + 1 +
= ζ 2 η t + 1 + 1 t η 2 ζ t + 1 + 1 t j = 2 t 1 j t + 1 j η j ζ ( t + 3 j ) .
For the case where t is even, which implies an odd weight  t + 3 ,  we can write, from (20), the explicit representation
1 2 S 2 , t + 1 + + ( 1 2 ) = t S 1 , t + 2 + + S 1 , t + 2 + + ( 1 2 ) + S 1 , t + 2 + + S 2 , t + 1 +
ζ 2 η t + 1 η 2 ζ t + 1 j = 2 t 1 j t + 1 j η j ζ ( t + 3 j ) .
All terms on the right of (21) are known from (6) and (8) and where, from [17] (p. 414),
S 1 , t + 1 + + ( 1 2 ) = S 1 , t + 1 + S 1 , t + 1 + + + j = 1 t 1 1 j + 1 2 j ζ ( j + 1 ) ζ t + 1 j .
Specifically, for  t = 4 ,  we have
S 2 , 5 + + ( 1 2 ) = 107 64 ζ ( 7 ) + ζ ( 4 ) ζ ( 3 ) 21 16 ζ ( 5 ) ζ ( 2 ) .
Considering the case where t is odd, let  t = 2 t * 1  (and rename  t * = t ), where  t N , therefore
( 1 2 t ) S 1 , 2 t + 1 + + S 1 , 2 t + 1 + + ( 1 2 ) + S 1 , 2 t + 1 + S 2 , 2 t + + 1 2 S 2 , 2 t + + ( 1 2 ) + S 2 , 2 t +
= ζ 2 η 2 t η 2 ζ 2 t j = 2 2 t 1 1 j 2 t j η j ζ ( 2 t + 2 j ) .
By rearranging, we obtain
S 2 , 2 t + S 2 , 2 t + + + 1 2 S 2 , 2 t + + ( 1 2 ) + ( 2 t 1 ) S 1 , 2 t + 1 + + ( 1 2 )
= ( 2 t 2 ) S 1 , 2 t + 1 + + + ζ 2 η 2 t η 2 ζ 2 t j = 2 2 t 1 1 j 2 t j η j ζ ( 2 t + 2 j ) .
For the specific case  t = 6 ,
S 2 , 12 + S 2 , 12 + + + 1 2 S 2 , 12 + + ( 1 2 ) + 11 S 1 , 13 + + ( 1 2 ) = 115581 8192 ζ ( 14 ) 37 64 ζ ( 7 ) 2 371 256 ζ ( 5 ) ζ ( 9 ) 3329 1024 ζ ( 11 ) ζ ( 3 ) .

3. Illustrative Examples of Euler Harmonic Sums

The next example investigates the case q = 2 from Theorem 2 and the case of q = 4 from Theorem 1.
Example 1.
Consider Theorem 2 and let  q = 2  with  t = 2 t * 1  (and rename  t * = t ), where  t N . This allows the investigation of Euler harmonic sums with even weight. From (18),
1 2 t 2 2 t S 1 , 2 t + 1 + + ( 2 ) S 1 , 2 t + 1 + + + S 1 , 2 t + 1 + ( 1 2 ) 2 1 2 t S 2 , 2 t + + ( 2 ) 1 2 S 2 , 2 t + + + 1 2 S 2 , 2 t + ( 1 2 )
= 1 2 ζ 2 η 2 t 2 1 2 t η 2 ζ 2 t j = 2 2 t 1 1 j 2 2 t + 1 j 2 t j η j ζ ( 2 t + 2 j ) .
By using the decomposition rule (14) and from [17] (p. 414), the half-argument identity, for  p N 0 , t N
1 2 p S p + 1 , t + ( 1 2 ) = S p + 1 , t + S t , , p + 1 + + η p + t + 1 η p + 1 η t
We can simplify (23) so that
2 2 t 1 S 2 , 2 t + + + 2 S 2 , 2 t + S 2 t , 2 + + 2 t S 1 , 2 t + 1 + = 1 2 ζ 2 η 2 t 2 1 2 t η 2 ζ 2 t
+ S 2 t + 1 , 1 + + 2 t 1 1 2 2 t S 1 , 2 t + 1 + + j = 2 2 t 1 1 j 2 2 t + 1 j 2 t j η j ζ ( 2 t + 2 j ) .
The Euler sum  S 2 t + 1 , 1 +  is determined from the Alzer–Choi identity (9) and  S 1 , 2 t + 1 + +  is the famous Euler identity (5). The identities (16) and (24) both represent Euler sums of even weight. For  t = 6 ,
2 S 2 , 12 + 4095 4096 S 2 , 12 + + S 12 , 2 + + 12 S 1 , 13 + = 75 , 989 , 887 1 , 720 , 320 ζ ( 14 ) 189 32 ζ ( 7 ) 2 375 32 ζ ( 5 ) ζ ( 9 ) 5883 512 ζ ( 11 ) ζ ( 3 ) .
Example 2.
From Theorem 1, consider the case  q = 4 ,  such that
1 t 4 t 1 + t S 1 , t + 2 + + ( 4 ) + S 2 , t + 1 + + ( 1 4 )
= 1 t + 1 4 t 1 S 2 , t + 1 + + ( 4 ) + 1 4 ζ ( 3 ) ζ ( t ) + 1 t t 4 1 t ζ ( 2 ) ζ ( t + 1 )
+ j = 3 t 1 j 1 t 4 t + 1 j t + 1 j ζ ( j ) ζ t + 3 j + 1 4 j 1 ζ ( j + 1 ) ζ t + 2 j ,
The Euler sum  S 2 , t + 1 + + ( 4 )  can be evaluated utilizing the method described in ref. [18], such that, for the case where t is even,
S 2 , t + 1 + + ( 4 ) = 1 8 3 × 2 2 t + 2 3 × 2 t + 1 + 3 × 2 t 1 ζ ( 2 ) ζ ( t + 1 )
+ 2 2 t 1 π t + 1 β t + 2 + 1 32 2 2 t + 1 t + 2 t + 1 t + 2 64 ζ ( t + 3 )
+ j = 1 t + 1 2 4 t + 1 2 j t + 2 2 j ζ ( 2 j ) ζ t + 3 2 j ,
Here, the Dirichlet beta function  β ( z )  (see Kölbig [19]) is defined by
β ( z ) = k = 1 ( 1 ) k + 1 ( 2 k 1 ) z ( ( z ) > 0 ) .
By substituting (27) and simplifying the expression (26), we conclude that the symmetrical reciprocal identity
1 4 t 1 + t S 1 , t + 2 + + ( 4 ) + S 2 , t + 1 + + ( 1 4 ) = 1 4 ζ ( 3 ) ζ ( t ) 2 π t + 1 β t + 2
+ 1 2 2 t 5 1 3 × 2 2 t + 2 + 3 × 2 t + 1 3 × 2 t + 8 t ζ ( 2 ) ζ ( t + 1 )
+ 1 2 2 t 2 2 2 t + 1 t + 2 + t + 1 t + 2 64 1 32 ζ ( t + 3 )
j = 1 t + 1 2 4 2 2 j t + 2 2 j ζ ( 2 j ) ζ t + 3 2 j
+ j = 3 t 1 j 1 4 t + 1 j t + 1 j ζ ( j ) ζ t + 3 j + 1 4 j 1 ζ ( j + 1 ) ζ t + 2 j .
For  t = 4 , we obtain
5 4 4 S 1 , 6 + + ( 4 ) + S 2 , 5 + + ( 1 4 ) = 49 , 159 1024 ζ ( 7 ) 10 π β 6 5031 512 ζ ( 2 ) ζ ( 5 ) + 3 16 ζ ( 4 ) ζ ( 3 ) .

4. Concluding Remarks

The scope of this study has been demonstrated in the development of Theorems 1 and 2, together with their respective corollaries. The examples have clearly demonstrated the applicability of the theorems and given new results for the linear arrangement of Euler sum identities with even weight. In particular, we have obtained closed-form identities of linear arrangements of Euler harmonic number sums of the type (4), (11), and (18) with arbitrary argument and with weight t + 3 for t N . The linear arrangement of Euler harmonic number sums with even weight, expressed in closed form, have been given in (15), (17), (22), (24), and (25). It is expected that further research will be undertaken in the evaluation of Euler linear harmonic sums identities with multiple arguments.

Funding

This research received no external funding.

Data Availability Statement

Data are contained within the article.

Acknowledgments

The author wishes to extend his profound thanks to the anonymous reviewers for their insightful feedback, which was constructive and has undoubtedly led to significant improvements in the quality of this paper.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Flajolet, P.; Salvy, B. Euler sums and contour integral representations. Exp. Math. 1998, 7, 15–35. [Google Scholar] [CrossRef]
  2. Srivastava, H.; Choi, J. Zeta and q-Zeta Functions and Associated Series and Integrals; Elsevier, Inc.: Amsterdam, The Netherland, 2012; p. xvi+657. ISBN 978-0-12-385218-2. [Google Scholar]
  3. Borwein, J.M.; Crandall, R.E. Closed forms: What they are and why we care. Not. Am. Math. Soc. 2013, 60, 50–65. [Google Scholar] [CrossRef]
  4. Euler, L. Opera Omnia; Series 1; Teubner: Berlin, Germany, 1917; Volume AT, pp. 217–267. [Google Scholar]
  5. Borwein, D.; Borwein, J.M.; Girgensohn, R. Explicit evaluation of Euler sums. Proc. Edinb. Math. Soc. 1995, 38, 277–294. [Google Scholar] [CrossRef]
  6. Sitaramachandrarao, R. A Formula of S. Ramanujan. J. Number Theory 1987, 25, 1–19. [Google Scholar] [CrossRef]
  7. Alzer, A.; Choi, J. Four parametric linear Euler sums. J. Math. Anal. Appl. 2020, 484, 123661. [Google Scholar] [CrossRef]
  8. Choi, J.; Srivastava, H.M. Explicit evaluation of Euler and related sums. Ramanujan J. 2005, 10, 51–70. [Google Scholar] [CrossRef]
  9. Sofo, A.; Nimbran, A.S. Euler sums and integral connections. Mathematics 2019, 7, 833. [Google Scholar] [CrossRef]
  10. Sofo, A. General order Euler sums with rational argument. Integral Transform. Spec. Funct. 2019, 30, 978–991. [Google Scholar] [CrossRef]
  11. Sofo, A. Reciprocal argument Euler sum identities. 2024; submitted. [Google Scholar]
  12. Sofo, A. General order Euler sums with multiple argument. J. Number Theory 2018, 189, 255–271. [Google Scholar] [CrossRef]
  13. Sofo, A.; Choi, J. Extension of the four Euler sums being linear with parameters and series involving the zeta functions. J. Math. Anal. Appl. 2022, 515, 126370. [Google Scholar] [CrossRef]
  14. Sofo, A.; Nimbran, A.S. Euler-like sums via powers of log, arctan and arctanh functions. Integral Transform. Spec. Funct. 2020, 31, 966–981. [Google Scholar] [CrossRef]
  15. Li, C.; Chu, W. Gauss’ Second Theorem for 2F1(1/2)-Series and Novel Harmonic Series Identities. Mathematics 2024, 12, 1381. [Google Scholar] [CrossRef]
  16. Chen, K.-W. On Some General Tornheim-Type Series. Mathematics 2024, 12, 1867. [Google Scholar] [CrossRef]
  17. Sofo, A.; Batir, N. Parameterized families of polylog integrals. Constuctive Math. Anal. 2021, 4, 400419. [Google Scholar] [CrossRef]
  18. Sofo, A. Generalization of the Flajolet Salvy identities. 2024; submitted. [Google Scholar]
  19. Kölbig, K.S. The polygamma function ψk(x) for x =  1 4 and x =  3 4 . J. Comput. Appl. Math. 1996, 75, 43–46. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Sofo, A. Linear Arrangement of Euler Sums with Multiple Argument. Symmetry 2024, 16, 1322. https://doi.org/10.3390/sym16101322

AMA Style

Sofo A. Linear Arrangement of Euler Sums with Multiple Argument. Symmetry. 2024; 16(10):1322. https://doi.org/10.3390/sym16101322

Chicago/Turabian Style

Sofo, Anthony. 2024. "Linear Arrangement of Euler Sums with Multiple Argument" Symmetry 16, no. 10: 1322. https://doi.org/10.3390/sym16101322

APA Style

Sofo, A. (2024). Linear Arrangement of Euler Sums with Multiple Argument. Symmetry, 16(10), 1322. https://doi.org/10.3390/sym16101322

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop