<br>

58 Integral Dquatlons

2.4 Volterra Integral Equation of the First Kind

Theorem. Volterra integral equation of the first kind can be converted to a non
homogeneous Volterra integral equation of the second kind.

An equation of the type

J K(r:5) (E) dE = f(r)

.(1)
is called a Volterra integral
equation of the first kind where K(x,5) and f() are given
function and (r) is the unknown function. Assume
that the kermel K(x,E) and all its
partial derivatives -K(*,),f(r),f(r) and (r) are continuous for 0 sx
f Sa,0sksa lt
is evident that the condition f(0) =0 is necessary for the equation
(1) to have a
continuous solution o (x).

Moréover, the kernel K(,E) possesses a

continuous derivative K(x,E), the first
dr
member of the equation (1) also possesses a continuous
derivative, the same must then
be true for f(r). Differentiating both sides -of
equation (1) by Leibnitz's rule of
differentiation with regard to x, we
obtain
K(x,)9(*) +f, K, (5) (G)
A=
f (1) ..(2)

Conversely, every solution of the equation

(2) also satisfies equation (1), since the two
members vanish for x =0, and their derivatives are
identical.
If K(x, ) does not vanish at any point in an
interval 0,a) then equation (2) may be
written as

(x) =.K(x,r) K,(r,)

K(x,x) ...3)
which reduces to a Volterra integral equation
of the second kind.

If K(x,x) =0 then the equation (2) again reduces to an

equation of the first kind,
which we can treat in the same way, provided
that K(x,E) possesses a continuous second
derivative. Again, differentiating (2) both the sides,
with regard to x, we have
K, (x,1) o(x)+ K,"(*,5) (E) d
-f'
(a)
...(4)

which represents a Volterra integral equation of the second kind

provided
<br>

Equatlons
Soutlon of Integral

=0 we adopt the
K(.) e0, If K,'(r)
successive
same procedure again, and so on. Thus, we
have the sequence of derivatives with
regard to x of the kernel K(x.5) until
derivative K(.) e0 for x ,
we get a In order that the equation (1)has a

intinuous solution, it would be necessary for f(r) to possess continuous derivatives

P)....f)which are all zero for x =0. The first (y
-I) equations obtained by
)
differentiating (1), both the sides, are satistied for x =0.

K()0)+J6 K(.)o () A = f')

(r)= K})
K(r.x)
which is an equation of the second kind provided
K'-x,x) 0.

2.5 Solution of the Fredholm Integral Equation by the

Method of Successive Substitutions
Consider the Fredholnm integral equation of second kind as

r) = F() + [" K(*.$) (G) A

...(1)
where,
(i) the kernel K(r,E) *0 is
real and continuous in the rectangle
Consider K(r,5) sP, where P is the maximum value in R.
R:asxsb,assb.
(i) the function F(r) 0 is real and continuous in an interval.
I:asxsb. Consider |F()| sQ. when Qis the maximum value in the interval.
(ii) A is a non-zero numerical parameter.

Since there exists a continuous solution

o(r) so substituting the unknown function
under an integral sign from the equation (1l) itself, we obtain

o()=F) + [ K() F() A

Proceeding in this manner successively for o(5), we get

o) = Fa)+à Ki.) F() ds + 22 | K(r.)f KG5)

<br>

60 Integral Equations

o (1)=
or F() + 2 [" K(r.E) F(G)d

In general, we get o(1) = F(r) + K(*, ) F() d

K
K(G,Š)[ (,Š2) (š2) džgdš, dž +

K(r,5)[ K(G,Š|)...[" K(šn-lšn) F») dE,n ...dš, dz ..2)

Consider the infinite series
=
o(*) F() + [" K(r.5) F) AF

+)* K(x,5) K(G,Š,) F(š,) d5,dE +t... ..3)

As the kernel K(x,E) and the known function F(E)are real
and continuous, so each term
of the above series represents a continuous, function in I
provided it converges
uniformly in that interval.

Since | K
(x,E) |sP and | F(r)|sQcontains the naximum value in R
and I respectively.
Assume S,() =2"[ K(r.2)[ K(G,Š)...

Then S,(r)|s|2"|Qpb-a)"
It will converge only if

|2|P(6-a) <l»||<p ...(4)

P(b- a)

Thus, the series (2) converges absolutely and uniformly when the relation (3) holds.
Again, let Sn(*) = ar* K(5) [ K(G,Š,)
<br>

Solution of Integral Equations 61

[KG šn) F(G„) dEnr..d5,dE

or |Sp()<|2r| MP (b-a)+l
where M is the max. value of the absolute value of the
function o (x) in I.
If |a|P(b-a) <lthen Lt S,l (x) =0
no
Thus. we notice that the function ¢
(r) which satisfies the relation (2) is the continuous
function given by the series (1). We can verify by diffract
substitution that the function
o(x) defined by (2) satisfies the integral
equation (1). Multiplying (2) both the sides
with Kx.ž) and integrating term by term
within the fixed domain, we have
K(«)[FE) + 1]| K($,) F(Š) d5i

+...+....]dE
The RH.S. may be expressed as

=à[ K(,5) FE) A + Ka.) K(G,

Š) F(Š) dš, dt..
= ¢ (r)-
F(r)
o() = Fr),+ 1[ K(x,) O (E) A

Particular Case
When A =1, the integral equation (l) reduces to
= F(r) +
() K(r.5) O E)
dF

The equation contains a continuous solution, even though relation (4) is not
satisfied,
c.g., the integral equation

has the continuous solution ¢ (x) = x although |2|P (b-«) = 2 1.

The method of solving integral equation of the second kind is based on an iterative
technique which yields a sequence of approximations leading to an infinite series
solution. This is known as Liouville Neumann series.

2.6 Iterated Kernels

The nth iterated kernel K,(x,E) is determined by the relation
<br>

62 Interal Equatione

K,(r.) = [ K(z,2) K-m (z.5) dz m <n

we know that K(z,E)= K(x,5) ...(1)
and K(r,5) =[ K(a,2) Ky(z.)dz,n = 23A,...
...2)
or K.š)=[" K(a.4) K
(4.5) dey
..3)
Substituting (n-1) for n in the equation (2), we have

Kr.3)=S K(z,2)K-(z.) dz

or Knlg.4)=[ K,4) Kygl2.5) de .(4)

Substituting K,- (4,5) in the equation (3), we have

K(.)-[ Ka.)j (42) Kz(g.)de, ds,

or K,(r.5)=[S {Ka4) K(z
4) Kiz (.3) }deyday ..5)
Substituting (n-1) for n in the equation (4), we have

K-gl2.5)=-f K(2.3) Kg (y.3) dzy ..6)

From the relation (5) and (6), we have

K,(z.) =[, SS K(r.z) K (42) K(y3)

K-g (zg 5) dzzdaydey'
..(7)
Performing thís operation successively, we
have

K(zj-5) dz,-|.... dzzdzydi

or K(x.) =[S. K(*,4,K(4.2) K(y.y).
(u-1)th order integral
KEm-liZm) K
(zmsZm+|)...
...
K(Z5) dzy-..dzzdzy dz ...(8)
Similarly, we obtain
=
K,(x,) [S. Kn,K(p.) K(p. P;).KPn-2)
(m -l)th order integral

dpm-..dp,dpydp .(9)
<br>

Solutlon of Integrat Equatlons 63

K-n(e.4)
and fS, S,
..5 Ke.4)K4-m)K( 6) K(aa-mo3)
(-m-1) th order integral

dgn-m-.dqg,day dq .(10)
From the relations (9) and (10), we
have Km(,z) Ky-m(z.E) dz

(m-1) th order integral

(1-m-1)th order integral

dan-m-..dqy dqy
d, }] dz
S K(x,2)Km(e,) dz =
or
fIS.. (n-1) th order
K.A) K(A-2)

K(pm-12) K(z.4)
K(q4.4)....K(4n-m-l.) dq,-m-1 .. dag dq
dz dpm-l .dp, dy ...(1I1)
Now without changing the limits of integration,
the variables of integration are changed
as follows:

Z1 Z2 Z3 -...Zm-1 Zm Zm+l Zm+2 Zm+3

.Z-l
Thus, we have

K(*:4yK(4.z)
(-) th order integral

K(E,,z3).... K(Zm-1m)
K(em "m-t)...K(ep-Š) de,- dzm.dz) d
..(12)

Hence from the relation (9) and (12), we have

K,(,z) =| K,, (x,z) K-m (z.ž) dz, m<n

<br>

64 Integral Equatio

2.7 Method
Solution of the Fredholm Integral Equation by the
of Successive Approximation
The method of successive approximations (iteration techrnique) applied in the soluti
of Volterra integral equation may be extended to the basic Fredholm equation of
second kind.

Consider the Fredholm integral equation of second kind as

o(r) = F()+ ^[" K(r.5) (E) A

where,
(i) the kernel K(r,E) £0 is real and continuous in the rectangle R for which a<x Shan
assb;
(ii) F(r) +0is real and continuous in an interval I, for which a sx sh,
(iii) 2 is a non-zero numerical parameter.

Consider the solution of the integral equation (1 )is in the form of apower series in as

o() = 4 (*) + A q () + 'o, (x)+...+ N'o, (x)+.. ...(2)

where the function , (r),ó (x),o, (x).... and so on are real and continuous in an interval
Let the series (2) is a solution of the integral equation (1 ),
then
o(x) + 2 (r) +...+'o,(x) +....
= F(r) + à [K(,) (4, (5) +à 4(E)+...+ N" o,
(E) +.... d

Equating the coefficients of like powers of , we have

...(3)

9, (x) = F)
<br>

Solutlon of Integral Equatlons 65

...(4)
=
where K(,) K(r,E)
K,(x,) =[ K(*,2) K(z,) A
K,(x,) =[ K(*,2) K,(e,$) dz

and, in general,

K„(x,) = K(x,2)
K- (z,) dz,(n= 23,..) ...5)
The functions K,(,) are called the successive iterated
kernels of the kernel K(x,).
K„(x,) = K(x,z) Kp-m(z.5) dz, n <n
a ...(6)
Consider the series
R(x,;2) = K(x,E)+ (r,) +..+
K, =
Kx,5) +..
R(x,;2)=) NKyx,5)
..(7)

The series on the R.H.S. is called the Neumann series of

the kernel K(x,E). If is an M

upper limit of | K(*,E)| then the iterated

kernel | K(,5)| < M"(b-a at every point
of the domain D. The series (7) is therefore uniformly convergent
in this region, if the
parameter |2|<l/M (b- a). Thus, the relation (2) which gives
the solution of the
integral equation (1 ) assumes the form

o(x) = F()+ [" R(x,5; A)

F()A ...8)
The function R(x, &; ) called the resolvent or the resolving kermel fcr the equation
is
(1). The equation (8) represents an explicit solution of the integral equation in terms of
the resolvent. In view of the relation (7), it follows that the resolvent kernel·R(x,5;2)
satisfies the functional equation:

R(x,;2) -K(x,) = ^ N'K,)

R(r,5;4) - K(r,) = K(x,2) ) NK(z.5) de

R(x,5;)- K(r,) = ^ K(,2) R

(2.5:2) dz

R(x,;2) = K(x,) + ^ | K(x,2) R (z,5:2) dz

<br>

66 Integral Dquatlon

In view of
the relation (7), it follows that the resolvent kermel R(x,;2) satisfies the function.
equation:

R(*,;^)- K(x,) = AS NKa(x,)

p0

R(x,5;2)- K(*,) = ^[ K(x,z)

S A'K„(z,5) V=0
dz

R(x.;^) - K(x,) = ^[ K(*,z) R (z.5;^) dz

R(x,5; 2) = K(r,5) +^ K(r,z) R

(z.5:^) dz
Similarly, we may obtain

R(r.5;2) = K(*,) + ^[K(z.) R (r,z;h) dz

..(9)

The first equation is established by transposing x

by z in (7), then multiplying both
sides by K(x,z) and integrating term
by term in the interval (a,b). Similarly, we may
prove the other relation.

Now, if the resolvent kernel R(x,E;2)

satisfies the equation (9), then
the function o()
in the equation (8) satisfies Fredholm
integral equation (1). Substituting o(x) from (8)
and (1), we have

F() + ^ [" R(r.5:^) F(E)

a d = F(r) + 2 a
K(x,E) +{F(E)

By transposing, and z, we get

Fr)+ ^ [ R(*;N) F(E) dE = F(*) +

{K(r.)

K(.2) R(a.k:A)Jde F()d5 ...(10)

From(9) and (10), we see that the functional
relation (8) is the unique solution of the
Fredholm integral equation (1) when the absolute
value of the parameter is sufficiently
small.
<br>

Solutlon of Integral Equations 67

2.8 Reciprocal Functions

If K(x,) and continuous in R then 3 a reciprocal function K(x,,) provided
is real
that
P(b-) <l, where P, is the maximum of]K(x, E)| in R.
We knów that the iterated kernels
of K(«,E)are represented as
K„(x,) = K(,)K-n (z,) dz, m<n
and K (x,) = K(r,)
Let -K(*ž) = K(x,) + K, (r.)+...+K(x,) + ...(1)
When K(x,E) is real and continuous in R,
the infinite series for k(x,E) is absolutely and
uniformly convergent if P(b-a) <1, Thus, we
have
-k(x,)- K(r,) = K,u,)+ Kg(*.)+...+ K(x,E) t....
Or
-k(x,) – K(*.) = K(,2) K(2, ) dz +...+

..+ a
K(,z)K(z, ) dz +....
Or
-k(x,)– K(*,) = K(x,z) K(z,)dz +...
t..
..+ Ky-l (x,z) K(z,5) dz+..
Or -k(x,) – K(*,) = K
(x,2)+...t...
K(x,z)+...]K(z,5) dz
From (1 ), we have
K(r.ž)+kK*, E)= [ K(a,z) k{z,) dz
(

K(x.) +k(x,5) = [ k{r,z2) K(z,) dz ...(2)

The functions K(x,E) and k(x,E) are said to be reciprocal if

they are both real and
continuous in R and satisfies the condition (2). A function k(x, E)
reciprocal to K(x, )
will exist, provided the series (1 ) converges uniformly.

2.9 Volterra's Solution of Fredholm's Equation

A function k(x,5)) reciprocal to K(x,5) exists where K(x,5) is real and continuous in
R, K(:,ž) £ 0.f() is real and continuous in 1,f) 0,
then the Fredholm's integral
equation of second kind has one and only one continuous solution in the interval
I,asx<h.
<br>

68 Integral Equatlons

Consider the Fredholm integral equation of second kind as

(1) = f(a) +[ K(*,) (E) d5. ..(1)

where the reciprocal function k(*,E) of K(x,E) is known. Let the equation (1)
has a
continuous solution (*), then

o(5) = f(E) +[ K(E.5|)o5) dš:

Multiplying by k(x,) and integrating over the interval (4,b), we have

Again K(.5) (5)ds, =o(*)- f()

=
o() f)x K(*:)f(E)d ...(2)

It follows that the integral equation (1) has a continuous solution and is given
by the
equation (2) which is unique.

Evample 15: Find the iterated kernels for the following kernels :

(i) K(x,E) = x-E, if a =0,b = 1

=
(ii) K(*, E) sin(r-2%);0sx <2r,0 sEs2
Solution: (i) Using formulas (§ 2.6; 6), we obtain in succession :

K(u,) =
(r,)
K =
x-
Or K,r.4)=f,(r-2) (e-) =*-2 3

or

or K,(:9)=- -)(e-})de m(-)*-)

12

K, (.3)--f, r-)--Jaa
or =(-11
<br>

Solutlon of Integral Equations 69

K,
or (:)-, a-9(e-)de =(-1 -*-)
and so, on.
It follows that the iterated kernels are of the form :

(a) Ifn = 2v -1then Kpl (x, E) =)"

l2-T(*)
(b) If n=2v then Klr. E) = )
12"-1 -."=1.2,3
2

(iüi) Using formulas ($ 2.6,6), we obtain in succession:

K (x,E) = K(x,) =
sin(x-2)
K, (*.3)
or =* sin(x-2z) sin(z -2) dz
2
K, (x,5) = cos(x +2-3z)- cos(x-2-z)]dz

K,(u:)=;- sin(r +28-32) + sin(x -28-a) =0

2
=
Or K, (*. ) K(x,z) K,z.) dz =0
or K4 (x,) = 0

Hence the resolvent kernel of the kernel is equal to the kernel itself:
=
R(x,5; 2) sin(x–24)
Thus, the Neumann series consists of one term and, obviously, converges for any
2.
:

Eample 16: Find the iterated kernels for the following kernels
(i) K(x,E)=x+ sin5;a=-,b =r
(ii) K(x,) =e cos ;a = 0,b= m
[Meerut 2008]

Solution: () Using formulas (S 2.6,6), we obtain in succession:

K(x,) =
K(x,) = x + sin
(x+ sin z) (z + sin ) dz = 2r(1+
x
or K, (x,E) =| sinE)

dz = (2n (l+xsin )
(x+sin z) (1 + z sin
E)

or Kg (x,E) =2r
<br>

70 Integral Equatlons

Or K4(.5) =4r
-T
(+ sin z) (r + sin ) dz = (2r)³ (l+xsin )
or Ks (r.5) = 8T J T (*+sin 2) (1 + z sin ) dz = (2n) (*+ sin E)
-T

and so on.
It follows that the iterated kernels are of the form
K(x,) =x+ sin & = K(r.)
K, (x, 5) =2n(l + xsin E)
K, (x,) = (2n' (r+ sin )
K(x,) = (2n)' (l+xsin )
Ks (x,E) = (2r)* (r+ sin E)
= (2n)° (1+ xsin
K(x,) E)

Kgu-1
(r,) = (2ni-2 (x+sin E)
and k,y(x,E) = (2n)- (1+ xsin E) when v = 12,3

(i) Using formulas (S 2.6,6), we obtain in succession:

K(:,) =e cos

or K, (x,) = -[o e cos z.e' cos Edz

K, (x, ) =-e cos

K, (r.ž) = e* cos z. e cos E dz

and so on.
Hence, the iterated kernel reduces
K,(n.)
=(-)r-cos e* cos ,ž,e=
v =
1,2,....
<br>

Solutlon of Integral Equatlons

Pvple 17: Find the first two iterated kernels of the kernel
K(r,5) = (r-E)*;a=-1,b =l

Solution: We know that

K (x,) = K(x,) = (* -E

or K,(x.) =f K(*,) K (e.¿)dz

K,(*,3) = [ (-z
Or (z- dz

4 2

and K,(«.5) =, K(x,2) K, (e,) dz

Ks (r.5)
or =[ (r-z(e+$ +22
+ž+dz
K,(x.) =
or ( +)+-+

Example 18: Find the resolvent kernels for the following

kernels:
= 0
(i) K(x,ž) =(l+ x) (1-);a=-1,b
(ii) K(r.5) =ets;a=0,b =l
[Meerut 2004]

Solution: (i) From the iterated kernels, we know that

= K(x,E) =(1l+
K
(x,) )(|-5)
or K,(x.) =, (l+)(-z) (l + z)
(Q-) dz

K,(r.3) =a +*) (Q-3)

or K,
(r.) = +)(-z) ((+ z) (Q-E) dz

2
K,(.4)-K,(.t)-a+a-9

Thus, in general, we have

K,(x.) =(2 /3)(1 + x)(-)
Hence the resolvent kernel is given by
<br>

72 Integral Eqúation

R(.5:2) =)K(r.5) =
4+ ) |-8) E()
R(xE:A) = (1-E) where
or 3l+ x)0-5) | 2| < 23
3-22
(ii) We know that
K{ (x,E) = K(x,E) = e+$

Or K,(r.5):

Or Kg(.) =, ede
K, (x.) =
2

In general, we have

K9-(
Thus, the resolvent kernel is given by
-

H=l n=l

R(r,;^) = 2e+$ 2
where|^|<2-1
2-2(e-l)
Evnple 19: Solve the following linear integral equations :
(i) -I/2
() =r+fo 5) d
(i) o
(r) = (sin x-) + /2 r o(E)
4 d
Solution: (i) Here K(r,E) =
1, 2.
F(r)=x
=l and
=
We know that K,(x.5) K(r.5) =1

Or K, (r.4) = fo dz =

or
<br>

Solution of Integral Equations

73

Kas.3)-G)

Thus, the resolvent kernel is given by

RU.k:3)
-) 2
2-2
or R(x,;1) =2, as
=l
Hence, the solution of
the integral equation is given by

o(1) = F() + R(*,5;2) F(E)

d
or ¢(*) =+ rl/2 25 d =x+*=x+ constant.

(ii) Here K(x,) =x, =,F(x)

A =
sin x
4 4

We know that K,
(x,) = K(r.) =r
or K,(x.) =
x T/2
zx.z dz

Kz(.) =**fo

3
or dz

3
or

Thus, the resolvent kernel is given by

R(x.5:)
=-r 24)
96
or
96-3
<br>

74 Integral Equatlons

Hence the solution of the integral equation is given by

o (1) = sin r- 4 I 96
496-10
24 x 96
or o (r) = sin
x–+
4 96-n3 96
= sin x

Example 20: Solve the integral

equation:

o)-(-eš0$)4$
[Kanpur 2005]
Solution: Here K(r.8) = 3
,à=;.F() =e-e-5
We knowthat K(x,5) = K(r,) =

or

or

or K,(x,E)
=gŠ
Hence, the resolvent kernel is given by

25
n=l 2-2

Or R(x.:4) =,= 4
3
Thus, the solution of the integral equation
is given by

or
---)
<br>

Solutlon of Integral Equatlons

75

E2

Or

EuDple 21: Solve the following

integral equations :

(i)

Solution: (i) Here K(x,E) = = 5x

x$,A =F(r)
2 6
We know that K(x,) =
K(x,) = x,
or

or

-1

Hence,the resolvent kernel is given

by
1-1
R(*.5;^) =x, 3x
3-4
or R(x,5:4)
=. 6

Thus, the solution of the integral equation is given

by
(x) = 5x,
1 6 1
5x

(ii) Here K(s,) =1,4 =.F(r) =e-et

We know that K{(x,) = K(x,) =1
or K, (x:ž) =fo dz =l

or
<br>

76 Integral Equatlons

or K„(,E) = 1

Thus, the resolvent kernel R(x, 5;2) is given by

1

R(*.;^)
=* -} =2

or
o) =(e-}e+t)+fo (e
-}o+)a4
Or

Exmple 22: Solve the integral equations;

=
(1) o(*) f(a) + S
ctšo(G) d5

(ii) (*) =l+ ^ [ sin(r+E) ¢ (5)

d
[Meerut 20031
Solution: (i) Here K (x,E) = K(x, E) = e*-s

Or

Proceeding in this way, we find that all the iterated kernels coincide with K(x,E). Thus,
the resolvent kernel becomes
r(x,5;^) = K(x,E)(l+à+2+....) =
r-3) /(-2)
The resolvent kernel is an analytic function of except at the point =l, which is a
simple pole of the kernel T.

The solution of the equation is given by

o(4) =
f)-fE)4
(ii) Here K(x,) = K(x,) = sin(x + 5)
K, (x, ) =sin(x
Or + t)
sin(t + E)dt = cos(x-)
or Kg
(x, ž) = sin(x + t) cos(t-ž)dt = ( n
sin(x + )
Proceeding in this way, we have
K4(x,) =(G n)' cos(x-), Kç
(x,) =() sin(x+)
K
(x,) =(G n) cos(x-), K(«,) =( n)°sin(x +8)
<br>

Solutlon of Integral Equations 77

The resolvent kernel R

(x,E;2) is given by

R(,5;A) =) K(x,E)

R(x,5;2) = K
(x,5) + K,(r,) +
or 2 Kg(xr.ž) +aK, (x.5)
+x* Ks(x.E)+K,K(x, E)+...

=
or R(*,;2) sin(x

2
or R(.:A)=|sin(r + B) + cos(x
2

Or R.:)=|sin(a+)+cos 1

Provided

= 2
or R(x,;2) 4-22,2l2 sin(x +) + Ar cos(r-E)1
Thus, the solution of the integral equation is
given by

(1) = f(r)+^[R(x,5:^)F(E) dE
22
(x) =1+ {2 sin(x+E) + Ar cos(x -E)}dE
4-12
22
(x) =1+ (-2 cos(x +E)-Ar sin(x-E)h
4-2
(r) =1+ 42 [2 cos x + Ar sin x] where| 2|<
4-'n2 T

Example 23: Solve the integral equation

o() =l+^[,1-3x) >(E)d

For what values of A the solution does not exist ?

Solution: The integral equation is given as

o(x) =1+af -3x) o(E) d5

...1)
=
Here fa) =1, K(.) =1-3x5,h I(say)

We know that = K(x,E)=1-3xE

K(x,5) ..(2)
<br>

78 Integral Equatlons

and K,(r,)=[, K(*-z) K(z.3) de ..(3)

Substituting p =2 in (3), we have

K, (x.ž) =J K(*,z) K (z,E)dz

or K, (1.) =Ja a-3xz)1 -32)de

Or K,(«.) = fo f1-3z(r +) +9x="}dz

or K,(*.3) =1-+)+3r$ ..(4)

Substituting p =3 in (4), we have

K, (x.) = o K(*,z) K, (. )dz
Or

or KsuJ-|; (-)--*-)-9-)=
or K,(r.) =(-3x) =4(,) ..(5)
Similarly K (x,) =[, K(a,2) K, z) dz

or K,(,5) =S, -3xz)(1-328) dz

4 J0

or K,(.9)--u+B)+318)-; Kz:3) ...(6)

K, (x,) =
and [, K(*,2) K4(2.)dz

or K,u.9)-|,0-3e)-e+)+3*h4e
Or K,(.9)-( 4-31j) -( K,a.)
.(7)

Thus, we have K, (x,) = l-3xk,

Kg(u.) =1-u+)+3
K,(*,) =(-32).
<br>

Solutlon of Integral Equations

79

Ku.)--u+)+3*
K,:.)-(a-314).
K0:9)-(4-u+-3*
The resolvent kernel is given by

R(r.:A) -) AK,(r.5)

or R(r.;A) =[K(r.) + 2? Kg (r,E) +

K(r.)+..]
+
[K, (r, ) + K(u5) + KG(r,E)+.....
or R(*,:2) =(|

+a-}*)*3:4}
or

or R(r,5:2)= 4
4-22 ...(8)
Hence, the solution of the integral
equation is given as
() = f() + ^ R(r,5;2) o () dE

o
(r) =1+ 42
4-22
or o 42 4+d-6x 4 +22(2 -3x)
(r) =1+ |2|<2
4-22 4 4-2 ..(9)
From the equation (9), we notice
that the solution exists only when |2|<2. The
solution of the integral equation does not exist
for||>2
Example 24: Solve the non-homogeneous Fredholm integral equations of the second
kind, by the method of successive approximations to the third order.
(i) () = 2r+ ^ [ (r+E) o () d, (x) =1

(ii) (r) =l+ ^ | (r+E) o

(5)d5,b(r) =1
<br>

80 sIntegral Equatione

Solution: The integral equation is given as

.(1)
,(x) denotes the nth order approximation,
Consider then,we have
o,(x) =2x + ^ f, (x+5) -1 (5) dE ..(2)
Substituting n =
1234,....n the equation (2), we have

or ...(3)

Also
, (x) =
2x+ f (*+E) g (5) 45

or

Or ) (*)=2x+ ^ f, (x+5) } +5(2 + 2)} dz

or
...(4)

and
, (*) = 2x +^ , (5)
f (x+5) d
or

or 722
3 12
722
or 9-2***f
12

(22 72 +2x +
+5
3
+
12 hr+x+
or

(ii)The integral equation is given as

o(x) =1+ (x+E) (5) dE , , (x) =1
...(1)

Consider ,(x)denotes the th order approximation, then, we have

,(x) =1+ f (x+E)0,(5)d5 ...2)

<br>

Solutlan of Integral Equatlons 81

Substituting n = 1,2,3,4,... in the equation (2), we have

.(3)

Also (r) =l+ (+8)4 (E)

d
Or

..(4)
and

Or

or
12*
+ xà + xà²
12

or

E
XERCISE-2.2

1. Find the iterated kernel of the following kernels:

(a) =
K(x,) sin(-E);a
=
0,b = n/2 for n = 23
[Kanpur 2002)
=
(b) K(*) xeš; a=0,b =1
= 2
(c) K(,E) =e-l; a=0,b =l, for

2. Construct the resolvent kernels of the following kernels:

(a) K(x,) =l ;a =0,b =1
(b) K(x,)= sin.x cos;a = 0,b = n/2
K(x.) x;a=-l,b =1
=
(c)
=l
(d) K(x,E)=;a=-1,b
<br>

82

3. Show that the resolvent kernel of the equation

O(x) = ( R(:A)0(5 is equal to R(x,5:A +#)
F(r+)+
4. Solve the following integral equations:
(a)

(b) o(r) =r+ ^ [l+ 3r) O(5)d

5. Solve the fredholm integral equation:
o(r) =l+ à i+3r)0(E)d
6. Prove that the mh iterated kernel K,(x,E) satisfies the relation
K(x) = [K,(xt)Km-, (t.)dt,
where r is any positive integer less than n.

ANSWERS-2.2
=
(a) Kg (r,ž) sin(x+)-*cos(r -E)
4

Kz (x,) =
4-n2
-sin(r-)
16
(b) K(x,5) =xe

eš2--]+(-x-)ei-,0sxsE
(c) K (x,) = +
eš-x-š] (r --I)e*,sxsI
(a) R(x,5:2) = 2
2 sin x cos
(b) R(r,5:2) = |<2
2
3xk
(c) R(x,:2)= 3-24^|<3/2

(d) R(r.:)- Sa' -|à|<5/2

5-22
(a) o(r) =x
3x
(b) o(x) = -+3
3-2.
o(x)=[4 +4A(2 -3x)] /(4-2)