Next Article in Journal
Fixed Point Results for New Types of Fuzzy Contractions via Admissible Functions and FZ -Simulation Functions
Next Article in Special Issue
Cayley Graphs Defined by Systems of Equations
Previous Article in Journal
On Λ-Fractional Analysis and Mechanics
Previous Article in Special Issue
Balances in the Set of Arithmetic Progressions
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Fourier Series for the Tangent Polynomials, Tangent–Bernoulli and Tangent–Genocchi Polynomials of Higher Order

by
Cristina Bordaje Corcino
1,2,† and
Roberto Bagsarsa Corcino
1,2,*,†
1
Research Institute for Computational Mathematics and Physics, Cebu Normal University, Cebu City 6000, Philippines
2
Mathematics Department, Cebu Normal University, Cebu City 6000, Philippines
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
Axioms 2022, 11(3), 86; https://doi.org/10.3390/axioms11030086
Submission received: 11 December 2021 / Revised: 11 February 2022 / Accepted: 13 February 2022 / Published: 22 February 2022
(This article belongs to the Special Issue Discrete Mathematics as the Basis and Application of Number Theory)

Abstract

:
In this paper, the Fourier series expansion of Tangent polynomials of higher order is derived using the Cauchy residue theorem. Moreover, some variations of higher-order Tangent polynomials are defined by mixing the concept of Tangent polynomials with that of Bernoulli and Genocchi polynomials, Tangent–Bernoulli and Tangent–Genocchi polynomials. Furthermore, Fourier series expansions of these variations are also derived using the Cauchy residue theorem.

1. Introduction

In recent decades, several mathematicians have been attracted to work on some well-known special functions, numbers and polynomials (e.g., zeta functions, Bernoulli, Euler and Genocchi numbers and polynomials and derivative polynomials [1,2,3,4,5,6,7]) due to their wide-ranging applications, from number theory and combinatorics to other fields of applied mathematics. With these, different variations and generalizations of these functions, numbers and polynomials have appeared in the literature. Some variations were constructed by mixing the concepts of two or three special functions, numbers or polynomials. For instance, the poly-Bernoulli numbers and polynomials in [6,8,9,10,11] were constructed by mixing the concepts of polylogarithm and Bernoulli numbers and polynomials. Moreover, the Apostol–Genocchi polynomials, Frobenius–Euler polynomials, Frobenius–Genocchi polynomials and Apostol–Frobenius-type poly–Genocchi polynomials in [12,13,14,15] are constructed by mixing the concepts of Apostol, Frobenius, Genocchi and Euler polynomials. Another interesting mixture of special polynomials can be constructed by joining the concepts of Tangent polynomials with Bernoulli and Genocchi polynomials.
The Tangent polynomials T n ( x ) of degree n with complex argument x are defined as coefficients of the following generating function (see [16,17])
n = 0 T n ( x ) z n n ! = 2 e 2 z + 1 e x z ,
where T n ( 0 ) = T n , the Tangent numbers defined as coefficient of the following series expansion of the tangent function tan x (see [12]):
tan z = n = 0 ( 1 ) n + 1 T 2 n + 1 z 2 n + 1 ( 2 n + 1 ) ! ,
with T 0 = 1 and T 2 n = 0 , n N . The first few values of Tangent polynomials are given as follows:
T 0 ( x ) = 1 , T 1 ( x ) = x 1 , T 2 ( x ) = x 2 2 x , T 3 = x 3 3 x 2 + 2 ,
T 4 ( x ) = x 4 4 x 3 + 8 x , T 5 ( x ) = x 5 5 x 4 + 20 x 2 16 ,
T 6 ( x ) = x 6 6 x 5 + 40 x 3 96 x ,
where
T 0 ( 0 ) = 1 , T 1 ( 0 ) = 1 , T 2 = 0 , T 3 ( 0 ) = 2 , T 4 ( 0 ) = 0 , T 5 = 16 , T 6 ( 0 ) = 0 ,
which are exactly the values of Tangent numbers.
For r N , the higher-order tangent polynomials, T n r ( x ) ( n 0 ) , are defined by the following generating function (see [18])
2 e 2 t + 1 r e x t = n = 0 T n r ( x ) t n n ! , | 2 t | < π .
When r = 1 , the above equation gives the generating function for the tangent polynomials in (2) (see [16]).
The Tangent–Berrnoulli polynomials and Tangent–Genocchi polynomials of higher order, which are variations of the Tangent polynomials, are defined below.
Definition 1.
The Tangent–Bernoulli polynomials of higher order denoted by ( T B ) n r ( x ) , r 2 are defined by means of the following generating function
t e 2 t 1 r e x t = n = 0 ( T B ) n r ( x ) t n n ! , | t | < π .
Definition 2.
The Tangent–Genocchi polynomials of higher order denoted by ( T G ) n r ( x ) , r 2 are defined by means of the following generating function
2 t e 2 t + 1 r e x t = n = 0 ( T G ) n r ( x ) t n n ! , | t | < π .
In this paper, the researchers derived the Fourier series expansion of the Tangent polynomials, Tangent–Bernoulli and Tangent–Genocchi polynomials of order r, r Z + .

2. Methods

This study was facilitated by the use of Residue Theory; the Cauchy Residue Theorem [19] was applied to solve the contour integral representation of the polynomials.

3. Results and Discussions

In this section, we provide a detailed discussion of the process of deriving the Fourier series expansion of higher-order Tangent polynomials as well as the Fourier series expansions of higher-order Tangent–Bernoulli and Tangent–Genocchi polynomials using the method of Bayad [20].

3.1. Higher Order Tangent Polynomials

To derive the Fourier series expansion of Tangent polynomials of higher order, the Cauchy Integral Formula is applied to (3) to yield
T n r ( x ) n ! = 1 2 π i C 2 e 2 t + 1 r e x t d t t n + 1 ,
where C is a circle about zero with radius < π / 2 . Consider the function
f ( t ) = 2 e 2 t + 1 r e x t t n + 1 .
This function has a pole of order n r + 1 at t = 0 and poles of order r at t k where t k are zeros of e 2 t + 1 . The values of t k are obtained as follows:
e 2 t + 1 = 0 log e 2 t = log 1 2 t = log ( 1 ) = ln | 1 | + i A r g ( 1 ) + 2 k π i = π i + 2 k π i t k : = t = ( 2 k + 1 ) π 2 i , f o r k Z
Notice that the nearest t k are t 0 and t 1 . Hence, the contour in (6) must have a radius of less than | t 0 | = | t 1 | = π 2 . Now, let C N be the circle about 0 of radius ( 2 N + 1 ) π 2 ϵ , where 0 < ϵ < 1 . This range of ϵ will ensure that the circle C N is between the circles C N 1 and C N . Then, using Cauchy Residue Theorem [19], we have
1 2 π i C N f ( t ) d t = k < N R e s f ( t ) , t = t k + R e s f ( t ) , t = 0 .
Taking the limit as N ,
lim N 1 2 π i C N f ( t ) d t = k Z R e s f ( t ) , t = t k + R e s f ( t ) , t = 0 .
We prove the following lemma:
Lemma 1.
For 0 < x r , n 1 and fixed r Z + , as N ,
C N f ( t ) d t 0 ,
where C N is a circle about 0 with radius R = ( 2 N + 1 ) π 2 ϵ , 0 < ϵ < 1 .
Proof. 
Taking the modulus of the integral gives
C N 2 e 2 t + 1 r · e x t t n + 1 d t 2 r C N | e x t | | e 2 t + 1 | r · | d t | | t n + 1 | , w i t h t C N , | t | = ( 2 N + 1 ) π 2 ϵ .
Note that, with t = a + i b , a = R e ( t ) and b = I m ( t ) , we have
| e 2 t + 1 | = | e 2 ( a + i b ) + 1 | = | e 2 a e 2 i b + 1 | = | e 2 a ( cos 2 b + i sin 2 b ) + 1 | = | e 2 a cos 2 b + 1 + e 2 a i sin 2 b | = ( e 2 a cos 2 b + 1 ) 2 + ( e 2 a i sin 2 b ) 2 | e 2 t + 1 | r = e 2 a r 1 + 2 cos 2 b e 2 a + 1 e 4 a r .
Then,
| e x t | | e 2 t + 1 | r = e x a e 2 a r 1 + 2 cos 2 b e 2 a + 1 e 4 a r 2 = 1 e a ( 2 r x ) 1 + 2 cos 2 b e 2 a + 1 e 4 a r 2 .
Now, with 0 < x r , a ( 2 r x ) a ( 2 r r ) = a r . Hence,
1 e a ( 2 r x ) 1 e a r .
Thus,
| e x t | | e 2 t + 1 | r 1 e a r 1 1 + 2 cos 2 b e 2 a + 1 e 4 a r 2 = 1 e 2 a + 2 cos 2 b + e 2 a r 2 .
Let B = e 2 a + 2 cos 2 b + e 2 a r 2 . With t = R e i θ = R ( cos θ + i sin θ ) , a = [ ( 2 N + 1 ) π 2 ϵ ] cos θ , 0 θ π , we consider three cases:
  • Case 1. When cos θ < 0 , e 2 a 0 and e 2 a + as N + , which means B + and so | e x t | | e 2 t + 1 | r is bounded;
  • Case 2. When cos θ > 0 , e 2 a + and e 2 a 0 as N + , which means B + and so | e x t | | e 2 t + 1 | r is bounded;
  • Case 3. When cos θ = 0 , e 2 a = e 2 a = 1 , which means B = 2 + 2 cos 2 b . For | e x t | | e 2 t + 1 | r to be bounded, 2 + 2 cos 2 b 0 , where
    b = ( 2 N + 1 ) π 2 ϵ sin θ = ( 2 N + 1 ) π 2 ϵ ( ± 1 ) .
    Note that
    2 + 2 cos 2 b = 2 + 2 cos 2 ( ± 1 ) ( 2 N + 1 ) π 2 ϵ = 2 + 2 cos 2 ( 2 N + 1 ) π 2 ϵ .
    The last expression is zero if, and only if,
    cos 2 ( 2 N + 1 ) π 2 ϵ = 1
    2 ( 2 N + 1 ) π 2 ϵ = ( 2 M + 1 ) π , M Z ( 2 N + 1 ) π 2 ϵ = ( 2 M + 1 ) π 2 2 ( N M ) π 2 = ϵ ( N M ) π = ϵ ,
    which is impossible because 0 < ϵ < 1 . Hence, B 0 . That is, | e x t | | e 2 t + 1 | r is bounded.
Thus, in any case,
| e x t | | e 2 t + 1 | r K ,
for some constant K, t C N . Consequently,
C N 2 e 2 t + 1 r · e x t t n + 1 d t 2 r C N | e x t | | e 2 t + 1 | r · | d t | | t n + 1 | 2 r C N K | d t | | t n + 1 | = 2 r K 2 N + 1 π 2 ϵ n + 1 · 2 N + 1 π 2 ϵ ( 2 π ) < 2 r + 1 K π 2 N + 1 π 2 ϵ n 0 a s N f o r n 1 .
Using Lemma 1, Equation (9) becomes
0 = k Z R e s f ( t ) , t = t k + R e s f ( t ) , t = 0 0 = R e s f ( t ) , t = 0 + k Z R e s f ( t ) , t = t k 0 = T n r ( x ) n ! + k Z R e s f ( t ) , t = t k .
Thus, we have
T n r = n ! k Z R e s f ( t ) , t = t k .
The Fourier series for the tangent polynomials of higher order is given in the following theorem.
Theorem 1.
For 0 < x r ,
T n r ( x ) = 2 · n ! 2 π r + n k = 0 j = 0 r 1 ( 1 ) j r + n j 1 r j 1 π j j ! B j r x 2 × cos ( 2 k + 1 ) π x / 2 ( r + n j ) π / 2 ( 2 k + 1 ) r + n j ,
where B j r x 2 denotes the Bernoulli polynomials of order r defined by
w e w 1 r e x w = j = 0 B j r ( x ) w n n ! .
Proof. 
For r 2 ,
R e s f ( t ) , t = t k = 1 ( r 1 ) ! lim t t k d r 1 d t r 1 ( t t k ) r 2 e 2 t + 1 r e x t t n + 1 .
Since e 2 t k = e 2 ( 2 k + 1 ) π 2 i = e ( 2 k + 1 ) π i = 1 , we can write e 2 t + 1 r as
e 2 t + 1 r = ( 1 ) r e 2 t ( 1 ) 1 r = ( 1 ) r e 2 t · e 2 t k 1 r = ( 1 ) r e 2 ( t t k ) 1 r .
Then, we have
( t t k ) r 2 e 2 t + 1 r e x t t n + 1 = 2 r ( t t k ) r ( 1 ) r e 2 ( t t k ) 1 r · e x t t n + 1 = ( 1 ) r 2 ( t t k ) r ( e 2 ( t t k ) 1 ) r · e x t t n + 1 = ( 1 ) r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n + 1 ) ,
where B n r denotes the Bernoulli numbers of order r defined in the theorem.
w e w 1 r = n = 0 B n r w n n ! .
Applying the Leibniz Rule on differentiation,
d r 1 d t r 1 ( t t k ) r 2 e 2 t + 1 r e x t t n + 1 = d r 1 d t r 1 ( 1 ) r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n + 1 ) = ( 1 ) r d r 1 d t r 1 e x t n = 0 B n r 2 ( t t k ) n n ! t ( n + 1 ) = ( 1 ) r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n + 1 ) · d j d t j e x t n = 0 B n r 2 ( t t k ) n n ! ,
d j d t j e x t n = 0 B n r 2 n ( t t k ) n ) n ! = l = 0 j j l x j l e x t n = l B n r 2 n n ! ( n ) l ( t t k ) n l = e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! ,
d r 1 d t r 1 ( t t k ) r 2 e 2 t + 1 r e x t t n + 1 = ( 1 ) r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n + 1 ) × e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! .
Thus,
R e s f ( t ) , t = t k = 1 ( r 1 ) ! lim t t k ( 1 ) r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n + 1 ) × lim t t k e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! .
Note that B n r ( t t k ) n l ( n l ) ! 0 as t t k except when n = l . Using the notation ( n ) l defined by
( n ) l = n ( n 1 ) ( n l + 1 ) , n l 1 , l = 0 0 , n < l ,
this gives
R e s f ( t ) , t = t k = 1 ( r 1 ) ! ( 1 ) r j = 0 r 1 r 1 j ( 1 ) r 1 j ( n + r 1 j ) r 1 j t k ( n + r j ) × e x t k l = 0 j j l x j l 2 l B l r = ( 1 ) r ( r 1 ) ! j = 0 r 1 ( r 1 ) ! j ! ( r 1 j ) ! ( 1 ) r 1 j ( n + r 1 j ) r 1 j t k ( n + r j ) × e x t k l = 0 j j l x j l 2 l B l r = j = 0 r 1 ( 1 ) j 1 n + r 1 j r 1 j t k j n r j ! e x t k 2 j l = 0 j j l x j l 2 j l B l r .
Recall from [21] that B j r x 2 = l = 0 j j l B l r x 2 j l . Thus,
R e s f ( t ) , t = t k = j = 0 r 1 ( 1 ) j 1 2 j n + r 1 j r 1 j t k j n r j ! e x t k B j r x 2 = j = 0 r 1 ( 1 ) j 1 2 j n + r 1 j r 1 j B j r x 2 j ! e x t k t k n + r j = j = 0 r 1 ( 1 ) j 1 2 j r + n j 1 r j 1 B j r x 2 j ! e x t k t k r + n j .
With t k = 1 2 ( 2 k + 1 ) π i , we get
R e s f ( t ) , t = t k = j = 0 r 1 ( 1 ) j 1 2 j r + n j 1 r j 1 B j r x 2 j ! e 1 2 ( 2 k + 1 ) π i x 1 2 ( 2 k + 1 ) π i r + n j
= 1 1 2 π i r + n j = 0 r 1 ( 1 ) j 1 2 j r + n j 1 r j 1 1 2 π i j j ! B j r x 2 e 1 2 ( 2 k + 1 ) π i x ( 2 k + 1 ) r + n j .
This gives
T n r ( x ) = n ! 2 π i r + n k Z j = 0 r 1 ( 1 ) j r + n j 1 r j 1 π i j j ! B j r x 2 e 1 2 ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) r + n j .
Now, from (12), we look at
i ( r + n j ) k Z e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) r + n j .
Noting that i ( r + n j ) = e ( r + n j ) π i / 2 and ( 1 ) r + n j = e ( r + n j ) π i , we see that
i ( r + n j ) k Z e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) r + n j
= i ( r + n j ) k = 0 e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) r + n j + ( 1 ) r + n j k = 0 e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) r + n j = k = 0 e ( 2 k + 1 ) x / 2 ( r + n j ) / 2 π i + e ( 2 k + 1 ) x / 2 ( r + n j ) / 2 π i ( 2 k + 1 ) r + n j = k = 0 2 cos ( 2 k + 1 ) π x / 2 ( r + n j ) π / 2 ( 2 k + 1 ) r + n j = 2 k = 0 cos ( 2 k + 1 ) π x / 2 ( r + n j ) π / 2 ( 2 k + 1 ) r + n j .
Replacing (13) with (14) and substituting to (12), we get the desired formula (11). □

3.2. Higher-Order Tangent–Bernoulli Polynomials

The Fourier series of higher-order Tangent–Bernoulli polynomials is derived similarly to that in Section 3.1. Applying the Cauchy Integral Formula [19] to (4) yields
( T B ) n r ( x ) n ! = 1 2 π i C t e 2 t 1 r e x t d t t n + 1 ,
where C is a circle about zero with radius < π . Consider the function
g ( t ) = t e 2 t 1 r e x t t n + 1 .
This function has a pole of order n + 1 at t = 0 and poles of order r at t k , where t k are zeros of e 2 t + 1 . The values of t k are obtained as follows:
e 2 t 1 = 0 log e 2 t = log 1 2 t = log ( 1 ) = ln | 1 | + i A r g ( 1 ) + 2 k π i = 2 k π i t k : = t = k π i , f o r k Z , k 0 .
Note that the poles nearest to zero are ± π . Hence, the circle C must have a radius < π .
Now, let C N be the circle about 0 of radius N π ϵ , where 0 < ϵ < 1 . This range of ϵ will ensure that the circle C N is between the circles C N 1 and C N . Then, using Cauchy Residue Theorem [19], we have
1 2 π i C N g ( t ) d t = k < N R e s g ( t ) , t = t k + R e s g ( t ) , t = 0 .
Taking the limit as N ,
lim N 1 2 π i C N g ( t ) d t = k Z R e s g ( t ) , t = t k .
We prove the following lemma.
Lemma 2.
For 0 < x r and n r , as N ,
C N g ( t ) d t 0 ,
where C N is a circle about 0 with radius R = N π ϵ , 0 < ϵ < 1 .
Proof. 
Taking the modulus of the integral gives
C N t e 2 t 1 r · e x t t n + 1 d t C N | e x t | | e 2 t 1 | r · | d t | | t n r + 1 | , w i t h t C N , | t | = N π ϵ .
Note that, with t = a + i b , a = R e ( t ) and b = I m ( t ) , we have
| e 2 t 1 | = | e 2 ( a + i b ) + 1 | = | e 2 a e 2 i b 1 | = | e 2 a ( cos 2 b + i sin 2 b ) 1 | = | e 2 a cos 2 b 1 + e 2 a i sin 2 b | = ( e 2 a cos 2 b 1 ) 2 + ( e 2 a i sin 2 b ) 2 | e 2 t 1 | r = e 2 a r 1 2 cos 2 b e 2 a + 1 e 4 a r .
Then,
| e x t | | e 2 t 1 | r = e x a e 2 a r 1 + 2 cos 2 b e 2 a + 1 e 4 a r 2 = 1 e a ( 2 r x ) 1 2 cos 2 b e 2 a + 1 e 4 a r 2 .
Now, with 0 < x r , a ( 2 r x ) a ( 2 r r ) = a r . Hence,
1 e a ( 2 r x ) 1 e a r .
Thus,
| e x t | | e 2 t 1 | r 1 e a r 1 1 2 cos 2 b e 2 a + 1 e 4 a r 2 = 1 e 2 a 2 cos 2 b + e 2 a r 2 .
Let B = e 2 a 2 cos 2 b + e 2 a r 2 . With t = R e i θ = R ( cos θ + i sin θ ) , a = [ N π ϵ ] cos θ , 0 θ π , we consider three cases:
  • Case 1. When cos θ < 0 , e 2 a 0 and e 2 a + as N + , which means B + and so | e x t | | e 2 t + 1 | r is bounded;
  • Case 2. When cos θ > 0 , e 2 a + and e 2 a 0 as N + , which means B + and so | e x t | | e 2 t + 1 | r is bounded;
  • Case 3. When cos θ = 0 , e 2 a = e 2 a = 1 , which means B = 2 2 cos 2 b . For | e x t | | e 2 t 1 | r to be bounded, 2 2 cos 2 b 0 , where
    b = [ N π ϵ ] sin θ = [ N π ϵ ] ( ± 1 ) .
    Note that
    2 2 cos 2 b = 2 2 cos 2 ( ± 1 ) [ N π ϵ ] = 2 2 cos 2 [ N π ϵ ] .
    The last expression is zero if and only if
    cos 2 [ N π ϵ ] = 1
    2 [ N π ϵ ] = 2 M π , M Z N π ϵ = M π ( N M ) π = ϵ
    which is impossible because 0 < ϵ < 1 . Hence, B 0 . That is, | e x t | | e 2 t 1 | r is bounded.
Thus, in any case,
| e x t | | e 2 t 1 | r K 1 ,
for some constant K 1 , t C N . Consequently,
C N t e 2 t 1 r · e x t t n + 1 d t K 1 C N | d t | | t n r + 1 | = K 1 ( N π ϵ ) 2 π ( N π ϵ ) n r + 1 = 2 K 1 π ( N π ϵ ) n r 0 a s N f o r n r .
Applying Lemma 2, Equation (18) becomes
0 = k Z , k 0 R e s g ( t ) , t = t k + R e s g ( t ) , t = 0 0 = R e s g ( t ) , t = 0 + k Z , k 0 R e s g ( t ) , t = t k 0 = ( T B ) n r ( x ) n ! + k Z , k 0 R e s g ( t ) , t = t k .
Thus, we have
( T B ) n r = n ! k Z , k 0 R e s g ( t ) , t = t k .
The Fourier series for the Tangent–Bernoulli polynomials ( T B ) n r ( x ) of higher order is given in the following theorem.
Theorem 2.
For 0 < x r ,
( T B ) n r ( x ) = n ! π n k = 0 j = 0 r 1 ( 1 ) r j n j 1 r j 1 π j 2 r j 1 j ! B j r x 2 × cos k x ( n j ) / 2 π k n j ,
where B j r x 2 denotes the Bernoulli polynomials of order r defined by
w e w 1 r e x w = n = 0 B n r ( x ) w n n ! .
Proof. 
For r 2 ,
R e s g ( t ) , t = t k = 1 ( r 1 ) ! lim t t k d r 1 d t r 1 ( t t k ) r t e 2 t 1 r e x t t n + 1 .
Since e 2 t k = e 2 k π i = 1 , we can write e 2 t 1 r as
e 2 t 1 r = e 2 t ( 1 ) 1 r = e 2 t · e 2 t k 1 r = e 2 ( t t k ) 1 r .
Then, we have
( t t k ) r 1 e 2 t 1 r e x t t n r + 1 = ( t t k ) r e 2 ( t t k ) 1 r · e x t t n r + 1 = 1 2 r ( 2 ( t t k ) ) r ( e 2 ( t t k ) 1 ) r · e x t t n r + 1 = 1 2 r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n r + 1 ) ,
where B n r denotes the Bernoulli numbers of order r defined by the generating function
w e w 1 r = n = 0 B n r w n n ! .
Applying the Leibniz Rule,
d r 1 d t r 1 ( t t k ) r 1 e 2 t 1 r e x t t n r + 1 = d r 1 d t r 1 1 2 r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n r + 1 ) = 1 2 r d r 1 d t r 1 e x t n = 0 B n r 2 ( t t k ) n n ! t ( n r + 1 ) = 1 2 r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) · d j d t j e x t n = 0 B n r 2 ( t t k ) n n ! ,
d j d t j e x t n = 0 B n r 2 n ( t t k ) n ) n ! = l = 0 j j l x j l e x t n = l B n r 2 n n ! ( n ) l ( t t k ) n l = e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! ,
d r 1 d t r 1 ( t t k ) r 1 e 2 t 1 r e x t t n r + 1 = 1 2 r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) × e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! ,
R e s g ( t ) , t = t k = 1 ( r 1 ) ! lim t t k 1 2 r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) × lim t t k e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! .
Note that B n r ( t t k ) n l ( n l ) ! 0 as t t k except when n = l .
R e s g ( t ) , t = t k = 1 ( r 1 ) ! 1 2 r j = 0 r 1 r 1 j ( 1 ) r 1 j ( n r + r 1 j ) r 1 j t k ( n r + r j ) × e x t k l = 0 j j l x j l 2 l B l r = 1 2 r ( r 1 ) ! j = 0 r 1 ( r 1 ) ! j ! ( r 1 j ) ! ( 1 ) r 1 j ( n 1 j ) r 1 j t k ( n j ) × e x t k l = 0 j j l x j l 2 l B l r = 1 2 r j = 0 r 1 ( 1 ) r j + 1 n 1 j r 1 j t k j n j ! e x t k 2 j l = 0 j j l x j l 2 j l B l r .
We recall that B j r x 2 = l = 0 j j l B l r x 2 j l . Hence,
R e s g ( t ) , t = t k = 1 2 r j = 0 r 1 ( 1 ) r j + 1 2 j n 1 j r 1 j t k j n j ! e x t k B j r x 2 = 1 2 r j = 0 r 1 ( 1 ) r j + 1 2 j n 1 j r 1 j B j r x 2 j ! e x t k t k n j = 1 2 r j = 0 r 1 ( 1 ) r j + 1 2 j n j 1 r j 1 B j r x 2 j ! e x t k t k n j .
Taking t k = k π i , we get
R e s g ( t ) , t = t k = 1 2 r j = 0 r 1 ( 1 ) r j + 1 2 j n j 1 r j 1 B j r x 2 j ! e k π i x k π i n j .
= 1 2 r ( π i ) n j = 0 r 1 ( 1 ) r j + 1 2 j n j 1 r j 1 ( π i ) j j ! B j r x 2 e k π i x k n j .
Consequently,
( T B ) n r ( x ) = n ! 1 2 r ( π i ) n k Z , k 0 j = 0 r 1 ( 1 ) r j 2 j n j 1 r j 1 ( π i ) j j ! B j r x 2 e k π i x k n j .
Now, we look at
i ( n j ) k Z , k 0 e k π i x k n j .
Noting that i ( n j ) = e ( n j ) π i / 2 and ( 1 ) n j = e ( n j ) π i , we see that
i ( n j ) k Z , k 0 e k π i x k n j = i ( n j ) k = 1 e k π i x k n j + ( 1 ) n j k = 1 e k π i x k n j = k = 1 e k x ( n j ) / 2 π i + e k x ( n j ) / 2 π i k n j = 2 k = 1 cos k x ( n j ) / 2 π k n j .
Replacing (22) with (23) in (21), we get the desired formula (20). □

3.3. Higher-Order Tangent–Genocchi Polynomials

Here, Fourier series of the higher-order Tangent–Genocchi polynomials will similarly be derived.
Applying the Cauchy Integral Formula [19] to (5) yields
( T G ) n r ( x ) n ! = 1 2 π i C 2 t e 2 t + 1 r e x t d t t n + 1 ,
where C is a circle about zero with the radius < π / 2 . Consider the function
h ( t ) = 2 t e 2 t + 1 r e x t t n + 1 .
This function has a pole of order n + 1 at t = 0 and poles of order r at t k where t k are zeros of e 2 t + 1 . More precisely,
t k : = t = ( 2 k + 1 ) π 2 i , f o r k Z .
Now, let C N be the circle about 0 of radius ( 2 N + 1 ) π 2 ϵ , where 0 < ϵ < 1 . This is the same C N used in Section 3.1. By the Cauchy Residue Theorem [19],
1 2 π i C N h ( t ) d t = k < N R e s h ( t ) , t = t k + R e s h ( t ) , t = 0 .
Taking the limit as N ,
lim N 1 2 π i C N h ( t ) d t = k Z R e s h ( t ) , t = t k .
We prove the following lemma:
Lemma 3.
For 0 < x r , n > r and r > 1 , as N ,
C N h ( t ) d t 0 .
where C N is a circle about 0 with radius R = ( 2 N + 1 ) π 2 ϵ , 0 < ϵ < 1 .
Proof. 
Taking the modulus of the integral gives
C N 2 t e 2 t + 1 r · e x t t n + 1 d t 2 r C N | e x t | | e 2 t + 1 | r · | d t | | t n r + 1 | , w i t h t C N , | t | = ( 2 N + 1 ) π 2 ϵ .
As shown in the proof of Lemma 1,
| e x t | | e 2 t + 1 | r K
for some constant K, t C N . Consequently,
C N 2 e 2 t + 1 r · e x t t n + 1 d t 2 r C N | e x t | | e 2 t + 1 | r · | d t | | t n r + 1 | 2 r C N K | d t | | t n r + 1 | = 2 r K 2 N + 1 π 2 ϵ n r + 1 · 2 N + 1 π 2 ϵ ( 2 π ) < 2 r + 1 K π 2 N + 1 π 2 ϵ n r 0 a s N f o r n r .
Applying Lemma 3, Equation (27) becomes
0 = k Z R e s h ( t ) , t = t k + R e s h ( t ) , t = 0 0 = R e s h ( t ) , t = 0 + k Z R e s h ( t ) , t = t k 0 = ( T G ) n r ( x ) n ! + k Z R e s h ( t ) , t = t k .
Thus, we have
( T G ) n r = n ! k Z R e s h ( t ) , t = t k .
The desired Fourier series for the tangent–Genocchi polynomials of higher order is given in the following theorem.
Theorem 3.
For 0 < x r , n > r and r > 1 ,
( T G ) n r ( x ) = n ! 2 n 1 k = 0 j = 0 r 1 ( 1 ) j n j 1 r j 1 1 π n j j ! B j r x 2 × cos ( 2 k + 1 ) π x / 2 ( n j ) π / 2 ( 2 k + 1 ) n j ,
where B j r x 2 denotes the Bernoulli polynomials of order r defined by
w e w 1 r e x w = n = 0 B n r ( x ) w n n ! .
Proof. 
For r 2 ,
R e s h ( t ) , t = t k = 1 ( r 1 ) ! lim t t k d r 1 d t r 1 ( t t k ) r 2 t e 2 t + 1 r e x t t n + 1 .
Since e 2 t k = e 2 ( 2 k + 1 ) π 2 i = 1 , we can write e 2 t + 1 r as
e 2 t + 1 r = ( 1 ) r e 2 t ( 1 ) 1 r = ( 1 ) r e 2 t · e 2 t k 1 r = ( 1 ) r e 2 ( t t k ) 1 r .
Then, we have
( t t k ) r 2 t e 2 t + 1 r e x t t n + 1 = ( 2 ( t t k ) ) r ( 1 ) r e 2 ( t t k ) 1 r · e x t t n r + 1 = ( 1 ) r ( 2 ( t t k ) ) r ( e 2 ( t t k ) 1 ) r · e x t t n r + 1 = ( 1 ) r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n r + 1 ) ,
where B n r denotes the Bernoulli numbers of order r defined by the generating function
w e w 1 r = n = 0 B n r w n n ! .
Applying the Leibniz Rule for differentiation,
d r 1 d t r 1 ( t t k ) r 2 t e 2 t + 1 r e x t t n + 1 = d r 1 d t r 1 ( 1 ) r n = 0 B n r 2 ( t t k ) n n ! e x t t ( n r + 1 ) = ( 1 ) r d r 1 d t r 1 e x t n = 0 B n r 2 ( t t k ) n n ! t ( n r + 1 ) = ( 1 ) r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) · d j d t j e x t n = 0 B n r 2 ( t t k ) n n ! ,
d j d t j e x t n = 0 B n r 2 n ( t t k ) n ) n ! = l = 0 j j l x j l e x t n = l B n r 2 n n ! ( n ) l ( t t k ) n l = e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! ,
d r 1 d t r 1 ( t t k ) r 2 t e 2 t + 1 r e x t t n + 1 = ( 1 ) r j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) × e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! ,
R e s h ( t ) , t = t k = ( 1 ) r ( r 1 ) ! lim t t k j = 0 r 1 r 1 j d r 1 j d t r 1 j t ( n r + 1 ) × lim t t k e x t l = 0 j j l x j l n = l 2 n B n r ( t t k ) n l ( n l ) ! .
Note that B n r ( t t k ) n l ( n l ) ! 0 as t t k except when n = l . Thus,
R e s g ( t ) , t = t k = ( 1 ) r ( r 1 ) ! j = 0 r 1 r 1 j ( 1 ) r 1 j ( n r + r 1 j ) r 1 j t k ( n r + r j ) × e x t k l = 0 j j l x j l 2 l B l r = ( 1 ) r ( r 1 ) ! j = 0 r 1 ( r 1 ) ! j ! ( r 1 j ) ! ( 1 ) r 1 j ( n 1 j ) r 1 j t k ( n j ) × e x t k l = 0 j j l x j l 2 l B l r = j = 0 r 1 ( 1 ) j 1 n 1 j r 1 j t k j n j ! e x t k 2 j l = 0 j j l x j l 2 j l B l r .
We recall that B j r x 2 = l = 0 j j l B l r x 2 j l . Hence,
R e s h ( t ) , t = t k = j = 0 r 1 ( 1 ) j 1 2 j n 1 j r 1 j t k j n j ! e x t k B j r x 2 = j = 0 r 1 ( 1 ) j 1 2 j n 1 j r 1 j B j r x 2 j ! e x t k t k n j = j = 0 r 1 ( 1 ) j 1 2 j n j 1 r j 1 B j r x 2 j ! e x t k t k n j .
Taking t k = ( 2 k + 1 ) π 2 i , we get
R e s h ( t ) , t = t k = j = 0 r 1 ( 1 ) j 1 2 j n j 1 r j 1 B j r x 2 j ! e ( 2 k + 1 ) π 2 i x ( 2 k + 1 ) π 2 i n j
= 2 π i n j = 0 r 1 ( 1 ) j 1 n j 1 r j 1 ( π i ) j j ! B j r x 2 e ( 2 k + 1 ) π 2 i x 2 k + 1 n j .
Hence,
( T G ) n r ( x ) n ! = 2 π n j = 0 r 1 ( 1 ) j 1 n j 1 r j 1 π j j ! B j r x 2 i ( n j ) k Z e ( 2 k + 1 ) π 2 . i x 2 k + 1 n j .
Now, we look at
i ( n j ) k Z e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) n j .
Noting that i ( n j ) = e ( n j ) π i / 2 and ( 1 ) n j = e ( n j ) π i , we have
i ( n j ) k Z e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) n j = i ( n j ) k = 0 e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) n j + ( 1 ) n j k = 0 e ( 2 k + 1 ) π i 2 x ( 2 k + 1 ) n j = k = 0 e ( 2 k + 1 ) x / 2 ( n j ) / 2 π i + e ( 2 k + 1 ) x / 2 ( n j ) / 2 π i ( 2 k + 1 ) n j = k = 0 2 cos ( 2 k + 1 ) π x / 2 ( n j ) π / 2 ( 2 k + 1 ) n j = 2 k = 0 cos ( 2 k + 1 ) π x / 2 ( n j ) π / 2 ( 2 k + 1 ) n j .
Replacing (31) with (32) in (30), we get the desired formula (29). □

4. Conclusions and Recommendation

The Fourier series expansion of the Tangent polynomials of higher order was derived using the Cauchy Residue Theorem. Moreover, two variations of higher-order tangent polynomials were defined, namely, Tangent–Bernoulli and Tangent–Genocchi polynomials of higher order, and their Fourier series expansions were derived using Cauchy Residue theorem. For future research work, the authors recommend to derive the Fourier series expansion of Apostol–Frobenius Tangent, Apostol–Frobenius Tangent–Bernoulli and Apostol–Frobenius Tangent–Genocchi polynomials of higher order using the Cauchy Residue Theorem.

Author Contributions

Conceptualization, C.B.C. and R.B.C.; Formal analysis, C.B.C. and R.B.C.; Funding acquisition, C.B.C. and R.B.C.; Investigation, C.B.C. and R.B.C.; Methodology, C.B.C. and R.B.C.; Supervision, C.B.C. and R.B.C.; Writing—original draft, R.B.C.; Writing—review and editing, C.B.C. and R.B.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Cebu Normal University through its Research Institute for Computational Mathematics and Physics.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The articles used to support the findings of this study are available from the corresponding author upon request.

Acknowledgments

The authors would like to thank the Research Institute for Computational Mathematics and Physics of Cebu Normal University for funding this research.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Abramowitz, M.; Stegun, I.A.; Miller, D.A.B. Handbook of Mathematical Functions With Formulas, Graphs and Mathematical Tables (National Bureau of Standards Applied Mathematics Series No. 55). J. Appl. Mech. 1965, 32, 239. [Google Scholar] [CrossRef] [Green Version]
  2. Araci, S.; Acikgoz, M.; Sen, E. Some New Formulae for Genocchi Numbers and polynomials involving Bernoulli and Euler polynomials. Int. J. Math. Math. Sci. 2014, 2014, 760613. [Google Scholar] [CrossRef] [Green Version]
  3. Kim, T. Euler Numbers and Polynomials Associated with Zeta Functions. Abstr. Appl. Anal. 2008, 2008, 581582. [Google Scholar] [CrossRef]
  4. Kim, T. Note on the Euler numbers and polynomials. Adv. Stud. Contemp. Math. 2008, 17, 109–116. [Google Scholar]
  5. Rademacher, H. Topics in Analytic Number Theory; Die Grundlehren der Mathematischen Wissenschaften; Springer: Berlin, Germany, 1973; Volume 169. [Google Scholar]
  6. Simsek, Y. On twisted generalized euler numbers. Bull. Korean Math. Soc. 2004, 41, 299–306. [Google Scholar] [CrossRef]
  7. Simsek, Y. New families of special numbers for computing negative order euler numbers and related numbers and polynomials. Appl. Anal. Discret. Math. 2018, 12, 1–35. [Google Scholar] [CrossRef]
  8. Arakawa, T.; Kaneko, M. On Poly-Bernoulli Numbers. Comment. Math. Univ. St. Pauli 1999, 48, 159–167. [Google Scholar]
  9. Bayad, A.; Hamahata, Y. Polylogarithms and poly-Bernoulli polynomials. Kyushu J. Math. 2011, 65, 15–24. [Google Scholar] [CrossRef] [Green Version]
  10. Kaneko, M. Poly-Bernoulli Numbers. J. Théorie Nombres Bordx. 1997, 9, 221–228. [Google Scholar] [CrossRef]
  11. Ryoo, C.S. On the twisted q-tangent numbers and polynomials. Appl. Math. Sci. 2013, 7, 4935–4941. [Google Scholar] [CrossRef]
  12. Ryoo, C. A numerical investigation on the zeros of the tangent polynomials. J. Appl. Math. Inform. 2014, 32, 315–322. [Google Scholar] [CrossRef]
  13. Ryoo, C.S. Explicit identities for the generalized tangent polynomials. Nonlinear Anal. Differ. Equ. 2018, 6, 43–51. [Google Scholar] [CrossRef] [Green Version]
  14. Ryoo, C.S. A note on the symmetric properties for the tangent polynomials. Int. J. Math. Anal. 2013, 7, 2575–2581. [Google Scholar] [CrossRef]
  15. Corcino, R.; Corcino, C. Higher Order Apostol-Frobenius-Type Poly-Genocchi Polynomials With Parameters a,b and c. J. Inequal. Spec. Funct. 2021, 12, 54–72. [Google Scholar]
  16. Ryoo, C.S. A note on the tangent numbers and polynomials. Adv. Stud. Theor. Phys. 2013, 7, 447–454. [Google Scholar] [CrossRef]
  17. Ryoo, C.S. On the analogues of tangent numbers and polynomials associated with p-adic integral on Zp. Appl. Math. Sci. 2013, 7, 3177–3183. [Google Scholar] [CrossRef] [Green Version]
  18. Ryoo, C.S. Multiple tangent zeta function and tangent polynomials of higher order. Adv. Stud. Theor. Phys. 2014, 8, 457–462. [Google Scholar] [CrossRef]
  19. Churchill, R.V.; Brown, J.W. Complex Variable and Applications, 8th ed.; McGraw- Hill: New York, NY, USA, 2008. [Google Scholar]
  20. Bayad, A. Fourier Expansions for Apostol-Bernoulli, Apostol-Euler and Apostol-Genocchi Polynomials. Math. Comp. 2011, 80, 2219–2221. [Google Scholar] [CrossRef] [Green Version]
  21. Kim, D.S.; Kim, T. A note on higher-order Bernoulli polynomials. J. Inequal. Appl. 2013, 2013, 111. [Google Scholar] [CrossRef] [Green Version]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Corcino, C.B.; Corcino, R.B. Fourier Series for the Tangent Polynomials, Tangent–Bernoulli and Tangent–Genocchi Polynomials of Higher Order. Axioms 2022, 11, 86. https://doi.org/10.3390/axioms11030086

AMA Style

Corcino CB, Corcino RB. Fourier Series for the Tangent Polynomials, Tangent–Bernoulli and Tangent–Genocchi Polynomials of Higher Order. Axioms. 2022; 11(3):86. https://doi.org/10.3390/axioms11030086

Chicago/Turabian Style

Corcino, Cristina Bordaje, and Roberto Bagsarsa Corcino. 2022. "Fourier Series for the Tangent Polynomials, Tangent–Bernoulli and Tangent–Genocchi Polynomials of Higher Order" Axioms 11, no. 3: 86. https://doi.org/10.3390/axioms11030086

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