Theoretical Mathematics & Applications, vol.3, no.3, 2013, 121-143 ISSN: 1792- 9687 (print), 1792-9709 (online) Scienpress Ltd, 2013 Asymptotic behaviour of boundary condition functions for a fourth-order boundary value problem Emmanuel K. Essel 1 , Ernest Yankson 2 and Samuel M. Naandam 3 Abstract In this paper, we prove that the Boundary Condition Constants for the Boundary Value Problem π ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( : 4 ) 1 ( 3 ) 2 ( 2 ) 4 ( x x x P x x P x x P x L λφ φ φ φ φ φ = + + + ≡ { } 0 ) ( ) ( 4 1 ) 1 ( ) 1 ( = + ≡ ∑ = − − s s rs s rs r b n a m U φ φ φ ( ) 4 , 1 ≤ ≤ s r can be replaced by Boundary Condition Functions and that the Boundary Condition Functions are asymptotically equivalent for large values of λ , to the Boundary Condition Functions for the corresponding Fourier problem , given by F π : ) ( ) ( ) 4 ( x x λφ φ = { } 0 ) ( ) ( 4 1 ) 1 ( ) 1 ( = + ≡ ∑ = − − s s rs s rs r b n a m U φ φ φ ( ) 4 , 1 ≤ ≤ s r . 1 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. 2 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. 3 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. Article Info: Received: June 30, 2013. Revised: August 30, 2013. Published online : September 1, 2013
23
Embed
Asymptotic behaviour of boundary condition functions for a ... 3_3_10.pdf · Asymptotic behaviour of . boundary condition functions . for a fourth-order boundary value problem . Emmanuel
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Emmanuel K. Essel1, Ernest Yankson2 and Samuel M. Naandam3
Abstract
In this paper, we prove that the Boundary Condition Constants for the Boundary
Value Problem
π )()()()()()()()(: 4)1(
3)2(
2)4( xxxPxxPxxPxL λφφφφφφ =+++≡
{ } 0)()(4
1
)1()1( =+≡ ∑=
−−
s
srs
srsr bnamU φφφ ( )4,1 ≤≤ sr
can be replaced by Boundary Condition Functions and that the Boundary
Condition Functions are asymptotically equivalent for large values of λ , to the
Boundary Condition Functions for the corresponding Fourier problem 𝜋, given
by
Fπ : )()()4( xx λφφ =
{ } 0)()(4
1
)1()1( =+≡ ∑=
−−
s
srs
srsr bnamU φφφ ( )4,1 ≤≤ sr .
1 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. 2 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. 3 Department of Mathematics and Statistics University of Cape Coast, Cape Coast Ghana. Article Info: Received: June 30, 2013. Revised: August 30, 2013. Published online : September 1, 2013
122 Asymptotic behaviour of boundary condition functions …
Mathematics Subject Classification: 35B40; 34B05
Keywords: Boundary condition functions; fourth order boundary value problem
and asymptotic behaviour
1 Introduction
Over the years quite a good number of mathematicians have studied
boundary condition functions. The use of boundary condition functions for
boundary value problems was first considered by Kodaira in [1].
In this paper Kodaira considered the replacement of boundary condition
constants of separated boundary conditions , associated with real differential
equations of arbitrary even order , by solutions of the differential equation. In [2]
E. C. Titchmarsh proved that, the self-adjoint boundary value problem:
)()()()()2( xxxqx λφφφ =+− (-∞< a ≤ x ≤ b < ); τσλ i+=∞
(1)
(1)
( ) cos ( )sin 0( )cos ( )sin 0a ab b
ϕ α ϕ α
ϕ β ϕ β
+ =
+ =
is equivalent asymptotically, for suitably large values of λ , to the corresponding
Fourier problem:
)()()2( xx λφφ =−
(1)
(1)
( ) cos ( )sin 0( )cos ( )sin 0a ab b
ϕ α ϕ α
ϕ β ϕ β
+ =
+ =
The coefficient q and the constantsα , β are real valued and [ ]baCq ,∈ . W.
N. Everitt in [3] also worked on self-adjoint boundary value problems. D. N. Offei
in [4] extended the use of boundary condition functions to non-self adjoint
boundary value problems with complex-valued coefficients and constants and
with boundary conditions separated or otherwise.
E.K. Essel, E. Yankson and S.M. Naandam 123
In [5], D. N. Offei proved that the boundary condition functions, for the
boundary value problem:
)()()()()()( 3)1(
2)3(3 xxxPxxPxiL λφφφφφ =++≡
0)()()( )1( === bba φφφ
are asymptotically equivalent, for suitably large values of λ , to the corresponding
functions, for the corresponding Fourier problem. In [9], M. B. Osei showed that
the boundary condition functions of the second order boundary value problem are
asymptotically equivalent to the boundary condition functions of the
corresponding Fourier problem of the boundary value problem.
2 Notation
In this section we give some properties of the linear differential expression L
and some notations used in subsequent sections of this paper.
1. For a suitable set of functions ),(xrφ ),41( ≤≤ r the symbol )(xΦ denotes
the 4 x 4 matrix )]([ )1( xsr
−φ )4,1( ≤≤ sr .
Thus,
)(xΦ .
)()()()()()()()()()()()(
)()()()(
)3(4
)3(3
)3(2
)3(1
)2(4
)2(3
)2(2
)2(1
)1(4
)1(3
)1(2
)1(1
4321
≡
xxxxxxxxxxxx
xxxx
φφφφφφφφφφφφφφφφ
Also )(ˆ xg represents the column vector with components
(1) ( 1)( ), ( ),..., ( ).ng x g x g x−
2. The symbol *( )xB denotes the conjugate transpose of the matrix ( )xB
whilst )(ˆ* xb denote the row vector with components (1) ( 1)( ), ( )... ( )nb x b x b x− .
3. Given the linear expression 𝐿 defined by
124 Asymptotic behaviour of boundary condition functions …
)()()()()()()()()()( 4)1(
3)2(
2)3(
1)4(
0 xxPxxPxxPxxPxxPL φφφφφφ ++++≡
)( bxa ≤≤ . The Lagrange adjoint of L is denoted by +L and defined as
ψψψψψψ )())(1()()1()()1()()1( 4)1(
3)2(
22)3(
13)4(
04 xPPPPPL +−+−+−+−≡+ .
4. (a) For suitable pairs of functions f and g
∫b
a
gLfLfg +− dx = )]([)]([ afgbfg − .
Here [fg](x) is a bilinear form in ),,,( )3()2()1( ffff and
),,,( )3()2()1( gggg given by
[ ]4 4
( 1) ( 1) *
1 1
ˆˆ( ) ( ) ( ) ( ) ( ) ( ) ( ) j kjk
j kx B x g x f x g x xfg f x− −
= =
= =∑∑ B
where,
−−
−−−−+−−+−−−+−
=
000)(00)()(3)(0)()(2)()(2)(2)(
)()()()()()()()()()()(
)(
0
0)1(
01
0)1(
01)2(
0)1(
12
0)1(
01)2(
0)1(
12)3(
0)2(
0)2(
1)1(
2)3(
3
xPxPxPxP
xPxPxPxPxPxPxPxPxPxPxPxPxPxPxPxPxP
xB
(b) If ),(1 xP )(2 xP and )(3 xP are identically zero in some neighbourhood of
a and b and 0P is a constant independent of x then
−
−==
000000000
000
)()(
0
0
0
0
PP
PP
ba BB .
(c) The Lagrange adjoint of +L is L and for suitable pair of functions g and f
{ }∫ −=−+b
a
agfbgfdxLfggLf )}({)}({
where
{gf }(x) =∑∑= =
−−4
1
4
1
)1()1( )()()(j k
kjjk xgxfxA = )(ˆ)()(ˆ * xgxxf A .
E.K. Essel, E. Yankson and S.M. Naandam 125
The jkA are dependent on the coefficients of the differential expression +L
and ][)( jkAx =A .
5. If ( , )xϕ λ is a solution of Lϕ λϕ= and ( , )xψ λ is a solution of Lψ λψ+ =
then
{ }
{ }
2
1
2 1 1 2[ ]( ) [ ]( ) ( )
0
x
x
x x L L dx a x x b
dx
ϕψ ϕψ ψ ϕ ϕ ψ
ψλϕ ϕλψ
+− = − ≤ ≤ ≤
= −
=
∫
∫
and hence,
)]([)]([ 12 xx φψφψ = .
Thus, ))](,(),([ xxx λψλφ is independent of ],[ bax∈ .
Similarly ))}(,(),({ xxx λφλψ is independent of ],[ bax∈ . This implies that
))](,(),([ xxx λψλφ and ))}(,(),({ xxx λφλψ may be denoted by ][φψ and
}{ψφ , respectively.
6. (a) If there is a constant 𝐾 such that )()( xKxf φ≤ for 0xx ≥ we write
).(φOf =
(b) If lxxf
→)()(
φ , ∞→x where 0≠l we write φlf ~ .
.
3 Preliminaries
The boundary value problem to be considered is of the form
π : )()()()()()()()( 4)1(
3)2(
2)4( xxxPxxPxxPxL λφφφφφφ =+++≡ (1)
{ } 0)()(4
1
)1()1( =+≡ ∑=
−−
s
srs
srsr bnamU φφφ ( ),4,1 ≤≤ sr (2)
126 Asymptotic behaviour of boundary condition functions …
where the functions )(2 xP , )(3 xP , )(4 xP , the constants rsm and rsn and the
parameter λ are complex- valued. The functions )(xPr ( r = 2, 3, 4) are of the
class )4( rC − on the closed bounded interval [a, b] and )(2 xP , )(3 xP are identically
zero in a neighbourhood of both a and b .
The corresponding Fourier boundary value problem for π is given by
Fπ : )()()4( xx λφφ = (3)
{ } 0)()(4
1
)1()1( =+≡ ∑=
−−
s
srs
srsr bnamU φφφ ( )4,1 ≤≤ sr . (4)
Let )},/();,/({ λχλψ xbxa rr )41( ≤≤ r be the boundary condition
functions for π and )},/();,/({ λχλψ xbxa FrFr )41( ≤≤ r the boundary condition
functions for Fπ . Then ),/( λψ xar and ),/( λχ xbr are solutions of ψλψ =+L
such that
*)()( MBΨ aa = and *,)()( NBΧ bb =
where
−
−==
0001001001001000
)()( ba ΒB , M= ,
44434241
34333231
24232221
14131211
mmmmmmmmmmmmmmmm
and N= .
44434241
34333231
24232221
14131211
nnnnnnnnnnnnnnnn
Likewise ),/( λψ xaFr and ),/( λχ xbFr are solutions of )()()4( xx ψλψ = such
that
*)()( MBΨ aa = and .*)()( NBΧ bb =
Let )},/(),,/({ λλ xbgxaf rr )41( ≤≤ r be the boundary condition
functions for the boundary value problem
E.K. Essel, E. Yankson and S.M. Naandam 127
),()()4( xx λφφ = and )(ˆ)(ˆ 44 bIaIU φφφ += =
0000
,
where I 4 is the 4 x 4 unit matrix. Then ),/(),,/( λλ xbgxaf rr are solutions of
)()()4( xx ψλψ = such that
F(a) = B(a)I 4 = B(a) and G(b) = B(b)I 4 = B(b).
4 Proof of Theorems
We now prove five Theorems that will enable us to prove our main results in
Theorem 6.
Theorem 1
(i) ),/( λψ xaFr = ),/(4
1λxafm s
srs∑
=
.
(ii) ),/( λχ xbFr = ),/(4
1λxbgn s
srs∑
=
.
(iii) Let ( ) ( / , )s sf x f a x λ= , ( ) ( / , )s sg x f a x λ= .
Then
)(xf s = )()1( )1(1
)1( xf ss −−− 42 ≤≤ s
)(xgs = )()1( )1(1
)1( xg ss −−− .42 ≤≤ s
Proof. (i) and (ii) ),/( λxaf s and ),/( λxbgs are solutions of
)()()4( xx ψλψ = (5)
such that
F(a) = B(a)I 4 = B(a) (6)
128 Asymptotic behaviour of boundary condition functions …
i.e.,
)()()()()()()()()()()()(
)()()()(
)3(4
)3(3
)3(2
)3(1
)2(4
)2(3
)2(2
)2(1
)1(4
)1(3
)1(2
)1(1
4321
afafafafafafafafafafafaf
afafafaf
=
−
−
0001001001001000
.
Similarly G(b) = B(b)I 4 = B(b).
Let ),( λxhr = ),/(4
1λxafm s
srs∑
=
, then ),(1 λxh , ),(2 λxh , ),(3 λxh and ),(4 λxh are
solutions of (5) such that
H(a) =
)()()()()()()()()()()()(
)()()()(
)3(4
)3(3
)3(2
)3(1
)2(4
)2(3
)2(2
)2(1
)1(4
)1(3
)1(2
)1(1
4321
ahahahahahahahahahahahah
ahahahah
=
)()()()()()()()()()()()(
)()()()(
)3(4
)3(3
)3(2
)3(1
)2(4
)2(3
)2(2
)2(1
)1(4
)1(3
)1(2
)1(1
4321
afafafafafafafafafafafaf
afafafaf
.
44342414
43332313
42322212
41312111
mmmmmmmmmmmmmmmm
This implies that
H(a) = F(a)M* .
But from (6) F(a) = B(a), therefore
H(a) = B(a)M* .
Now ),/( λψ xaFr )41( ≤≤ r are solutions of the same (5) such that
*)()( Maa B=Ψ .
Hence we have ),/( λψ xaFr = ),( =λxhr ),/(4
1λxafm s
srs∑
=
).41( ≤≤ r
Similarly if
),( λxqr = ),/(4
1λxbgn s
srs∑
=
, (7)
E.K. Essel, E. Yankson and S.M. Naandam 129
then ),( λxqr )41( ≤≤ r are solutions of (5) such that Q(b) = B(b)N* and so
),/( λχ xbFr = ),( λxqr = ),/(4
1λxbgn s
srs∑
=
.
This proves (i) and (ii).
(iii) ),/( λxaf r )41( ≤≤ r are solutions of )()()4( xx ψλψ = such that
F(a) = B(a)I 4 = B(a)
)()()()()()()()()()()()(
)()()()(
)3(4
)3(3
)3(2
)3(1
)2(4
)2(3
)2(2
)2(1
)1(4
)1(3
)1(2
)1(1
4321
afafafafafafafafafafafaf
afafafaf
=
−
−
0001001001001000
(8)
The general solution of the equation )()()4( xx ψλψ = can be obtained as follows:
Let )sin(cos4 θθλρ ir +== for πθπ ≤<− , .τσρ i+=
Then,
+
+
+
=42sin
42cos4
1 kikrkπθπθρ (k = 0, 1, 2, 3).
Hence, for 𝑘 = 0 we have
+
=
4sin
4cos4
1
0θθρ ir (9)
Since the fourth roots of unity are 1, -1, i and –i we see that the four roots of the
equation λρ =4 are 0ρ ,- 0ρ , -i 0ρ and i 0ρ . Put 0ρ=P then the four roots are P,
-P , iP and –iP. The general solution of the equation
0)()( 4)4( =− xx ψρψ )( 40 λρ = (10)
is thus given by .)( )(4
)(3
)(2
)(1
axiPaxipaxpaxp eAeAeAeAx −−−−−− +++=ψ
Now ),/( λxaf r is a solution of (10) such that F(a) = B(a), and so
Then the boundary condition functions for the boundary value problem (1) and (2)
are given by
),/( λψ xar = ),/( λψ xaFr + ∫x
ar dttxtR )(),/( ψλ )41( ≤≤ r
),/( λχ xbr = ),/( λχ xbFr + ∫x
ar dttxtS )(),/( χλ ).41( ≤≤ r
Proof. As a function of x, ),/(1 λxtf (x, t ]),[ ba∈ is a solution of
)()4( xψ = )(xψλ such that from (8)
),/( 001 λxxf = ),/( 00)1(
1 λxxf = ),/( 00)2(
1 λxxf = 0 (18)
and
),/( 00)3(
1 λxxf = −1 where 0x ].,[ ba∈ (19)
It follows from (17), (18) and (19) that ),/( λxtR is a solution of )()4( xψ = )(xψλ
such that
0 0(1)
0 0 2 0(2)
0 0 3 0(3)
0 0 4 0
( / , ) 0
( / , ) ( )
( / , ) ( )
( / , ) ( )
R x xR x x P xR x x P xR x x P x
λ
λ
λ
λ
=
= −
= = −
(20)
If ),/( λψ xar = ),/( λψ xaFr + ∫x
ar dttxtR )(),/( ψλ ).41( ≤≤ r (21)
E.K. Essel, E. Yankson and S.M. Naandam 133
Then using the formula
( , ) ( , ) ( , )x
a
d FF x t dt x t F x xdx x
δδ
= +∫ ∫ and (20) we obtain the following equations
from (21)
(1) (1) (1)( / , ) ( ) ( / , ) ( ) ( / , ) ( )x
r Fr r ra
a x x R t x t dt R x x xψ λ ψ λ ψ λ ψ= + +∫
(2) (2) (2) (1)
(2) (2)2
( / , ) ( ) ( / , ) ( ) ( / , ) ( )
( ) ( / , ) ( ) ( ) ( )
x
r Fr r rax
Fr r ra
a x x R t x t dt R x x x
x R t x t dt P x x
ψ λ ψ λ ψ λ ψ
ψ λ ψ ψ
= + +
= + −
∫
∫
(3) (3) (3) (2) (1)2
(3) (3) (1)3 2
( / , ) ( ) ( / , ) ( ) ( / , ) ( ) [ ( ) ( )]
( ) ( / , ) ( ) ( ) ( ) [ ( ) ( )]
x
r Fr r r rax
Fr r r ra
a x x R t x t dt R x x x P x x
x R t x t dt P x x P x x
ψ λ ψ λ ψ λ ψ ψ
ψ λ ψ ψ ψ
= + + −
= + + −
∫
∫
(4) (4) (4) (3)
(1)3 2
(4) (4)4
(2)(1)3 2
( / , ) ( ) ( / , ) ( ) ( / , ) ( )
( ) ( ) [ ( ) ( )]
( ) ( / , ) ( ) ( ) ( )
[ ( ) ( )] ( ) ( ) .
x
Fr r ra
r rx
Fr r ra
r r
a x x R t x t dt R x x x
P x x P x x
x R t x t dt P x x
P x x P x x
ψ λ ψ λ ψ λ ψ
ψ ψ
ψ λ ψ ψ
ψ ψ
= + +
+ −
= + −
+ −
∫
∫
By definition, ),/( λxtR being a solution of )()4( xψ = )(xψλ implies that (4) ( / , ) ( / , )R t x R t xλ λ λ= and so
(4) (4) (4)
(2)(1)4 3 2
( ) ( ) ( / , ) ( )
( ) ( ) [ ( ) ( )] ( ) ( )
x
r Fr ra
r r r
x x R t x t dt
P x x P x x P x x
ψ ψ λ λ ψ
ψ ψ ψ
= +
− + −
∫ (22)
Substituting (21) into (22) we have (2)(4) (4) (1)
4 3 2( ) ( ) [ ( ) ( )] ( ) ( ) [ ( ) ( )] ( ) ( ) .r Fr r Fr r r rx x x x P x x P x x P x xψ ψ λ ψ ψ ψ ψ ψ = + − − + − But, )()4( xψ = )(xψλ implies that )()4( xFrψ = )(xFrψλ and hence,
134 Asymptotic behaviour of boundary condition functions …