Next Article in Journal
Parseval–Goldstein-Type Theorems for Lebedev–Skalskaya Transforms
Previous Article in Journal
Elephant Random Walk with a Random Step Size and Gradually Increasing Memory and Delays
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

On Closed Forms of Some Trigonometric Series

by
Slobodan B. Tričković
1,* and
Miomir S. Stanković
2
1
Department of Mathematics, University of Niš, 18000 Niš, Serbia
2
Mathematical Institute of Serbian Academy of Sciences and Arts, 11000 Belgrade, Serbia
*
Author to whom correspondence should be addressed.
Axioms 2024, 13(9), 631; https://doi.org/10.3390/axioms13090631
Submission received: 15 August 2024 / Revised: 11 September 2024 / Accepted: 13 September 2024 / Published: 14 September 2024
(This article belongs to the Special Issue Special Functions and Related Topics)

Abstract

:
We have derived alternative closed-form formulas for the trigonometric series over sine or cosine functions when the immediate replacement of the parameter appearing in the denominator with a positive integer gives rise to a singularity. By applying the Choi–Srivastava theorem, we reduce these trigonometric series to expressions over Hurwitz’s zeta function derivative.

1. Preliminaries

All particular cases of the general summation formula for the trigonometric series
T α f = n = 1 ( s ) n 1 f ( ( a n b ) x ) ( a n b ) α ,
with parameters s = 1 or s = 1 , f = sin or f = cos , a = 1 2 b = 0 1 , α > 0 , we derived in [1] and expressed as a single formula via the power series
T α f = c π x α 1 2 Γ ( α ) f ( π α 2 ) + k = 0 ( 1 ) k F ( α 2 k δ ) ( 2 k + δ ) ! x 2 k + δ ,
and one can obtain each of them by looking up in Table 1 and taking the corresponding parameters, including s , a , b , f , c and δ , with F standing for Riemann’s ζ function, initially defined by the series [2]
ζ ( z ) = n = 1 1 n z , Re z > 1 ,
or Dirichlet η , λ , β functions defined as follows:
η ( z ) = n = 1 ( 1 ) n 1 n z , λ ( z ) = n = 1 1 ( 2 n 1 ) z , β ( z ) = n = 1 ( 1 ) n 1 ( 2 n 1 ) z .
Riemann’s ζ function satisfies the functional equation [3]
ζ ( 1 z ) = 2 ζ ( z ) ( 2 π ) z Γ ( z ) cos π z 2 ,
where Γ ( s ) is the gamma function [4]
Γ ( z ) = 0 x z 1 e x d x , Re z > 0 ,
introduced by Euler. On the other hand, integration by parts of the integral (6) gives rise to the basic relation
Γ ( z + 1 ) = z Γ ( z ) ,
which can be extended for arbitrary n N :
Γ ( z + n ) = ( z + n 1 ) ( z + n 2 ) z Γ ( z ) ,
and this formula one can use to express the Pochhammer symbol:
( z ) n = z ( z + 1 ) ( z + 2 ) ( z + n 1 ) = Γ ( z + n ) Γ ( z ) .
Rewriting integral (6) as
Γ ( z ) = 0 x z 1 e x d x = n = 0 ( 1 ) n n ! 0 1 x z + n 1 d x + 1 x s 1 e x d x = n = 0 ( 1 ) n n ! 1 z + n + 1 x z 1 e x d x
provides the analytic continuation of the gamma function for all complex numbers, except for integers less than or equal to zero.
For z Z , Euler’s reflection formula [4] (p. 35) holds:
Γ ( 1 z ) Γ ( z ) = π sin π z .
The first logarithmic derivative of the gamma function [5] is the function ψ (also known as the digamma), i.e.,
ψ ( s ) = d d s log Γ ( s ) = Γ ( s ) Γ ( s ) ,
whence we find ψ ( 1 ) = Γ ( 1 ) . It is connected to the nth harmonic number H n , given as the sum of reciprocal values of the first n positive integers by the equality
H n 1 = ψ ( n ) + γ , n N , H 0 = 0 ,
where γ is the Euler–Mascheroni constant, obtained as the limiting value of the sequence H n log n as n .
As an extension of the Riemann zeta function the Hurwitz zeta function appears, initially defined by the series [2]
ζ ( z , a ) = k = 0 1 ( k + a ) z , Re z > 1 , , a 0 , 1 , 2 , .
Obviously, ζ ( z , 1 ) = ζ ( z ) . A straightforward consequence of (12) is
ζ ( z , a ) = 1 a z + 1 ( 1 + a ) z + 1 ( 2 + a ) z + = a z + ζ ( z , a + 1 ) ,
and because of that, for most purposes, it suffices to restrict a to 0 < Re a 1 . In addition, most references treat a as a real number with 0 < a 1 . Considering a function of a, with z 1 fixed, ζ ( z , a ) is analytic in the half-plane Re a > 0 .
The Hurwitz zeta function satisfies an identity that generalizes the functional equation of the Riemann zeta function [3] (p. 257, Theorem 12.6)
ζ ( 1 z , a ) = Γ ( z ) ( 2 π ) z e i π z / 2 k = 1 e 2 k π i a k z + e i π z / 2 k = 0 e 2 k π i a k z
but we rewrite it in a more suitable form:
ζ ( 1 z , a ) = 2 Γ ( z ) ( 2 π ) z n = 1 1 n z cos ( π z 2 2 n π a ) .
For z = 1 in (14), by relying on Euler’s famous formula e i φ = cos φ + i sin φ and elementary properties of the complex logarithm, we evaluate
ζ ( 0 , a ) = 1 π n = 1 sin 2 n π a n = i 2 π log 1 e 2 i π a 1 e 2 i π a = 1 2 a .
Multiplying ζ ( z , a ) by Γ ( z ) , taking account of (12) and (6), we have
ζ ( z , a ) Γ ( z ) = k = 0 1 ( k + a ) z 0 x z 1 e x d x .
Introducing the substitution x = ( k + a ) t , then interchanging the sum and integral, we obtain the integral representation of the Hurwitz zeta function:
ζ ( z , a ) = 1 Γ ( z ) 0 t z 1 e a t k = 0 e k t d t = 1 Γ ( z ) 0 t z 1 e a t 1 e t d t .
Given (4), one easily derives the relations of the ζ function to η and λ :
η ( z ) = ( 1 2 1 z ) ζ ( z ) , λ ( z ) = ( 1 2 z ) ζ ( z ) .
We conclude that the η function extends analytically to the whole complex plane, including z = 1 , since
η ( 1 ) = lim z 1 η ( z ) = lim z 1 ( 1 2 1 z ) ζ ( z ) = lim z 1 1 2 1 z z 1 ( z 1 ) ζ ( z ) .
Further, using (5), then setting a = 1 in (15), we find
η ( 1 ) = log 2 lim z 1 π 2 ( z 1 ) sin π 2 ( 1 z ) ( 2 π ) z ζ ( 1 z ) π Γ ( z ) = 2 ζ ( 0 ) log 2 = log 2 .
However, (17) tells us that, as for the ζ function, λ is analytical for all complex numbers except z = 1 , where it has a non-removable singularity. In addition, (17) immediately gives rise to η ( 2 n ) = λ ( 2 n ) = 0 , n N , because by setting s = 2 n + 1 , n N , in the functional equation for the Riemann zeta function (5), we find
ζ ( 2 n ) = 2 ( 2 n ) ! ( 2 π ) z ζ ( 2 n + 1 ) cos π 2 ( 2 n + 1 ) = 0 .
Because of
n = 0 ( 1 ) n ( 2 n + 1 ) s = 1 1 3 s + 1 5 s 1 7 s + = n = 0 1 ( 4 n + 1 ) z n = 0 1 ( 4 n + 3 ) z = 1 4 z n = 0 1 ( n + 1 4 ) z n = 0 1 ( n + 3 4 ) z = 1 4 z ζ s , 1 4 ζ s , 3 4 ,
we express the β function through a difference of two Hurwitz functions:
β ( z ) = 4 s ζ s , 1 4 ζ s , 3 4 .
Using (16), we find its integral, represented as
β ( z ) = 1 4 z ζ s , 1 4 ζ s , 3 4 = 1 Γ ( z ) 0 t z 1 e t 4 1 + e t 2 d t = 1 Γ ( s ) 0 x s 1 e x 1 + e 2 x d x ,
which defines β as an analytical function for Re s > 0 , but employing the functional equations conjectured by Euler in 1749 and proved by the Swedish mathematician Malmsten in 1842,
β ( 1 z ) = 2 π z sin π s 2 Γ ( z ) β ( z ) .
β extends to the whole complex plane. By virtue of (9), Equation (20) can be rewritten as follows:
β ( z ) = π 2 z 1 β ( 1 z ) Γ ( 1 z ) cos π z 2 ,
whence we find β ( 2 n + 1 ) = 0 , n N .
Thus, owing to
ζ ( 2 n ) = η ( 2 n ) = λ ( 2 n ) = β ( 2 n + 1 ) = 0 , n N
the right-hand side series in (2) truncates. Thus, we obtain [1], one type of closed-form formula. So, for α = 2 m + p 1 , where m N and p = 0 or p = 1 , we have all these cases comprised by the general formula [1]
T 2 m + p 1 f = n = 1 ( s ) n 1 f ( ( a n b ) x ) ( a n b ) 2 m + p 1 = c π x 2 m + p 2 2 ( 2 m + p 2 ) ! f ( m π + π 2 ( p 1 ) ) + k = 0 m + p 1 ( 1 ) k F ( 2 m + p 1 2 k δ ) ( 2 k + δ ) ! x 2 k + δ .
We can obtain each from Table 1 by choosing the corresponding parameters.

2. Alternative Closed-Form Formulas

For a = 1 , b = 0 , s = 1 in (1), and putting f = sin , then f = cos , we have
n = 1 sin n x n α = π x α 1 2 Γ ( α ) sin π 2 α + k = 0 ( 1 ) k ζ ( α 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 , n = 1 cos n x n α = π x α 1 2 Γ ( α ) cos π 2 α + k = 0 ( 1 ) k ζ ( α 2 k ) ( 2 k ) ! x 2 k , 0 < x < 2 π ,
where ζ presents Riemann’s zeta function.
However, by taking limits α 2 m and α 2 m 1 in the first and the second formula of (22), respectively, one encounters singularities of the ζ function, so we have to act differently.
Theorem 1. 
Letting α 2 m in the first formula of (22), one brings the sine series in closed form:
n = 1 sin n x n 2 m = ( 1 ) m ( 2 π ) 2 m 1 ( 2 m 1 ) ! ζ 1 2 m , 1 x 2 π ζ 1 2 m , x 2 π ,
where 0 < x < 2 π . This series defines the Clausen function Cl 2 m [1] (p. 452).
Proof. 
We are not allowed to replace α with 2 m in the first formula of (22) since sin π 2 α = 0 and in the term for k = m 1 in the series, and we encounter singularities. We recall the Hurwitz formula connecting a trigonometric series and the Hurwitz zeta function.
In the first derivative of the Formula (14) for z = 2 m and a = x 2 π , we have
ζ 1 2 m , x 2 π = 2 ( 1 ) m Γ ( 2 m ) ( 2 π ) 2 m ψ ( 2 m ) log 2 π n = 1 cos n x n 2 m 2 ( 1 ) m Γ ( 2 m ) ( 2 π ) 2 m n = 1 cos n x n 2 m log n π ( 1 ) m 1 Γ ( 2 m ) ( 2 π ) 2 m n = 1 sin n x n 2 m .
For a = 1 x 2 π , we have
ζ 1 2 m , 1 x 2 π = 2 ( 1 ) m Γ ( 2 m ) ( 2 π ) 2 m ψ ( 2 m ) log 2 π n = 1 cos n x n 2 m 2 ( 1 ) m Γ ( 2 m ) ( 2 π ) 2 m n = 1 cos n x n 2 m log n π ( 1 ) m Γ ( 2 m ) ( 2 π ) 2 m n = 1 sin n x n 2 m .
By subtracting these equalities, we obtain
ζ 1 2 m , 1 x 2 π ζ 1 2 m , x 2 π = ( 1 ) m ( 2 m 1 ) ! ( 2 π ) 2 m 1 n = 1 sin n x n 2 m .
Hence, we obtain the sum of the sine series. □
Theorem 2. 
Letting α 2 m 1 in the second formula of (22), one brings the cosine series in closed form:
n = 1 cos n x n 2 m 1 = ( 1 ) m 1 ( 2 π ) 2 m 2 ( 2 m 2 ) ! ζ 2 2 m , 1 x 2 π + ζ 2 2 m , x 2 π ,
where 0 < x < 2 π . This series defines the Clausen function Cl 2 m 1 [1] (p. 452).
Proof. 
Similarly, one cannot immediately replace α with 2 m 1 , so it is necessary to perform a different method. Taking the first derivative of Formula (14) with respect to z for a = x 2 π , we have
ζ 1 s , x 2 π = 2 Γ ( s ) ( 2 π ) s ψ ( s ) log 2 π n = 1 cos π s 2 n x n s 2 Γ ( s ) ( 2 π ) s n = 1 cos π s 2 n x n s log n π Γ ( s ) ( 2 π ) s n = 1 sin π s 2 n x n s .
Setting here z = 2 m 1 , we find
ζ 2 2 m , x 2 π = 2 ( 1 ) m 1 Γ ( 2 m 1 ) ( 2 π ) 2 m 1 ψ ( 2 m 1 ) log 2 π n = 1 sin n x n 2 m 1 2 ( 1 ) m 1 Γ ( 2 m 1 ) ( 2 π ) 2 m 1 n = 1 sin n x n 2 m 1 log n π ( 1 ) m 1 Γ ( 2 m 1 ) ( 2 π ) 2 m 1 n = 1 cos n x n 2 m 1 .
For a = 1 x 2 π , we have
ζ 2 2 m , 1 x 2 π = 2 ( 1 ) m Γ ( 2 m 1 ) ( 2 π ) 2 m 1 ψ ( 2 m 1 ) log 2 π n = 1 sin n x n 2 m 1 2 ( 1 ) m Γ ( 2 m 1 ) ( 2 π ) 2 m 1 n = 1 sin n x n 2 m 1 log n π ( 1 ) m 1 Γ ( 2 m 1 ) ( 2 π ) 2 m 1 n = 1 cos n x n 2 m 1 .
By adding these equalities, we obtain
ζ 2 2 m , 1 x 2 π + ζ 2 2 m , x 2 π = ( 1 ) m 1 ( 2 m 2 ) ! ( 2 π ) 2 m 2 n = 1 cos n x n 2 m 1 .
Thence, we obtain the sum of the cosine series. □
For a = 1 , b = 0 , s = 1 and putting f = sin in (1), based on (22) and using its relation to the zeta function, (17), we obtain the alternating series over the sines and cosines expressed in terms of the Dirichlet eta function:
n = 1 ( 1 ) n 1 sin n x n α = n = 1 ( 1 ) n 1 sin n x n α 1 2 α 1 n = 1 sin 2 n x n α = k = 0 ( 1 ) k η ( α 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 , π < x < π .
Similarly, for a = 1 , b = 0 , s = 1 and putting f = cos in (1), we find
n = 1 ( 1 ) n 1 cos n x n α = k = 0 ( 1 ) k η ( α 2 k ) ( 2 k ) ! x 2 k , π < x < π .
Theorem 3. 
For α = 2 m , the series (25) takes the closed form
n = 1 ( 1 ) n 1 sin n x n 2 m = ( 1 ) m π 2 m 1 ( 2 m 1 ) ! ( 2 2 m 1 ζ 1 2 m , 1 x 2 π 2 2 m 1 ζ 1 2 m , x 2 π ζ 1 2 m , 1 x π + ζ 1 2 m , x π ) ,
Proof. 
These replacements are legitimate since we encounter no singularities in (25).
Further, from (25), we have
n = 1 ( 1 ) n 1 sin n x n 2 m = k = 0 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 ,
and split the right-hand side series as follows:
k = 0 m 2 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 + ( 1 ) m 1 x 2 m 1 ( 2 m 1 ) ! log 2 + k = m ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 .
Taking account of η ( 2 m 2 k 1 ) = ( 1 2 1 ( 2 m 2 k 1 ) ) ζ ( 2 m 2 k 1 ) , we can express the last right-hand side series in terms of the zeta function as a difference of two series:
k = m ( 1 ) k ζ ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 k = m ( 1 ) k ζ ( 2 m 2 k 1 ) 2 2 m 2 k 2 ( 2 k + 1 ) ! x 2 k + 1 .
If we set z = 2 n in (5), we can determine values of Riemann’s zeta function at negative odd integers:
ζ ( 1 2 n ) = ( 1 ) n 2 ( 2 n 1 ) ! ζ ( 2 n ) ( 2 π ) 2 n .
By shifting indices in both series of (29) and applying (30) and (8), the difference (29) becomes
2 ( 1 ) m 1 x 2 m 1 k = 1 ζ ( 2 k ) ( 2 k ) 2 m x 2 π 2 k k = 1 ζ ( 2 k ) ( 2 k ) 2 m x π 2 k .
where ( 2 k ) 2 m denotes Pochhammer’s symbol, given by (8). We refer now to the theorem in [6] (p. 419), which we state here in a slightly modified form.
Theorem 1. 
For every non-negative integer n there holds
k = 2 ζ ( k , a ) ( k ) n + 1 t n + k = ( 1 ) n n ! ( ζ ( n , a t ) ζ ( n , a ) + k = 1 n ( t ) k n k × ζ ( k n , a ) ( H n H n k ) ζ ( k n , a ) ) + H n + ψ ( a ) t n + 1 ( n + 1 ) ! ,
with | t | < | a | , n N 0 .
For the first series of (31), we set in (32) a = 1 , n = 2 m 1 , t = x 2 π , then again, a = 1 , n = 2 m 1 , t = x 2 π , and subtract the second equality from the first. In this way, we obtain
2 ( 1 ) m 1 x 2 m 1 k = 1 ζ ( 2 k ) ( 2 k ) 2 m x 2 π 2 k = ( 1 ) m ( 2 π ) 2 m 1 ( 2 m 1 ) ! × ( ζ 2 m + 1 , 1 x 2 π ζ 2 m + 1 , 1 + x 2 π 2 k = 1 m 2 m 1 2 k 1 × x 2 π 2 k 1 ( H 2 m 1 H 2 m 2 k ) ζ ( 2 k 2 m ) ζ ( 2 k 2 m ) ) .
Knowing that ζ ( 2 k 2 m ) = 0 for k = 1 , , m 1 and ζ ( 0 ) = 1 / 2 , H 0 = 0 , the right-hand side sum becomes
x 2 π 2 m 1 H 2 m 1 + 2 k = 1 m 2 m 1 2 k 1 x 2 π 2 k 1 ζ ( 2 k 2 m ) .
Taking the derivative with respect to s on both sides of (13), then putting there a = x 2 π and s = 1 2 m , we find
ζ 1 2 m , 1 + x 2 π = x 2 π 2 m 1 log x 2 π ζ 1 2 m , x 2 π .
We obtain the same structure for the second series, with the sole difference that instead of x 2 π , there appears x π . The subtraction yields the sum of (31), which is the sum of the right-hand side series in (28). So, for (31) we find
( 1 ) m π 2 m 1 ( 2 m 1 ) ! ( 2 2 m 1 ζ 1 2 m , 1 x 2 π 2 2 m 1 ζ 1 2 m , x 2 π ζ 1 2 m , 1 x π + ζ 1 2 m , x π + x π 2 m 1 log 2 + 2 2 m k = 1 m 2 m 1 2 k 1 x 2 π 2 k 1 ( 1 2 1 ( 2 m 2 k + 1 ) ) ζ ( 2 k 2 m ) ) .
By differentiating (5) we can evaluate ζ ( 2 n ) for positive integers n. So, the left-hand side is ζ ( 1 s ) , but on the right-hand side, we obtain a sum of four terms; and notice that in three of them, cos π s 2 appears while only one contains sin π s 2 , i.e.,
π ζ ( s ) ( 2 π ) s Γ ( s ) sin π s 2 .
With s = 2 n + 1 , all the terms except for the latter become zero, and we find
ζ ( 2 n ) = ( 1 ) n ( 2 n ) ! ζ ( 2 n + 1 ) 2 ( 2 π ) 2 n , n N .
Because of (34) and η ( s ) = ( 1 2 1 s ) ζ ( s ) , for the last row of (33), we have
( 1 ) m ( 2 m 1 ) ! π 2 m 1 k = 0 m 2 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 .
Adding this modified Formula (33) to (28), after a rearrangement, we arrive at (27). □
Theorem 5. 
For α = 2 m 1 , the series (26) takes the closed form
n = 1 ( 1 ) n 1 cos n x n 2 m 1 = ( 1 ) m 1 π 2 m 2 ( 2 m 2 ) ! ( 2 2 m 2 ζ 2 2 m , 1 x 2 π + 2 2 m 2 ζ 2 m + 2 , x 2 π ζ 2 2 m , 1 x π ζ 2 2 m , x π ) .
Proof. 
Following a similar procedure as in the proof of the preceding theorem, we replace α with 2 m 1 in (26) and have
n = 1 ( 1 ) n 1 cos n x n 2 m 1 = k = 0 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k .
The function η ( s ) is analytic in the whole complex plane, so to bring the latter series into closed form, we express it first as a sum comprising the value of η ( 1 ) , obtained for k = m 1 , and its remainder, i.e.,
k = 0 m 2 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k + ( 1 ) m 1 x 2 m 2 ( 2 m 2 ) ! log 2 + k = m ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k .
Further, we act in the same manner as in the previous proof. Relying on the relation η ( s ) = ( 1 2 1 s ) ζ ( s ) and (30), we determine the remainder in (36) as a difference of the two series
k = m ( 1 ) k ζ ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k k = m ( 1 ) k ζ ( 2 m 2 k 1 ) 2 2 m 2 k 2 ( 2 k ) ! x 2 k = 2 ( 1 ) m 1 x 2 m 2 k = 1 ζ ( 2 k ) ( 2 k ) 2 m 1 x 2 π 2 k k = 1 ζ ( 2 k ) ( 2 k ) 2 m 1 x π 2 k .
Applying again Theorem 4, where we take n = 2 m 2 and a = 1 , then in succession set t = x 2 π and t = x π , we obtain two formulas of the same structure. Subtracting them gives rise to the sum of the right-hand side series in (36), which is
( 1 ) m 1 π 2 m 2 ( 2 m 2 ) ! ( 2 2 m 2 ζ 2 2 m , 1 x 2 π + 2 2 m 2 ζ 2 2 m , x 2 π ζ 2 2 m , 1 x π ζ 2 2 m , x π x π 2 m 2 log 2 ( 1 ) m 1 ( 2 m 2 ) ! π 2 m 2 k = 0 m 2 ( 1 ) k η ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k ) .
Adding (37) to (36), we obtain (27). □
By using the identity ζ ( s ) + η ( s ) = 2 λ ( s ) , and summing up the first and second equations of (22) with (25) and (26), respectively, we obtain
n = 1 sin ( 2 n 1 ) x ( 2 n 1 ) α = π x α 1 4 Γ ( α ) sin π 2 α + k = 0 ( 1 ) k λ ( α 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 , n = 1 cos ( 2 n 1 ) x ( 2 n 1 ) α = π x α 1 4 Γ ( α ) cos π 2 α + k = 0 ( 1 ) k λ ( α 2 k ) ( 2 k ) ! x 2 k , 0 < x < π ,
Theorem 6. 
If α 2 m in the first formula of (38), the following holds:
n = 1 sin ( 2 n 1 ) x ( 2 n 1 ) 2 m = ( 1 ) m π 2 m 1 2 ( 2 m 1 ) ! ( 2 2 m ζ 1 2 m , 1 x 2 π 2 2 m ζ 1 2 m , x 2 π ζ 1 2 m , 1 x π + ζ 1 2 m , x π ) .
Proof. 
Since sin 2 m π 2 = 0 , in the first formula of (38) for α = 2 m singularity is encountered, and for k = m 1 there is singularity of the lambda function at 1, so we are not permitted to immediately replace α with 2 m but have to take the limit
L 2 m = lim α 2 m π x α 1 4 Γ ( α ) sin π 2 α + ( 1 ) m 1 λ ( α 2 m + 1 ) ( 2 m 1 ) ! x 2 m 1 .
Further, by bringing the fractions to the same denominator, we have
lim α 2 m π x α 1 ( 2 m 1 ) ! + 4 ( 1 ) m 1 x 2 m 1 λ ( α 2 m + 1 ) Γ ( α ) sin π 2 α 4 ( 2 m 1 ) ! Γ ( α ) sin π 2 α .
The denominator tends to zero as α tends to 2 m .
As for the numerator, knowing that λ ( s ) = ( 1 2 s ) ζ ( s ) , we have
lim α 2 m π x α 1 ( 2 m 1 ) ! + 4 x 2 m 1 Γ ( α ) λ ( α 2 m + 1 ) ( 1 ) m 1 sin π 2 α = π x 2 m 1 ( 2 m 1 ) ! 1 2 lim α 2 m ( α 2 m ) λ ( α 2 m + 1 ) sin π 2 ( α 2 m ) π 2 ( α 2 m ) = 0 ,
where, relying on the functional equation for the Riemann zeta function (5), we make use of the limiting value
lim z 0 z λ ( z + 1 ) = lim z 1 ( z 1 ) λ ( z ) = lim z 1 ( 2 π ) z ( z 1 ) ( 1 2 z ) ζ ( 1 z ) 2 Γ ( z ) cos z π 2 = lim z 1 π z 1 π 2 ( 1 z ) ζ ( 1 z ) Γ ( z ) sin π 2 ( 1 z ) = ζ ( 0 ) = 1 2 .
So, we can apply L’Hopital’s rule here. For the limiting value of the first derivative of the denominator, we find
4 ( 2 m 1 ) ! lim α 2 m π 2 cos π α 2 Γ ( α ) + sin π α 2 Γ ( α ) ψ ( α ) = ( 1 ) m 2 π ( 2 m 1 ) ! 2 ,
Now, we take the limit of the first derivative of the numerator
π ( 2 m 1 ) ! x 2 m 1 log x + ( 1 ) m 1 x 2 m 1 lim α 2 m ( 4 λ ( α 2 m + 1 ) Γ ( α ) sin π α 2 + 4 λ ( α 2 m + 1 ) Γ ( α ) ψ ( α ) sin π α 2 + 2 π λ ( α 2 m + 1 ) Γ ( α ) cos π α 2 ) .
First of all, using the relation λ ( z ) = ( 1 2 z ) ζ ( z ) , we find
λ ( z ) λ ( z ) = ζ ( z ) ζ ( z ) + log 2 2 z 1 .
It is easy to see that in the neighborhood of z = 1 , there holds
log 2 2 z 1 = log 2 + O ( | z 1 | ) ,
and looking up in [7] (p. 23), we read
ζ ( s ) ζ ( z ) = 1 z 1 + γ + O ( | z 1 | ) .
Thus, there follows
λ ( α 2 m + 1 ) = λ ( α 2 m + 1 ) α 2 m + ( γ + log 2 ) λ ( α 2 m + 1 ) + O ( | α 2 m | ) .
By replacing this in (41), omitting O ( | α 2 m | ) , it remains to calculate
lim α 2 m Γ ( α ) λ ( α 2 m + 1 ) ( 2 π ( 1 ) m sin π 2 ( α 2 m ) π 2 ( α 2 m ) + 4 sin π α 2 ( ψ ( α ) + γ + log 2 ) + 2 π cos π α 2 ) .
Because of lim α 2 m cos π α 2 = cos π m = ( 1 ) m , we conclude that the limiting value of the sum of only the first and third terms is zero, and finding the limit reduces to
lim α 2 m 2 ( 1 ) m π Γ ( α ) sin π 2 ( α 2 m ) π 2 ( α 2 m ) ( α 2 m ) λ ( α 2 m + 1 ) ( ψ ( α ) + γ + log 2 ) = ( 1 ) m π ( 2 m 1 ) ! ( ψ ( 2 m ) + γ + log 2 ) = ( 1 ) m π ( 2 m 1 ) ! ( H 2 m 1 + log 2 ) .
We have applied here the relation (11).
Thus, taking account of all of this, the value of (40) is
L 2 m = ( 1 ) m x 2 m 1 ( log x H 2 m 1 log 2 ) 2 ( 2 m 1 ) ! ,
and the trigonometric series (39) can now be expressed as
n = 1 sin ( 2 n 1 ) x ( 2 n 1 ) 2 m = ( 1 ) m x 2 m 1 2 ( 2 m 1 ) ! log x 2 H 2 m 1 + k = 0 m 2 ( 1 ) k λ ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 + k = m ( 1 ) k λ ( 2 m 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 ,
and after shifting the summation index in the last series, making use of the relation λ ( 2 m 2 k 1 ) = ( 1 2 ( 2 m 2 k 1 ) ) ζ ( 2 m 2 k 1 ) and applying (30), one can represent it as a difference of two series in terms of the zeta function:
( 1 ) m 1 x 2 m 1 2 k = 1 ζ ( 2 k ) ( 2 k ) 2 m x 2 π 2 k k = 1 ζ ( 2 k ) ( 2 k ) 2 m x π 2 k .
We have already dealt with these series in Formula (31), but here we express this difference as follows:
2 ( 1 ) m 1 x 2 m 1 k = 1 ζ ( 2 k ) ( 2 k ) 2 m x 2 π 2 k ( 1 ) m 1 x 2 m 1 k = 1 ζ ( 2 k ) ( 2 k ) 2 m x π 2 k = ( 1 ) m π 2 m 1 2 ( 2 m 1 ) ! ( 2 2 m ζ 1 2 m , 1 x 2 π 2 2 m ζ 2 m + 1 , 1 + x 2 π + x π 2 m 1 ( H 2 m 1 log 2 π ) ζ 1 2 m , 1 x π + ζ 1 2 m , 1 + x π 2 ( 1 ) m ( 2 m 1 ) ! π 2 m 1 k = 0 m 2 x 2 k + 1 ( 2 k + 1 ) ! ( 1 ) k λ ( 2 m 2 k 1 ) ) .
Adding the right-hand side to (42), after a rearrangement, and using (13), we obtain (39). □
Theorem 7. 
If α 2 m 1 in the second formula of (38), it holds that
n = 1 cos ( 2 n 1 ) x ( 2 n 1 ) 2 m 1 = ( 1 ) m 1 π 2 m 2 2 ( 2 m 2 ) ! ( 2 2 m 1 ζ 2 2 m , 1 x 2 π + 2 2 m 1 ζ 2 2 m , 1 + x 2 π ζ 2 2 m , 1 x π ζ 2 2 m , 1 + x π ) .
Proof. 
After taking the limit α 2 m 1 in the second formula of (38), acting in the same manner as in the proof of the previous theorem, we first find
n = 1 cos ( 2 n 1 ) x ( 2 n 1 ) 2 m 1 = ( 1 ) m x 2 m 2 2 ( 2 m 2 ) ! log x 2 H 2 m 2 + k = 0 m 2 ( 1 ) k λ ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k + k = m ( 1 ) k λ ( 2 m 2 k 1 ) ( 2 k ) ! x 2 k .
We deal again with the remainder and rewrite it as follows:
( 1 ) m 1 x 2 m 2 2 k = 1 ζ ( 2 k ) ( 2 k ) 2 m 1 x 2 π 2 k k = 1 ζ ( 2 k ) ( 2 k ) 2 m 1 x π 2 k .
Employing a similar procedure as in the case of the preceding theorem, we come to the closed-form Formula (43). □
For alternating series related to (38) we can express them as a power series involving Dirichlet’s beta function
n = 1 ( 1 ) n 1 sin ( 2 n 1 ) x ( 2 n 1 ) α = k = 0 ( 1 ) k β ( α 2 k 1 ) ( 2 k + 1 ) ! x 2 k + 1 , n = 1 ( 1 ) n 1 cos ( 2 n 1 ) x ( 2 n 1 ) α = k = 0 ( 1 ) k β ( α 2 k ) ( 2 k ) ! x 2 k , π 2 < x < π 2 .
Theorem 8. 
If α is replaced with 2 m 1 in the first formula of (45), we obtain the following closed form:
n = 1 ( 1 ) n 1 sin ( 2 n 1 ) x ( 2 n 1 ) 2 m 1 = ( 1 ) m 1 ( 2 π ) 2 m 2 2 ( 2 m 2 ) ! ( ζ 2 2 m , 1 4 x 2 π ζ 2 2 m , 3 4 x 2 π ζ 2 2 m , 1 4 + x 2 π + ζ 2 2 m , 3 4 + x 2 π ) ,
Proof. 
These replacements are legitimate because we do not encounter singularities since the function β is analytic in the whole complex plane; knowing that β ( 1 ) = arctg 1 = π 4 and by using (20), we easily calculate
β ( 0 ) = 2 π sin π 2 Γ ( 1 ) β ( 1 ) = 2 π · π 4 = 1 2 .
Further, from (45), we have
n = 1 ( 1 ) n 1 sin ( 2 n 1 ) x ( 2 n 1 ) 2 m 1 = k = 0 ( 1 ) k β ( 2 m 2 k 2 ) ( 2 k + 1 ) ! x 2 k + 1 ,
splitting the right-hand side series in three, but writing it for brevity now as
k = 0 m 2 ( x 2 ) k x β ( 2 m 2 k 2 ) ( 2 k + 1 ) ! ( 1 ) m x 2 m 1 2 ( 2 m 1 ) ! + k = m ( x 2 ) k x β ( 2 m 2 k 2 ) ( 2 k + 1 ) ! .
By shifting the summation index and making use of (21), the last series becomes
k = m ( x 2 ) k x β ( 2 m 2 k 2 ) ( 2 k + 1 ) ! = ( 1 ) m 1 x 2 m 2 k = 1 β ( 2 k + 1 ) ( 2 x π ) 2 k + 1 ( 2 k + 1 ) 2 m 1 .
Taking account of (18), we further change (47) to
( 1 ) m 1 x 2 m 2 k = 1 ζ 2 k + 1 , 1 4 ζ 2 k + 1 , 3 4 ( 2 k + 1 ) 2 m 1 x 2 π 2 k + 1
That means we are dealing with two series over the Hurwitz zeta functions and set n = 2 m 2 and t = x 2 π in (32), considering, apart from t, the same formula for t as well.
Replacing a in succession with 1 4 and 3 4 in (32), then subtracting these equalities, we obtain (48) on the left-hand side. The right-hand side consists of a sum of four terms, the first one being
( 1 ) m 1 ( 2 π ) 2 m 2 2 ( 2 m 2 ) ! ( ζ 2 m + 2 , 1 4 x 2 π ζ 2 m + 2 , 3 4 x 2 π ζ 2 m + 2 , 1 4 + x 2 π + ζ 2 m + 2 , 3 4 + x 2 π )
then of the two sums
( 1 ) m π 2 2 m 2 k = 1 m 1 2 x π 2 k 1 H 2 m 2 H 2 m 2 k 1 ( 2 m 2 k 1 ) ! ( 2 k 1 ) ! β ( 2 k 2 m + 1 ) + ( 1 ) m ( 2 π ) 2 m 2 ( 2 m 2 ) ! k = 1 m 1 x 2 π 2 k 1 2 m 2 2 k 1 4 2 k 2 m + 1 β ( 2 k 2 m + 1 )
and finally of
ψ ( 1 4 ) ψ ( 3 4 ) ( 1 ) m 1 ( 2 π ) 2 m 2 ( 2 m 1 ) ! x 2 π 2 m 1 .
For k = 1 , , m 1 , the expressions 2 k 2 m + 1 present negative odd integers, which means β ( 2 k 2 m + 1 ) = 0 , so the first sum in (49) equals zero.
Differentiating the relation (21) at z = 2 k 2 m + 1 , for 1 k m 1 , yields
β ( 2 k 2 m + 1 ) = π 2 2 k 2 m + 1 ( 1 ) k m Γ ( 2 m 2 k ) β ( 2 m 2 k ) ,
whereby the second sum in (49) becomes
k = 1 m 1 ( 1 ) k 1 β ( 2 m 2 k ) ( 2 k 1 ) ! = k = 0 m 2 ( 1 ) k β ( 2 m 2 k 2 ) ( 2 k + 1 ) ! .
Expression (50) is transformed into
4 β ( 1 ) ( 1 ) m 1 ( 2 π ) 2 m 2 ( 2 m 1 ) ! x 2 π 2 m 1 = 4 · π 4 ( 1 ) m x 2 m 1 2 π ( 2 m 1 ) ! = ( 1 ) m x 2 m 1 2 ( 2 m 1 ) ! ,
and this is obtained by setting n = 1 in the relation
β ( n ) = ( 1 ) n 2 2 n ( 2 n 1 ) ! ψ ( n 1 ) 1 4 ψ ( n 1 ) 3 4 , n N ,
which in turn we obtain from (18) by relying on the identity [5]
ψ ( n ) ( a ) = ( 1 ) n 1 n ! ζ ( n + 1 , a ) .
Collecting all the cases, we arrive at the closed-form Formula (46). □
Theorem 9. 
If α is replaced with 2 m in the second formula of (45), we obtain the following closed forms:
n = 1 ( 1 ) n 1 cos ( 2 n 1 ) x ( 2 n 1 ) 2 m = ( 1 ) m 1 ( 2 π ) 2 m 1 2 ( 2 m 1 ) ! ( ζ 1 2 m , 1 4 x 2 π ζ 1 2 m , 3 4 x 2 π + ζ 1 2 m , 1 4 + x 2 π ζ 1 2 m , 3 4 + x 2 π ) .
Proof. 
Acting in the same manner as in deducing Formula (46), we come to Formula (51). □
We can organize all these formulas by writing them as a general closed-form formula comprising the Formulas (23), (24), (27), (35), (39), (43), (46), and (51), i.e.,
n = 1 ( s ) n 1 f ( ( a n b ) x ) ( a n b ) 2 m + p 1 = ( 1 ) m + p 1 π 2 m + p 2 2 2 m + r 2 ( 2 m + p 2 ) ! × ( 2 2 k 2 ζ 2 p 2 m , q x 2 π + 2 2 k 2 δ ζ 2 p 2 m , 1 q + x 2 π j ζ 2 p 2 m , 1 q x 2 c π + δ ζ 2 p 2 m , q + x 2 c π ) ,
and one can obtain all its particular cases from Table 2 by choosing the corresponding parameters.
Thus, we have two types of closed-form formulas for each case of trigonometric series.

Author Contributions

Both authors contributed equally to the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Data are contained within the article.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Tričković, S.B.; Stanković, M.S.; Vidanović, M.V. On the summation of trigonometric series. Int. Trans. Spec. Func. 2008, 19, 441–452. [Google Scholar] [CrossRef]
  2. Abramowitz, M.; Stegun, A. Handbook of Mathematical Functions; Cambridge Universiy Press: New York, NY, USA, 2010. [Google Scholar]
  3. Apostol, T.M. Introduction to Analytic Number Theory; Springer: New York, NY, USA, 1976. [Google Scholar]
  4. Olver, F.W.J. Asymptotics and Special Functions; A K Peters: Wellesley, MA, USA, 1997. [Google Scholar]
  5. Bateman, H.; Erdélyi, A. Higher Transcendental Functions Volume I; McGraw-Hill Book Company, Inc.: New York, NY, USA; Toronto, ON, Canada; London, UK, 1953. [Google Scholar]
  6. Choi, J.; Srivastava, H.M. A certain family of series associated with the Zeta and related functions. Hiroshima Math. J. 2002, 32, 417–429. [Google Scholar] [CrossRef]
  7. Titchmarsh, E.C. The Theory of the Riemann Zeta-Function; Clarendon Press: Oxford, UK, 1986. [Google Scholar]
Table 1. Particular and closed-form cases of (2).
Table 1. Particular and closed-form cases of (2).
abscFf δ pConvergence Region
1011ζsin100 < x < 2π
cos01
−10ηsin10π < x < π
cos01
211 1 2 λsin100 < x < π
cos01
−10βsin11 π 2 < x < π 2
cos00
Table 2. Closed-form cases, α = 2 m + p 1 .
Table 2. Closed-form cases, α = 2 m + p 1 .
absfprc δ qkjConvergence Region
101sin11 −1110 0 < x < 2 π
cos00 1110
−1sin1 2 2 m 1 2 −11 m + 1 2 −1 π < x < π
cos0 2 2 m 1 2 11m1
211sin1 1 2 m 1 2 −11 m + 1 −1 0 < x < π
cos0 1 2 m 1 2 11 m + 1 2 1
−1sin0−111 1 4 11 π 2 < x < π 2
cos101−1 1 4 1−1
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

Tričković, S.B.; Stanković, M.S. On Closed Forms of Some Trigonometric Series. Axioms 2024, 13, 631. https://doi.org/10.3390/axioms13090631

AMA Style

Tričković SB, Stanković MS. On Closed Forms of Some Trigonometric Series. Axioms. 2024; 13(9):631. https://doi.org/10.3390/axioms13090631

Chicago/Turabian Style

Tričković, Slobodan B., and Miomir S. Stanković. 2024. "On Closed Forms of Some Trigonometric Series" Axioms 13, no. 9: 631. https://doi.org/10.3390/axioms13090631

APA Style

Tričković, S. B., & Stanković, M. S. (2024). On Closed Forms of Some Trigonometric Series. Axioms, 13(9), 631. https://doi.org/10.3390/axioms13090631

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