Initial and Boundary Value Problems Involving the Inhomogeneous Weber Equation and the Nield-Kuznetsov Parametric Functions

The International Journal Of Engineering And Science (IJES)
|| Volume || 6 || Issue || 3 || Pages || PP 09-24|| 2017 ||
ISSN (e): 2319 – 1813 ISSN (p): 2319 – 1805
DOI : 10.9790/1813-0603010924 www.theijes.com Page 9
 Initial and Boundary Value Problems Involving the
Inhomogeneous Weber Equation and the
Nield-Kuznetsov Parametric Functions
S.M. Alzahrani1
, I. Gadoura2
, M.H. Hamdan3
1
Dept. of Mathematics and Statistics, University of New Brunswick, P.O. Box 5050, Saint John, New Brunswick,
CANADA E2L 4L5 On leave from University of Umm Al-Qura, Kingdom of Saudi Arabia
2
Dept. of Electrical and Computer Engineering, University of New Brunswick, P.O. Box 5050, Saint John, New
Brunswick, CANADA E2L 4L5
3
Dept. of Mathematics and Statistics, University of New Brunswick, P.O. Box 5050, Saint John, New Brunswick,
CANADA E2L 4L5
--------------------------------------------------------ABSTRACT-----------------------------------------------------------
Initial and boundary value problems of the inhomogeneous Weber differential equation are treated in this work.
General solutions are expressed in terms of the parametric Nield-Kuznetsov functions of the first and second
kinds, and are computed when the forcing function is a constant or a variable function of the independent
variable.
Keywords Weber inhomogeneous equation, parametric Nield-Kuznetsov functions.
------------------------------------------------------------------------------------------------------------------------------------
Date of Submission: 09 February 2017 Date of Accepted: 05 March 2017
--------------------------------------------------------------------------------------------------------------------------------------
I. INTRODUCTION
In a recent article, Alzahrani et al. [1] discussed how Weber’s differential equation can be used in modelling
flow through porous layers with variable permeability. In particular, they showed that the inhomogeneous Weber
ordinary differential equations (ODE) of the form
)()
4
(
2
yfua
y
u  ; a …(1)
wherein prime notation denotes ordinary differentiation with respect to the independent variable y, plays an
important role in the study of transition layer introduced by Nield and Kuznetsov, [2].
Equation (1) and its more general forms have received considerable attention in the literature due to its many
applications in mathematical physics, (cf. [3] and the references therein). The elegant discussion of equation (1)
by Temme, [3], illustrates that the homogeneous part of (1) possesses the parabolic cylindrical functions,
),( yaW  , as solutions, where y and a are real numbers. These solutions represent a linearly independent,
numerically satisfactory pair of solutions for all y , with a Wronskian given by
W( ),( xaW , ),( xaW  )=1. Computational techniques of the functions ),( yaW  have been provided by
many authors, and include the efficient algorithms developed by Temme, [4], and Gil et al., [5].
Computations of solution to (1) when the forcing function )( yf is a constant function of the independent
variable have been obtained by Abu Zaytoon et al. [6], who used computational procedures outlined in Gil et al.,
[5] to obtain solutions to initial and boundary value problems involving (1) with a constant forcing function f(y).
In the work of Alzahrani et al. [1], general solutions to the inhomogeneous equation (1) were expressed in terms
of the parametric Nield-Kuznetsov integral functions of the first and second kinds, valid for constant and
variable forcing functions, respectively. Series expressions for the said Nield-Kuznetsov functions were also
developed.
In the current work, we provide solutions to initial and boundary value problems involving equation (1) and
subject to the initial conditions:
)0(u and  )0(u …(2)
or subject to the boundary conditions:

Preparation of Papers for the International Journal of Engineering and Science..
11
)( bau  and 22
)( bau  …(3)
where  ,  , 1
a , 2
a , 1
b and 2
b are real numbers.
Solutions and computations are provided for the cases of )( yf being either a constant function or a variable
function of the independent variable, in order to study the effects of the forcing function on the behavior of the
solutions obtained.
II. SOLUTION PROCEDURE
General solution to equation (1) is given by, [1], [6]:
),(),(),( 21
yaKyaWcyaWcu www
 …(4)
where w
c1
, w
c 2
are arbitrary constants, ),( yaW and ),( yaW  are the Weber functions with parameter a,
[3], and ),( yaK w
is the parametric Nield-Kuznetsov function of the second-kind, defined by, [1]
dttaWtFyaWdttaWtFyaWyaK
yy
w  
00
),()(),(),()(),(),( …(5)
with first derivative given by
)(),()(),(),()(),(),(
00
yFdttaWtFyaWdttaWtFyaWyaK
yy
w

 …(6)
wherein )()( yfyF  .
When the forcing function is a constant, say )( yf , general solution (4) takes the form
),(),(),( 21
yaNyaWcyaWcu w
 …(7)
where ),( yaN w
is the parametric Nield-Kuznetsov function of the first-kind, defined by, [1]:
 
yy
w
dttaWyaWdttaWyaWyaN
00
),(),(),(),(),( …(8)
with first derivatives given by
 
yy
w
dttaWyaWdttaWyaWyaN
00
),(),(),(),(),( . …(9)
In order to obtain solutions to the initial and boundary value problems, general solutions (4) and (7) must satisfy
condition (2) for initial value problem, and condition (3) for boundary value problem. This requires
determination of the arbitrary constants appearing in (4) and (7). In what follows, the arbitrary constants are
determined for cases of constant and variable forcing functions.
In arriving at the expressions for the arbitrary constants, use is made of the following values at zero of the Nield-
Kuznetsov functions of the first and second kinds, obtained directly from equations (5), (6), (8) and (9), namely:
0)0,()0,(  aNaN ww
…(10)
0)0,( aK w
; )0()0,( FaK w
 …(11)
and values of the Weber functions )0,(aW and )0,(aW  , reported in the literature [3,4,5], as:
2/1
4/3
)
24
3
(
)
24
1
(
)2(
1
)0,(
ia
ia
aW


 …(12)

2/1
4/1
)
24
1
(
)
24
3
(
)2(
1
)0,(
ia
ia
aW


 …(13)
2/1
4/3
)4/3(
)4/1(
)2(
1
)0,0( 









W …(14)
2/1
4/1
)4/1(
)4/3(
)2(
1
)0,0( 









W . …(15)
III. SOLUTION TO THE INITIAL VALUE PROBLEM (IVP)
III.1. Constant Forcing Function:
When )( yf , where  is a given constant, the use of initial conditions (2) in the general solution (7)
yields the following values for the arbitrary constants w
c 1
, w
c 2
:
)0,()0,(2
)0,()0,(
1
aWaW
aWaW
c w




…(16)
)0,()0,(2
)0,()0,(
2
aWaW
aWaW
c w




. …(17)
III.2. Variable Forcing Function:
When )( yf is a variable function of y, using initial conditions (2) in gerenal solution (4) yields the following
values for the arbitrary constants w
c 1
, w
c 2
:
 
)0,()0,(2
)0,()0,()0,(
1
aWaW
aKaWaW
c w
w




…(18)
 
)0,()0,(2
)0,()0,()0,(
2
aWaW
aKaWaW
c w
w




. …(19)
IV. SOLUTION TO THE BOUNDARY VALUE PROBLEM (BVP)
IV.1. Constant Forcing Function:
When )( yf , using boundary conditions (3) in gerenal solution (7) yields the following values for the
arbitrary constants w
c 1
and w
c 2
:
 
),(),(),(),(
),(),(),(),(),(),(
1221
12211221
1
aaWaaWaaWaaW
aaWaaNaaWaaNaaWbaaWb
c ww
w



 …(20)
 
),(),(),(),(
),(),(),(),(),(),(
1221
12211221
2
aaWaaWaaWaaW
aaWaaNaaWaaNaaWbaaWb
c ww
w



 . …(21)
IV.2. Variable Forcing Function:
When )( yf is a variable function of y, using boundary conditions (3) in general solution (4) yields the
following values for the arbitrary constants w
c 1
and w
c 2
:

),(),(),(),(
),(),(),(),(),(),(
1221
12211221
1
aaWaaWaaWaaW
aaWaaKaaWaaKaaWbaaWb
c
ww
w


 …(22)
),(),(),(),(
),(),(),(),(),(),(
1221
12211221
2
aaWaaWaaWaaW
aaWaaKaaWaaKaaWbaaWb
c
ww
w


 . …(23)
V. COMPUTATIONS OF SOLUTIONS TO IVP AND BVP
V.1. Series Expressions for the Parametric Nield-Kuznetsov Functions
In order to determine values of the arbitrary constants, given by expressions (16)-(23), and to evaluate solutions
to IVP and BVP, we need to evaluate Weber functions and the parametric Nield-Kuznetsov functions at the
given values of the independent variable, y. These have been developed in [1] and in [6], as follows.
The following series expressions have been developed in [3,4,5], for the Weber functions ),( yaW and
),( yaW  :




 

0
12
0
2
)!12(
)()0,(
)!2(
)()0,(),(
n
n
n
n
n
n
n
y
aaW
n
y
aaWyaW  …(24)




 

0
12
0
2
)!12(
)()0,(
)!2(
)()0,(),(
n
n
n
n
n
n
n
y
aaW
n
y









1
2
1
12
)!2(
)()0,(
)!12(
)()0,(),(
n
n
n
n
n
n
n
y
aaW
n
y




 

1
12
1
2
)!12(
)()0,(
)!2(
)()0,(),(
n
n
n
n
n
n
n
y
aaW
n
y
aaWyaW  …(27)
nnn
nna  )12)(1(
2
1
12
  …(28)
)()32)(1(
2
1
)()( 12
annaaa nnn
   …(29)
1)(0
a ; aa )(1
 ; 1)(0
a ; aa )(1
 . …(30)
Using equations (24)-(30) in the definitions of ),( yaN w
and ),( yaK w
, given by (5) and (8), we obtain the
following series expressions:




















































0
22
0
2
0
12
0
12
)!22(
)(
)!2(
)(
)!12(
)(
)!12(
)(
)0,()0,(2),(
n
n
n
n
n
n
n
n
n
n
n
n
w
n
y
a
n
y
a
n
y
a
n
y
a
aWaWyaN


…(31)
 





y
n
n
n
n
n
nw
dt
n
t
tFa
n
y
aaWaWyaK
0 1
2
0
2
]
)!2(
)()([]
)!2(
)()[0,()0,(2),(  . …(32)
V.2. Numerical Experiment
Assume that it is required to solve (1) subject to initial conditions 1)0(,2)0(  uu or boundary conditions
.2)1(,1)0(  uu The forcing function is either constant,


1
)( yf and


1
)( yf , or
variable yyf )( ,
2
)( yyf  and yyf sin)(  . Then, expressions (16)-(23) are evaluated for the

arbitrary constants using expressions (24)-(32). Solutions to IVP, and to BVP, are then evaluated and plotted
over the interval 10  y for the various values of parameter a = 0, 1, 2, -1, as discussed in the next section.
V.3. Solution to the Initial Value Problem (IVP)
Values of the arbitrary constants, computed using (16)-(19) are given in Table 1 for the various values of
parameter a tested. These values are independent of the forcing function, and they have the same values for both
constant and variable forcing function in equation (1).
Table 1. Values of arbitrary constants in IVP with constant forcing function
for different values of parameter a .
All )( yf w
c1 w
c 2
0a 5120.04502460 92.00050673
1a 0.6356082490 92.09857042
2a 21.06562955 82.26617879
1a 0.6356082490 92.09857042
Solution curves for  /1)( yf are shown in Figure 1(a-f). Figure 1(a-d) provide a comparison
between the solution profiles of /1)( yf and /1)( yf , for each of the values of parameter a
tested, and show the relative positions of the solution curves. For positive values of a, the solution curves
corresponding to both forcing functions are exponential, and get closer to each other as a increases. This might
point to the conclusion that when a gets large enough one might expect the solution to (1), with constant forcing
function, to become independent of the forcing function.
When a = 0, Figure 1(a) shows an almost linear solution curve for /1)( yf , and a parabolic curve for
/1)( yf whose vertex does not fall within the interval 10  y . Figure 1(d) shows parabolic curves,
when a = -1, with vertices representing maximum points over the interval 10  y .
Comparison of the solution curves for different values of a is given in Figure 1(e), for /1)( yf , and in
Figure 1(f), for /1)( yf . The solution curve corresponding to a = 0 is an almost linear, increasing curve
for both constant forcing functions tested. Above this curve lie the exponentially increasing solution curves
corresponding to a > 0, and below which lies the parabolic curve that corresponds to a < 0.
Fig. 1(a). Solution to IVP for Constant Forcing Function  /1)( yf ; a=0.

Fig. 1(b). Solution to IVP for Constant Forcing Function  /1)( yf ; a=1.
Fig. 1(c). Solution to IVP for Constant Forcing Function  /1)( yf ; a=2.
Fig. 1(d). Solution to IVP for Constant Forcing Function  /1)( yf ; a=-1.

Fig. 1(e). Solution to IVP for Constant Forcing Function  /1)( yf and Different Values of a.
Fig. 1(f). Solution to IVP for Constant Forcing Function  /1)( yf and Different Values of a.
Solution curves for the IVP with variable forcing function are shown in Figures 2(a-g). Relative positions of the
curves for each value of a are illustrated in Figures 2(a-d), and their shapes depend on the value of a in much
the same way as for the case of constant forcing function.
Effects of changing a for each of the variable forcing functions yyf )( ,
2
)( yyf  and yyf sin)(  ,
are illustrated in Figures 2(e,f,g), respectively. Again, the solution curve corresponding to a = 0 is an almost
linear, increasing curve for all variable forcing functions tested. Above this curve lie the exponentially increasing
solution curves corresponding to a > 0, and below which lies the parabolic curve that corresponds to a < 0. This
pattern persists for all forcing functions tested.
Fig. 2(a). Solution to IVP for Variable Forcing Function )( yf ; a=0

Fig. 2(b). Solution to IVP for Variable Forcing Function )( yf ; a=1.
Fig. 2(c). Solution to IVP for Variable Forcing Function )( yf ; a=2.
Fig. 2(d). Solution to IVP for Variable Forcing Function )( yf ; a=-1.

Fig. 2(e). Solution to IVP for yyf )( and Different Values of a.
Fig. 2(f). Solution to IVP for
2
)( yyf  and Different Values of a.
Fig. 2(g). Solution to IVP for yyf sin)(  and Different Values of a.

V.4. Solution to the Boundary Value Problem (BVP)
Values of the arbitrary constants, computed using (20) and (21), are given in Table 2 for the various values of
parameter a and constant forcing functions.
Table 2. Values of arbitrary constants in BVP with constant forcing function
for different values of parameter a .


1
)( yf


1
)( yf
0a 7.40481466201
w
c
11.382555732
w
c
6.73173801801
w
c
61.709479082
w
c
1a 4.48746543101
w
c
5.87962390802
w
c
2.27148655901
w
c
11.095602782
w
c
2a 8.98051530701
w
c
2.68538886502
w
c
2.81546727701
w
c
0.85043689602
w
c
1a 8.48890875201
w
c
11.855998092
w
c
9.74455659601
w
c
52.111645932
w
c
Solution curves for  /1)( yf are shown in Figures 3(a-f). Figures 1(a-d) provide a comparison of
the solution profiles of /1)( yf and /1)( yf for each of the values of parameter a tested, and
show the relative positions of the solution curves. For positive values of a, the solution curves corresponding to
both forcing functions are parabolic with vertices representing minimum points. For negative a, solution curves
are parabolic with vertices representing maximum points. When a = 0, the solution curve corresponding to
/1)( yf is parabolic with a vertex representing a maximum point while the solution curve corresponding
to /1)( yf is an increasing, almost linear curve. In all cases, the curves corresponding
to /1)( yf are higher than the curves corresponding to /1)( yf .
Comparison of the solution curves for different values of a is given in Figure 3(e), for /1)( yf , and in
Figure 3(f), for /1)( yf . The solution curve corresponding to a = 0 is an almost linear, increasing curve
for both constant forcing functions tested. Above this curve lie the parabolic solution curves corresponding to a
< 0, and below which lies the parabolic curve that corresponds to a > 0.
Fig. 3(a). Solution to BVP for Constant Forcing Function  /1)( yf ; a=0.

Fig. 3(b). Solution to BVP for Constant Forcing Function  /1)( yf ; a=1.
Fig. 3(c). Solution to BVP for Constant Forcing Function  /1)( yf ; a=2.
Fig. 3(d). Solution to BVP for Constant Forcing Function  /1)( yf ; a=-1.

Fig. 3(e). Solution to BVP for Constant Forcing Function  /1)( yf and Different Values of a.
Fig. 3(f). Solution to BVP for Constant Forcing Function  /1)( yf and Different Values of a.
Values of the arbitrary constants, computed using (22) and (23), are given in Table 3 for the various values of
parameter a and variable forcing functions.
Table 3. Values of arbitrary constants in BVP with variable forcing function for different values of parameter a .
yyf )( 2
)( yyf  yyf sin)( 
0a 7.39669035001
w
c
91.374431412
w
c
0.48235535201
w
c
01.460096422
w
c
7.40484814601
w
c
51.382589212
w
c
1a 9.48917065001
w
c
9.87791868802
w
c
9.43350076501
w
c
3.93358857402
w
c
8.48377440201
w
c
5.88331493702
w
c
2a 4.97920837801
w
c
6.68669579402
w
c
7.93734016501
w
c
8.72856400702
w
c
7.97562781201
w
c
6.69027636002
w
c
1a 1.47789875601
w
c
51.844988092
w
c
5.54603431001
w
c
91.913123642
w
c
8.48493027801
w
c
71.852019612
w
c

Solution curves for the BVP with variable forcing function are shown in Figures 4(a-g). Relative positions of
the curves for each value of a are illustrated in Figures 4(a-d), and their shapes depend on the value of a in
much the same way as for the case of constant forcing function. For the three variable forcing functions tested
for 10  y , solution curves are very close to each other and assume similar parabolic shapes. This might be
indicative of lesser influence the forcing functions have on the solution, as compared to the influence of
parameter a, as illustrated in Figures 4(e,f,g), for each of the variable forcing functions yyf )( ,
2
)( yyf  and yyf sin)(  . For each of the variable forcing functions, the solution curve corresponding to
a = 0 is almost linear, increasing curve for all variable forcing functions tested. Above this curve lies the
parabolic solution curve corresponding to a < 0, and below which lie the parabolic curve that corresponds to a >
0. This pattern persists for all forcing functions tested.
Fig. 4(a). Solution to BVP for Variable Forcing Function )( yf ; a=0.
Fig. 4(b). Solution to BVP for Variable Forcing Function )( yf ; a=1.

Fig. 4(c). Solution to BVP for Variable Forcing Function )( yf ; a=2.
Fig. 4(d). Solution to BVP for Variable Forcing Function )( yf ; a=-1.
Fig. 4(e). Solution to BVP for yyf )( and Different Values of a.

Fig. 4(f). Solution to BVP for
2
)( yyf  and Different Values of a.
Fig. 4(g). Solution to BVP for yyf sin)(  and Different Values of a.
VI. CONCLUSION
In this work we considered initial value and boundary value problems associated with Weber’s inhomogeneous
ordinary differential equation with real parameter a. Both cases of constant and variable forcing functions have
been treated using five selected functions. Solutions have been expressed in terms of the parametric Nield-
Kuznetsov functions of the first- and second-kind, and evaluated using computational procedures based on
power series for the Weber and the Nield-Kuznetsov functions. Computations of the arbitrary constants and the
solutions are tabulated or graphed in this work, and support the following conclusions.
(a) Values of the arbitrary constants in the case of initial value problem are independent of the forcing function
in Weber’s equation. Rather, they depend on the real parameter a.
(b) Values of the arbitrary constants in the case of boundary value problem depend on both the forcing function
in Weber’s equation and on the real parameter a.
(c) For the initial value problem with constant or variable forcing functions, the solution curve corresponding to
a = 0 is an almost linear, increasing curve for all variable forcing functions tested. Above this curve lie the
exponentially increasing solution curves corresponding to a > 0, and below which lies the parabolic curve
that corresponds to a < 0.
(d) For the boundary value problem with constant or variable forcing functions, the solution curve
corresponding to a = 0 is an almost linear, increasing curve. Above this curve lies the parabolic solution
curve corresponding to a < 0, and below which lie the parabolic curve that corresponds to a > 0.

REFERENCES
[1]. S.M. Alzahrani, I. Gadoura, and M.H. Hamdan, Nield- Kuznetsov functions of the first- and second kind, IOSR Journal of Applied
Physics, 8(3), 2016, 47-56.
[2]. D.A. Nield and A.V. Kuznetsov, The effect of a transition layer between a fluid and a porous medium: shear flow in a channel,
Transport in Porous Media, 78, 2009, 477-487.
[3]. N. M. Temme, Parabolic cylinder functions, in Frank W. J. Olver, Daniel W. Lozier, Ronald F. Boisvert and Charles W. Clark (eds.),
NIST Handbook of mathematical functions, 12, (U.S. Dept. Commerce, Washington, DC, Cambridge University Press, 2010) 303-
319.
[4]. N. M. Temme, Numerical and asymptotic aspects of parabolic cylinder functions, J. Computational and Applied Mathematics, 121,
2000, 221-246.
[5]. A. Gil, J. Segura and N.M. Temme, Fast and accurate computation of the Weber parabolic cylinder function W(a, x), IMA Journal of
Numerical Analysis, 31, 2011, 1194−1216.
[6]. M.S. Abu Zaytoon, T.L. Alderson, and M.H. Hamdan, Weber’s inhomogeneous differential equation with initial and boundary
conditions, Int. J. Open Problems Compt. Math., 9(2), 2016, 1-11.
Biographies and Photographs
Mr. S.M. Alzahrani is a PhD candidate in Applied Mathematics, supervised by Drs. I. Gadoura and M.H.
Hamdan at the University of New Brunswick, Saint John, N.B., Canada.

Initial and Boundary Value Problems Involving the Inhomogeneous Weber Equation and the Nield-Kuznetsov Parametric Functions

More Related Content

What's hot (16)

Viewers also liked (20)

Similar to Initial and Boundary Value Problems Involving the Inhomogeneous Weber Equation and the Nield-Kuznetsov Parametric Functions (20)

Recently uploaded (20)

Initial and Boundary Value Problems Involving the Inhomogeneous Weber Equation and the Nield-Kuznetsov Parametric Functions