Professional Documents
Culture Documents
net/publication/361316080
Puiseux and Taylor Series of the Einstein Functions and Their Role in the
Solution of Inhomogeneous Airy's Equation
CITATIONS
2 authors, including:
M. H. Hamdan
University of New Brunswick, Saint John, New Brunswick, Canada
175 PUBLICATIONS 966 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Project # 4 PARALLEL FLOW OF A PRESSURE-DEPENDENT VISCOSITY FLUID DOWN INCLINED COMPOSITE POROUS LAYERS View project
All content following this page was uploaded by M. H. Hamdan on 15 June 2022.
Puiseux and Taylor Series of the Einstein Functions and Their Role in
the Solution of Inhomogeneous Airy’s Equation
M.H. HAMDAN
Department of Mathematics and Statistics
University of New Brunswick
100 Tucker Park Road, Saint John, N.B., E2L 4L5
CANADA
D.C. ROACH
Department of Engineering
University of New Brunswick
100 Tucker Park Road, Saint John, N.B., E2L 4L5
CANADA
Abstract: - The Einstein functions in generalized Puiseux and Taylor series are used as forcing functions in Airy’s
inhomogeneous equation, and particular and general solutions are obtained. Comparison are made with solutions
obtained using the Nield-Kuznetsov functions’ approach. For each of the Einstein’s functions, the standard Nield-
Kuznetsov function of the second kind is expressed in terms of Bessel functions. Computations and graphs in
this work were produced using Wolfram Alpha.
Received: August 9, 2021. Revised: May 3, 2022. Accepted: May 21, 2022. Published: June 15, 2022.
solutions to Airy’s inhomogeneous ordinary value problems. In the process of this work, the
differential equation (ODE) of the form existing relationships between Einstein’s functions
and the standard Nield-Kuznetsov functions,
𝑦 ′′ − 𝑥𝑦 = 𝐸𝑗(𝑥) (2) [1,9,10], will be utilized to express the standard
Nield-Kuznetsov function of the second kind in terms
wherein “prime” notation denotes ordinary of Bessel functions, [11].
differentiation with respect to the independent
variable, and 𝐸𝑗(𝑥) is one of the four functions in (1).
The particular solution to (2) is given by, [1]: 2 Einstein and Bessel Functions
Roach and Hamdan, [1], obtained particular solution
𝑥 𝜏
(3) (3) through the following alternate form, introduced
𝑦𝑝 = ∫ {∫ 𝐸𝑗(𝑡)𝑑𝑡} 𝑑𝜏
0 0 by Hamdan and Kamel, (2011):
and the general solution is given by
𝑦𝑝 = 𝜋𝐾𝑖(𝑥) − 𝜋𝑓(𝑥)𝑁𝑖(𝑥) (7)
𝑥 𝜏
𝑦 = 𝑐1 𝐴𝑖(𝑥) + 𝑐2 𝐵𝑖(𝑥) + ∫ {∫ 𝐸𝑗(𝑡)𝑑𝑡} 𝑑𝜏 (4) where 𝑁𝑖(𝑥) and 𝐾𝑖(𝑥) are the standard Nield-
0 0
Kuznetsov functions of the first and second kind,
where 𝑐1 and 𝑐2 are arbitrary constants, and respectively, defined as, (Nield and Kuznetsov, 2009,
𝐴𝑖(𝑥) and 𝐵𝑖(𝑥) are the linearly independent Airy’s Hamdan and Kamel, [9]:
functions of the first and second kind, respectively,
[6], with a non-zero Wronskian given by, [2,6]: 𝑥 𝑥
𝑁𝑖(𝑥) = 𝐴𝑖(𝑥) ∫0 𝐵𝑖(𝑡)𝑑𝑡 − 𝐵𝑖(𝑥) ∫0 𝐴𝑖(𝑡)𝑑𝑡 (8)
𝑑𝐵𝑖 (𝑥) 𝑑𝐴𝑖 (𝑥)
𝑊(𝐴𝑖 (𝑥), 𝐵𝑖 (𝑥)) = 𝐴𝑖 (𝑥) − 𝐵𝑖 (𝑥) 𝐾𝑖(𝑥) = 𝑓(𝑥)𝑁𝑖(𝑥) −
𝑑𝑥 𝑑𝑥
𝑥 𝑥
1 {𝐴𝑖(𝑥) ∫0 𝑓(𝑡)𝐵𝑖(𝑡) 𝑑𝑡− 𝐵𝑖(𝑥) ∫0 𝑓(𝑡)𝐴𝑖(𝑡) 𝑑𝑡}(9)
= (5)
𝜋
where 𝑓(𝑥) = 𝐸𝑗(𝑥) in the work of Roach and
Roach and Hamdan, [1], obtained the following Hamdan, [1], and in the current work.
particular solutions corresponding to (1), The following relationships between Einstein
respectively, by evaluating (3): functions, 𝐸𝑖(𝑥), Airy’s functions, 𝐴𝑖(𝑥) and 𝐵𝑖(𝑥),
and the standard Nield-Kuznetsov functions 𝑁𝑖(𝑥)
𝜋2 𝑥 and 𝐾𝑖(𝑥) were established in a theorem introduced
𝑥𝐿𝑖2 𝑒−𝑥 + 2𝐿𝑖3 𝑒−𝑥 − − 2𝜁(3).
6 by Roach and Hamdan, [1]:
𝑥𝜋2
− (𝐿𝑖3 (𝑒−𝑥 ) + ) − 𝜁(3).
6
1 𝑥 𝜏
𝜋 3 𝐾𝑖(𝑥) = 𝐸𝑖(𝑥)𝑁𝑖(𝑥) + ∫ {∫ 𝐸𝑗(𝑡)𝑑𝑡} 𝑑𝜏 (10)
𝑦𝑝 = − {𝑥[𝐿𝑖2 (𝑒−𝑥 )] + 3𝐿𝑖3 (𝑒−𝑥 ) + 𝑥} (6) 𝜋 0 0
3
−6 𝜁(3). 𝐾𝑖(𝑥) =
𝑥 𝑥
−𝑥 𝑙𝑜𝑔(1 − 𝑒𝑥 ) − 4𝑥𝐿𝑖2 (𝑒𝑥 ) +
2
𝐸𝑗(𝑥) {𝐴𝑖(𝑥) ∫ 𝐵𝑖(𝑡)𝑑𝑡 − 𝐵𝑖(𝑥) ∫ 𝐴𝑖(𝑡)𝑑𝑡}
𝑥𝜋2 0 0
6𝐿𝑖3 (𝑒𝑥 ) − − 6 𝜁(3). 1 𝑥 𝑥
{ 3 + ∫ {∫ 𝐸𝑗(𝑡)𝑑𝑡} 𝑑𝑡 (11)
𝜋 0 0
where 𝜁(𝑥) is the zeta function and 𝜁(3) = Using (6) and (7), the following relationships
1.2020569, and 𝐿𝑖2 (𝑥) and 𝐿𝑖3 (𝑥) are involving the polylogarithm functions are developed:
polylogarithmic functions, [7,8].
In this work, an alternative method is offered in 𝑥 2
𝐾𝑖(𝑥) = 𝐸1(𝑥)𝑁𝑖(𝑥) + 𝐿𝑖2 𝑒 −𝑥 + 𝐿𝑖3 𝑒 −𝑥
which the series form of Einstein functions is used in 𝜋 𝜋
the evaluation of particular solution (3). The use of 𝜋𝑥 2
− − 𝜁(3) (12)
series and tis approach might offer an easier 6 𝜋
alternative to computing particular and general
solutions when dealing with initial and boundary
1 𝑥𝜋 𝜇
𝐾𝑖(𝑥) = 𝐸2(𝑥)𝑁𝑖(𝑥) − [ 𝐿𝑖3 (𝑒 −𝑥 ) + ]
𝜋 6 𝐼 1 (𝜇). ∫ 𝐼1 (𝑡)𝑑𝑡 −
−
1 2𝑥 √𝑥 3
0
3
− 𝜁(3) (13) 𝐾𝑖(𝑥) = 𝜇 +
𝜋 3√3(𝑒 𝑥 − 1)
𝐼1 (𝜇) ∫ 𝐼 1 (𝑡)𝑑𝑡
−
{ 3 3 }
0
𝐾𝑖(𝑥) = 𝐸3(𝑥)𝑁𝑖(𝑥) − 𝑥 2 𝜋𝑥 2
𝐿𝑖2 𝑒 −𝑥 + 𝐿𝑖3 𝑒 −𝑥 − − 𝜁(3) (21)
𝑥 3 𝜋2 6 𝜋 𝜋 6 𝜋
(𝑒 −𝑥 ) (𝑒 −𝑥 )
{ 𝐿𝑖2 + 𝐿𝑖3 + 𝑥} − 𝜁(3) (14)
𝜋 𝜋 3 π
𝑥2 𝐾𝑖(𝑥) =
𝐾𝑖(𝑥) = 𝐸4(𝑥)𝑁𝑖(𝑥) − 𝑙𝑜𝑔(1 − 𝑒 𝑥 )
𝜋 𝜇
4𝑥 6 𝑥𝜋 6
− 𝜋 𝐿𝑖2 (𝑒 𝑥 ) + 𝐿 (𝑒 𝑥 ) − 3 − π 𝜁(3)
𝜋 𝑖3
(15) 𝐼 1 (𝜇). ∫ 𝐼1 (𝑡)𝑑𝑡 −
−
2√𝑥 3
0
3
With the knowledge of the expressions of 𝐴𝑖(𝑥) 𝑙𝑜𝑔(1 − 𝑒 −𝑥 ) 𝜇 −
3√3
and 𝐵𝑖(𝑥), and their integrals in terms of Bessel’s
𝐼1 (𝜇) ∫ 𝐼 1 (𝑡)𝑑𝑡
function of the first kind as, [11]: 3
−
3
{ 0 }
1 𝑥𝜋 1
√𝑥 [ 𝐿𝑖3 (𝑒 −𝑥 ) + ] − 𝜁(3) (22)
𝐴𝑖(𝑥) = [𝐼 1 (𝜇) − 𝐼1 (𝜇)] (16) 𝜋 6 𝜋
3 −3 3
𝑥
𝐵𝑖(𝑥) = √ [𝐼 1 (𝜇) + 𝐼1 (𝜇)] (17)
3 −3 3 2𝑥√𝑥 2√𝑥
𝐾𝑖(𝑥) = [ − log(1 − 𝑒 −𝑥 )]
𝑥 𝜇 3√3(𝑒 𝑥 − 1) 3√3
1 𝜇 𝜇
∫ 𝐴𝑖(𝑡)𝑑𝑡 = ∫ [𝐼 1 (𝑡) − 𝐼1 (𝑡)] 𝑑𝑡 (18)
3 −
3 3 {𝐼 1 (𝜇). ∫ 𝐼1 (𝑡)𝑑𝑡 − 𝐼1 (𝜇) ∫ 𝐼 1 (𝑡)𝑑𝑡}
0 0 − −
3 3 3 3
0 0
𝑥 𝜇
1 𝑥 3 𝜋2
∫ 𝐵𝑖(𝑡)𝑑𝑡 = ∫ [𝐼 1 (𝑡) + 𝐼1 (𝑡)] 𝑑𝑡 (19) − { 𝐿𝑖2 (𝑒 −𝑥 ) + 𝐿𝑖3 (𝑒 −𝑥 ) + 𝑥}
− 𝜋 𝜋 3
√3 3 3
0 0 6
− 𝜁(3) (23)
2 3/2 π
wherein 𝜇 = 3
𝑥 ,then using (16)-(19) in (8), the
function 𝑁𝑖(𝑥) can be expressed in terms of Bessel’s
function, as, [11]:
𝜇
𝜇 2√𝑥𝑥 2 𝑒 𝑥
2√𝑥
𝐾𝑖(𝑥) = {𝐼−1 (𝜇). ∫ 𝐼1 (𝑡)𝑑𝑡
𝑁𝑖(𝑥) = [𝐼 1 (𝜇). ∫ 𝐼1 (𝑡)𝑑𝑡 − 3√3(𝑒 𝑥 − 1)2 3 3
0
3√3 − 𝜇
3 3
0
𝜇
− 𝐼1 (𝜇) ∫ 𝐼 1 (𝑡)𝑑𝑡}
−
3 3
𝐼1 (𝜇) ∫ 𝐼 1 (𝑡)𝑑𝑡] (20) 0
−
3 3
0 𝑥2
− 𝑙𝑜𝑔(1 − 𝑒 𝑥 )
𝜋
Using (16)-(19) in (12)-(15), the function 𝐾𝑖(𝑥)
in (12)-(15) can be expressed, respectively, in terms
of Bessel’s function, as
4𝑥 6 𝑥𝜋 6
− 𝐿𝑖2 (𝑒 𝑥 ) + 𝐿𝑖3 (𝑒 𝑥 ) − − 𝜁(3) (24)
𝜋 𝜋 3 π
𝑥
∫ log(1 − 𝑒 −𝑥 ) 𝑑𝑥 = 𝐿𝑖2 (𝑒 −𝑥 ) + 𝐶 (34) lim [ 𝑥 − log(1 − 𝑒 −𝑥 )] = 0 (38)
𝑥→∞ 𝑒 − 1
Domain of 𝐸3 (𝑥) is the set of positive real numbers and the general solution can be obtained by
and its range is the set of positive real numbers. Its substituting (42) in (4).
graph is shown in Fig. 3.
𝒙 𝟐 𝒆𝒙
Case 4: 𝑬𝟒 (𝒙) = (𝒆𝒙 −𝟏)𝟐
∞
𝑥 𝜋2 The following improper integral of 𝐸4 (𝑥) converges:
∫ [ 𝑥 − log(1 − 𝑒 −𝑥 )] 𝑑𝑥 = (37)
0 𝑒 −1 3 ∞
𝑥2𝑒 𝑥 𝜋2
∫ 𝑑𝑥 = (43)
and the function has a horizontal asymptote at 0, 0 (𝑒 𝑥 − 1)2 3
namely:
and the function has a horizontal asymptote at 0, their numerical values are different. Similar behavior
namely: is observed in Fig. 8(a) and 8(b). Although no
solution to initial or boundary value problems has
𝑥2𝑒 𝑥 been obtained in this work, hence solutions based on
lim =0 (44) the general solutions have not been computed, it is
𝑥→∞ (𝑒 𝑥 − 1)2
expected that numerical values of the general
solutions in both approaches should be close.
For particular solutions in Fig. 6(a) and 6(b), and
First derivative of 𝐸4 (𝑥) and its indefinite integral are in Fig. 8(a) and 8(b), differences occur in both the
given by: graphical trends and in the numerical values. This
might be attributed to the use of Puiseux series in
𝑥𝑒 𝑥 [𝑒 𝑥 (𝑥 − 2) + (𝑥 + 2)] these cases, or it might be possible that solutions to
𝐸′4 = (45) initial and boundary value problems based on the
(𝑒 𝑥 − 1)3
general solutions would adjust themselves
𝑥2𝑒 𝑥 numerically. While this is inconclusive at present,
∫ 𝑑𝑥 = 2𝐿𝑖2 (𝑒 𝑥 ) + 2𝑥𝑙𝑜𝑔(1 − 𝑒 𝑥 ) graphs of the particular solutions are meant to
(𝑒 𝑥 − 1)2 provide information with regard to how close the
𝑥2 polylogarithm expressions are to series expressions.
+ +𝐶 (46)
1 − 𝑒𝑥
𝑥2 𝑥4 𝑥6 𝑥8
𝐸4 (𝑥) = 1 − + − +
12 240 6048 172800
𝑥 10 691𝑥 12
− + (47)
5322240 118879488000
Using (47) in (3), the following particular solution is
obtained for (2) when its forcing function is 𝐸4 (𝑥):
𝑥2 𝑥4 𝑥6 𝑥8
𝑦𝑝 = − + − + Fig. 5(a)
2 144 7 200 338 688 𝒚𝒑 Corresponding to 𝑬𝟏 (𝒙) in Equation (6)
𝑥 10
−
15 552 000
𝑥 12 691𝑥 14
+ (48)
702 535 680 21 636 066 816 000
and the general solution can be obtained by
substituting (48) in (4).
Fig. 6(a)
𝒚𝒑 Corresponding to 𝑬𝟐 (𝒙) in Equation (6)
Fig. 8(a)
𝒚𝒑 Corresponding to 𝑬𝟏 (𝒙) in Equation (6)
5 Conclusion
In this work, a method of solution of the
inhomogeneous Airy’s equation when the right-
hand-side is one of Einstein’s functions is
investigated. The method is based on expressing the
Einstein functions in series form followed by
Fig. 7(a) obtaining particular solutions. The main conclusion
𝒚𝒑 Corresponding to 𝑬𝟏 (𝒙) in Equation (6) that can be drawn from this work is that Taylor series
expansions of Einstein’s functions produce particular
solutions with a trend that is similar to that of
particular solutions expressed in polylogarithmic
functions.
References: