A NOTE ON BESSEL FUNCTION DUAL INTEGRAL

A NOTE ON BESSEL

FUNCTION

DUAL INTEGRAL

EQUATION

WITH

WEIGHT FUNCTION

B.N. MANDAL

Department

of Mathematics

University of Calcutta 92 Acharya Prafulla Chandra Road

Calcutta 700009

INDIA

(Received February 16, 1987 and in revised form

May

15,

1987)

ABSTRACT.

An

elementary procedure based on

Sonine’s

integrals has been used to reduce dual integral equations with Bessel functions of different orders as kernels and an arbitrary weight function to a Fredholm integral equation of the second kind. The result obtained here encompasses many results concerning dual integral equations with Bessel functions as kernels known in the literature.

KEYS WORDS

AND

PHRASES. Dual integral equations, Bessel function, Sonine’s integrals, Hankel inversion, Fredholm integral equation.

1980 AMS SUBJECT CLASSIFICATION CODE. 45F.

I.

INTRODUCTION.

In

a recent paper Nasim

[I]

considered the following class of dual integral

equa-tions involving Bessel function of first kind of different orders as kernels and arbitrary weight function

I

t-2

_J9

(xt)

[l+w(t)] #(t)

dt f(x), 0 < x <

I,

o

I

t

-28

J

(xt) #(t)

dt-

g(x),

x >

I.

o

An

operational procedure based on exploiting the properties of Mellin transform is used in

[I]

to reduce (I.i) to a single integral equation involving Bessel function as kernel.

By

Hankel inversion this was then reduced to a Fredholm integral equation of second kind. However, the general result in

[I]

was obtained under the assumptions that n-

81

<

+I

and E 0, although the second assumption was not clearly stated. This is necessary to make % >

-I

(where %

+

-a

+

8)

for the Hankel inversion to be valid.

It

is not clear whether this operational procedure involving Mellin transform, can be used if these restrictions on the parameters p,9,,8 are not

Earlier Noble

[2]

used the

’multiplying-factor method’

based on

Sonine’s

integrals to a class of dual integral equations with Bessel functions of same order

(

)

as kernels to a Fredholm integral equation of second kind. Schmeltzer and Lewin

[3]

also used this multiplying-factor method to reduce another class of dual integral equations involving first kind Bessel functions of different orders and a particular weight function to dual integral equations with trigonometric sine function as kernels in closed form by the function-theoretic method. Later Ross

[4]

gave a simplified method of construction of explicit solution to the class of dual integral equations considered in

[3].

Also Rose and de

Hoog

[5]

considered another class of dual integral equations with some particular weight functions for explicit solution. These dual integral equations arose when standard integral transform approach was used

to solve some mixed boundary value problems were solved by using a complex variable technique. The explicit solution to the dual integral equations were then deduced from the solutions to the boundary value problems.

Except

for these special class of dual integral equations considered in

[3,4,5],

the general class of dual integral equations (I.I) does not appear to admit of explicit solution. References to most of the works on different dual integral equations with Bessel function kernels can be found in

[1,2,4,6].

While in

[2]

dual integral equations with first kind Bessel functions of the same order as kernels were considered, in the present paper as well as in

[I],

integral equations with Bessel functions of different orders as kernels are considered. Here we employ the multiplying-factor method of Noble

[2]

to reduce

(I.I)

to a single integral equation with Bessel function as kernel. Invoking Hankel inversion, this single integral equation is then reduced to a Fredholm integral equation of second kind. The result is obtained for all arbitrary values of the parameters

,u,a,B.

From this result many known results concerning dual integral equations with Bessel function kernels can be deduced.

As

special cases, three sets of values of the para-meters are considered. The general result obtained here then reduces to the known results for these special cases.

2. REDUCTION TO

A

FREDHOLM INTEGRAL

EQUATION

OF SECOND KIND.

We assume that the parameters

,,a,B

appearing in

(I.I)

are most general. By using the formulae

Id p

( )

[x

-

_J

(xt)]

(-I)

p x--p tp

_J+p

(xt),

(2.1)

d q

x-q q

(xt)

(2.2)

(

_x

[x

_J

(xt)

t

_{J_q}

where p,q are non negative integers, it is always possible to change the orders

,

to

’,’

respectively such that

’

I,

’

>

-3/2

(of course,

a,B

are then also

changed)

by choosing p suitably. Thus without any loss of generality, we can assume that the orders

,

of the Bessel functions in

(I.I)

satisfy the restrictions

>

-I,

>

-3/2.

These restrictions are necessary in the analysis that follows. 2

x2

+I

Multiplying each side of the first equation in (i.I) by

(r

x where

r

I

(r

2-x

2)

x

+I

J (xt) dx o

2

i(+l)

t--i

r

+V+l

_Ju++l(rt),

>-I,

>-I,

(2.3)

we obtain

++I

-2--I

2

F(+I)

r

f

t J

k+l

(rt)

[l+w(t)] (t)

dt o r

x+l

2

f

(r2-x)

f(x) dx, 0 < r <

I.

(2.4)

o

Similarly, multiplying each side of the second equation in

(I.I)

by

(x2-r2)

D x

-+I

where >

-I,

integrating from r(>l) to and using Sonine’s second integral in the form

I

(x2-r2)

x

-+I

_J

(xt)

dx r 2

F(+I)

t

-n-I

r

-+n+l

J

--I

(rt), -I <

(2.5)

we obtain 2rl

r

(n+l)

r

-+rl+l

I

t

-2B-rl-1

J

(rt) (t)

dt o

--I

I

(x2_r2)n

x-+l

g(x)

dx, < r <

.

r

(2.6)

Using

(2.2)

(with q replaced by and x replaced by

r)

in

(2.4),

we obtain

f

t-2a--l+E

_Jv++l-

(rt)

[l+w(t)] (t)

dt o

?

r(+l)

r

(7

o 2

_x2)

(r

f(x)dx, O<

r< (2.7)

where is a nonnegative integer. Similarly using (2.

I)

(with p replaced by m and x replaced by

r)

in

(2.6)

we obtain

f

t

-28--I+m

J (rt)

(t)

dt

-n-l+m

o

(-I)

m 2

-n

_r-n-l+m

m r(n+l)

(3

r

f

(x2-r2)n x-V+l

o

where m is a nonnegative integer.

In

(2.7)

and

(2.8)

we now equate the powers of t and the orders of the Bessel functions. This gives two equations to determine

,q

as

g(x)

dx, r >

I, (2.8)

-2a-

C

+

=-28

n

+

m

u++

I-

,= -n-

l+m.

2 u

+

a

B

+

m

2

(2.9)

The requirements that and

n

must be greater than

-I

(cf (2.3) and

(2.5))

can be satisfied by choosing the nonnegative integers and m appropriately. Thus we obtain the single integral equation

f

t

-

_J

(rt) (t)

dt 0 F(r)

f

t

-

_J%(rt)

w(t) (t)

dt, 0 < r < 0 G(r) r >

(2.10)

where

+v

2

+

-9

+

Y ₂ (2.11)

and

F(r)

and G(r) are given by the right sides in (2.7) and

(2.8)

respectively. Now in the single integral equation

(2.10)

we require % >

-I

for Hankel inver-sion. In general, this requirement may not be satisfied. To overcome this

difficulty, we use

(2.1)

(with p replaced by s and x replaced by r) in (2.10)

to obtain

f

t-Y+s

_J%+s

(rt) (t)

dt 0

Fl(r)-(-l)S

_of

t-Y+s

_J+s

(rt) w(t) (t)dt,

0 < r < G

(r)

r >

(2.12)

where F

_l(r)

(-l)S

rl+S

₍

r

s

[r

-)"

F(r)],

G

_l(r)

(-l)S

rl+S

₍

r

s

[r

-

G(r)],

(2.13)

and s is a nonnegative integer. Whatever be

a,B

we can always choose %+s > -I. Then the Hankel inversion can be invoked in

(2.12)

to give a Fredholm integral

equa-tion of second kind in

(x)

as

(x)

x

Y-s+l

o

r F

_{l(r) Jx+s(rX)dr}

+

f

r

_{l(r) J)t+s(rX)}

dr -y+s u

J%+s+l(u)

J%+s(X)

xJ%+s+

(x)

J%+s(U)

f

u

w(u) (u)

2 2 du

O U X

where

I,

are given in

(2.11)

and s is a nonnegative integer to be chosen appropriately so as to make

t+s

> -1.

3. SPECIAL CASES.

3.1.

,

-a/2,

B

0 where > max (-I,

a)

and 0

<

a

<

2. The dual integral equations are then

f

ta

[l+w(t)] (t)

_J(xt)

dt f(x), 0 < x <

I,

o

(xt)

dt

g(x)

x >

f

(t)

_J

o

(3.1)

These dual integral equations were considered by Noble

[2].

The conditions >

-I

and >

_-3/2

are satisfied automatically. Now from

(2.9)

a a

+

+

,,

n

=-I-

+

m

where E,m are nonnegative integers to be chosen such that > -I, >

-I.

We

a a

choose

O,

m so that

-I

+

>

-I

and

-

>-i. Also

a a

%

+

>-I,

y

-.

Hence we take s 0 in

(2.14).

Then a

f

2_u2)_l+a/2

21-a/2

-v-u

+I

(r

f(u) du,

Fl(r)

F(r)

_r

(a/2)

r _o

2a/2

_+a G

_l(r)

G(r)

r(l-a/2)

r

-I

d $

(u2_r2)-a/2

l-v

d-

u

g(u)

du, o so that the Fredholm integral equation in

(x)

is

(x)

(2x)

l-a/2

rl_V_a/2

(r

F(a/2)

f

fr

u+l

2

_u2)-l+a/2

o o

f(u)du}

_J+a/2(rx)

dr

l-a/2

_2a/2

_-+a/2

_d

2)-a/2

I-x

r

dr

f

(u2-r

u

g(u)

du

}"

J+a/2(rx)

dx

r(l-a/2)

r

x

l-a/2

I

u

a/2

o

u

_J+a/2+l

(u)

_{J+a/2 (x)-xJ+a/2+l (x)Jv+a/2}

(u)

w(u)

(u)

2 2 du

u x

(3.2)

This result was obtained in

[2]

(cf.eq.(4.7)).

Putting

g(x)

O, w(u)

O,

we obtain the explicit solution of the dual integral equations

f

ta

(t)

_J(xt)

dt f(x), 0 < x <

I,

o

f

(t)

_J(xt)

dt O, x >

o

where >

_max(-l,-a)

and 0 < a < 2, as

This coincides with the result given in

[7].

3.2.

I-81

<

+

and 0. This case was considered in

[I].

Then > -i, and we may mention here that the restriction 0 was not stated explicitly in

[I].

Unless this restriction holds good, it is not possible to use Hankel inversion with the order of the Bessel function as

.

To make >

-I,

n

>

-I,

we choose

I,

m in

(2.9).

Also as >

-I,

we choose s 0 in

(2.14)

so that the dual integral equations

(2.1)

in this case reduce to

#(x)

xY+l

E

2-Y+2s

-y+2-r

ug+l

2

_u2)-2

I

r

_Jl(rx)d

I

(r

f(u)

du}

r(-2+l)

o o

2-Y+28

_I

_r

A

J%(rx)d

I

(u2-r2)

Y-28

u

I-M

g(u)

du} F(Y-28+l) r u

_{Jl+l(U) Jl(x)}

_xJl+

_{l(x) J%(u)}

I

u-Y

w(u) #(u)

2 2 du o u x

(3.5)

This result does not seem to coincide with the general result given in

[I].

However, the known results for the following two special cases given in

[6]

as well as in

[I]

are deduced from

(3.5)

implying that

(3.5)

is also correct.

When 9 0,

1/2, 8

0,

(2.11)

gives

=-1/2

and y

.

Substitut-ing in

(3.5)

we obtain after simplification

r

#(x)

2__x

/ cos rx d

I

u f(u) o o

(r2_u2)

I/2

du

+ 2__x

I

COS rx

I

u

5(u)

o o

(u2_r

2)I/2

du dr x

f

w(u) #(u)

sln(u+x)

+

sln(u-x) du

w u

u+x

u-x

o

(3.6)

This result can also be deduced directly from

(2.14)

by choosing s 0, I, m 0. When O,

-I/2,

8 0, the integral equation for

#(x)

can similarly be deduced from

(3.5).

In this case

I/2,

y

-I/2.

After simplification we obtain r

#(x)

--2

I

sin rx

I

u

f(u)

o o

(r2_u

2)

1/2

du dr 2

I

sin rx

I

u

(u)

r

(u2_r

2)112

du

-1

f w(u) #(u)

sin(u-x)-

sin(u+x) dx

w _o u-x

u+x

(3.7)

(3.7)

can also be deduced directly from

(2.14)

by choosing s 0, O, m

I.

4. DISCUSSION.

complicated and lengthy compared to

Noble’s

multlplying-factor method based on

Sonlne’s

integrals as demonstrated in this note. Further, the method outlined here can be used

to attack the dual integral equations

(I.I)

for arbitrary values of the parameters 9,,=,8 and the weight function.

ACKNOWLEDGEMENT. This work was supported in part by the UGC project

No.F.8-8/85

(SR III).

REFERENCES

I.

NASIM, C. On dual integral equations with Hankel kernel and arbitrary weight function, Internat. J. Math. and Math. Sci. 9 (1986), 293-300.

2. NOBLE, D. The solution of Bessel function dual integral equations by a multiply-ing-factor method, Proc. Camb. Phil. Soc. 59 (1963), 351-362.

3.

SCHMELTZER,

R.A. and LEWIN, M. Function-theoretic solution to a class of dual integral equations and application to diffraction theory,

Quart.

Appl. Math. 21 (1964), 259-283.

4. ROSS, R.A.

An

integral equation arising in diffraction theory, Wave phenomenon: modern theory and application, North Holland Math. Stud. 97 (1983), 453-459. 5. ROSE, R.F. and

DE

HOOG, F.R. Exact solutions of certain dual integral equations and their asymptotic properties,

q.

Jl. Mech.

_AppI.

Math. 36

(1983),

420-436. 6.

SNEDDON, I.N.

Mixed

_bounar

value

_problems

in

potential theory,

John Wiley and

Sons, New York, 1966.

7. TITCHMSH, E.C. Introduction to the

theory

of Fourier integrals, Oxford

Special Issue on

Intelligent Computational Methods for

Financial Engineering

Call for Papers

As a multidisciplinary field, financial engineering is

becom-ing increasbecom-ingly important in today’s economic and financial

world, especially in areas such as portfolio management,

as-set valuation and prediction, fraud detection, and credit risk

management. For example, in a credit risk context, the

re-cently approved Basel II guidelines advise financial

institu-tions to build comprehensible credit risk models in order

to optimize their capital allocation policy. Computational

methods are being intensively studied and applied to

im-prove the quality of the financial decisions that need to be

made. Until now, computational methods and models are

central to the analysis of economic and financial decisions.

However, more and more researchers have found that the

financial environment is not ruled by mathematical

distribu-tions or statistical models. In such situadistribu-tions, some attempts

have also been made to develop financial engineering

mod-els using intelligent computing approaches. For example, an

artificial neural network (ANN) is a nonparametric

estima-tion technique which does not make any distribuestima-tional

as-sumptions regarding the underlying asset. Instead, ANN

ap-proach develops a model using sets of unknown parameters

and lets the optimization routine seek the best fitting

pa-rameters to obtain the desired results. The main aim of this

special issue is not to merely illustrate the superior

perfor-mance of a new intelligent computational method, but also

to demonstrate how it can be used e

ffectively in a financial

engineering environment to improve and facilitate financial

decision making. In this sense, the submissions should

es-pecially address how the results of estimated computational

models (e.g., ANN, support vector machines, evolutionary

algorithm, and fuzzy models) can be used to develop

intelli-gent, easy-to-use, and/or comprehensible computational

sys-tems (e.g., decision support syssys-tems, agent-based system, and

web-based systems)

This special issue will include (but not be limited to) the

following topics:

•

Computational methods

: artificial intelligence,

neu-ral networks, evolutionary algorithms, fuzzy inference,

hybrid learning, ensemble learning, cooperative

learn-ing, multiagent learning

•

Application fields

: asset valuation and prediction,

as-set allocation and portfolio selection, bankruptcy

pre-diction, fraud detection, credit risk management

•

Implementation aspects

: decision support systems,

expert systems, information systems, intelligent

agents, web service, monitoring, deployment,

imple-mentation

Authors should follow the Journal of Applied

Mathemat-ics and Decision Sciences manuscript format described at

the journal site

http://www.hindawi.com/journals/jamds/

.

Prospective authors should submit an electronic copy of their

complete manuscript through the journal Manuscript

Track-ing System at

http://mts.hindawi.com/

, according to the

fol-lowing timetable:

Manuscript Due

December 1, 2008

First Round of Reviews

March 1, 2009

Publication Date

June 1, 2009

Guest Editors

Lean Yu, Academy of Mathematics and Systems Science,

Chinese Academy of Sciences, Beijing 100190, China;

Department of Management Sciences, City University of

Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong;

[email protected]

Shouyang Wang, Academy of Mathematics and Systems

Science, Chinese Academy of Sciences, Beijing 100190,

China; [email protected]

K. K. Lai, Department of Management Sciences, City

University of Hong Kong, Tat Chee Avenue, Kowloon,

Hong Kong; [email protected]