College of Engineering and Computer Science

Mechanical Engineering Department

Engineering Analysis Notes
Last updated: July 7, 2004 Larry Caretto
Introduction to Orthogonal Functions and Eigenfunction Expansions
Goal of these notes
Function sets can form vector spaces and the notions of vectors and matrix operations
orthoonality, !asis sets, eienvalues, can !e carried over into analysis of functions that are
important in enineerin applications" #$n dealin %ith functions %e have eienfunctions in place
of eienvectors"& 'hese notes lin( the previous notes on vector spaces to the application to
functions"
'he eienfunctions %ith %hich %e %ill !e dealin are solutions to differential e)uations"
*ifferential e)uations, !oth ordinary and partial differential e)uations, are an important part of
enineerin analysis and play a ma+or role in the ,0-./ course" 0tudents are expected to have
a !ac(round in ordinary differential e)uations so that %e %ill !ein %ith a !rief revie% of this
topic" 1e %ill then discuss po%er series solutions to differential e)uations and apply this
techni)ue to /essel2s differential e)uation" 'he series solutions to this e)uation, (no%n as
/essel functions, usually occur in cylindrical eometries in the solution to the same pro!lems that
produce sines and cosines in rectanular eometries"
1e %ill see that /essel functions, li(e sines and cosines, form a complete set so that any function
can !e represented as an infinite series of these functions" 1e %ill discuss the 0turm3Loiuville
e)uation, %hich is a eneral approach to eienfunction expansions, and sho% that sines, cosines,
and /essel functions are special examples of functions that satisfy the 0turm3Liouville e)uation"
'he /essel functions are +ust one example of special functions that arise as solutions to ordinary
differential e)uations" .lthouh these special functions are less %ell (no%n than sines and
cosines, the idea that these special functions !ehave in a similar eneral manner to sines and
cosines in the solution of enineerin analysis pro!lems, is a useful concept in applyin these
functions %hen the pro!lem you are solvin re)uires their use"
'hese notes !ein !y revie%in some concepts of differential e)uations !efore discussin po%er
series solutions and Fro!enius method for po%er series solutions of differential e)uations" 1e
%ill then discuss the solution of /essel2s e)uation as an example of Fro!enius method" Finally
%e %ill discuss the 0turm3Liouville pro!lem and a eneral approach to special functions that form
complete sets"
What is a differential equation?
. differential e)uation is an e)uation, %hich contains a derivative" 'he simplest (ind of a
differential e)uation is sho%n !elo%:
0 0
) ( x x at y y with x f
dx
dy
4-5
$n eneral, differential e)uations have an infinite num!er of solutions" $n order to o!tain a uni)ue
solution one or more initial conditions #or !oundary conditions& must !e specified" $n the a!ove
example, the statement that y 6 y0 at x 6 x0 is an initial condition" #'he difference !et%een initial
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
and !oundary conditions, %hich is really one of namin, is discussed !elo%"& 'he differential
e)uation in 4-5 can !e ?solved@ as a definite interal"

x
x
dx x f y y
0
) ( -
0
425
'he definite interal can !e either found from a ta!le of interals or solved numerically,
dependin on f#x&" 'he initial #or !oundary& condition #y 6 y0 at x 6 x0& enters the solution directly"
Chanes in the values of y0 or x0 affect the ultimate solution for y"
. simple chane ma(in the riht hand side a function of x and y, f#x,y&, instead of a function of
x alone ives a much more complicated pro!lem"
0 0
) , ( x x at y y with y x f
dx
dy
495
1e can formally %rite the solution to this e)uation +ust as %e %rote e)uation 425 for the solution to
e)uation 4-5"

x
x
dx y x f y y
0
) , ( -
0
445
Aere the definite interal can no loner !e evaluated simply" 'hus, alternative approaches are
needed" 7)uation 445 is used in the derivation of some numerical alorithms" 'he #un(no%n&
exact value of f#x,y& is replaced !y an interpolation polynomial %hich is a function of x only"
$n the theory of differential e)uations, several approaches are used to provide analytical solutions
to the differential e)uations" 8eardless of the approach used, one can al%ays chec( to see a
proposed solution is correct !y su!stitutin a proposed solution into the oriinal differential
e)uation and determinin if the solution satisfies the initial or !oundary conditions"
Ordinary differential equations involve functions, %hich have only one independent varia!le"
'hus they contain only ordinary derivatives" Partial differential equations involve functions %ith
more than one independent varia!le" 'hus, they contain partial derivatives" 'he a!!reviations
B*7 and ;*7 are used for ordinary and partial differential e)uations, respectively"
$n an ordinary differential e)uation %e are usually tryin to solve for a function, y#x&, %here the
e)uation involves derivatives of y %ith respect to x" 1e call y the dependent variale and x the
independent variale" #Bther varia!les are conventionally used as the independent and
dependent varia!lesC often t is used as the independent varia!le and x is used as the dependent
varia!le" $n any physical pro!lem, the appropriate varia!le for the physical )uantities under
consideration are used"&
'he order of the differential equation is the order of the hihest derivative in the e)uation"
7)uations 4-5 and 495 are first3order differential e)uations" . differential e)uation %ith first, second
and third order derivatives only %ould !e a third order differential e)uation"
$n a linear differential equation, the terms involvin the dependent varia!le and its derivatives
are all linear terms" 'he independent varia!le may have nonlinear terms" 'hus x
9
d
2
yDdx
2
E y 6 0
is a linear, second3order differential e)uation" ydyDdx E sin#y& 6 0 is a nonlinear first3order
differential e)uation" #7ither term in this e)uation ydyDdx or sin#y& %ould ma(e the differential
e)uation nonlinear"&
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
*ifferential e)uations need to !e accompanied !y initial or oundary conditions" .n n
th
order
differential e)uation must have n initial #or !oundary& conditions in order to have a uni)ue
solution" .lthouh initial and !oundary conditions !oth mean the similar thins, the term ?initial
conditions@ is usually used %hen all the conditions are specified at one initial point" 'he term
?!oundary conditions@ is used %hen the conditions are specified at t%o different values of the
independent varia!le" For example, in a second order differential e)uation for y#x&, the
specification that y#0& 6 a and y2#0& 6 !, %ould !e called t%o initial conditions" 'he specification
that y#0& 6 c and y#L& 6 d, %ould !e called !oundary conditions" 'he initial or !oundary conditions
can involve a value of the varia!le itself, lo%er3order derivatives of the varia!le, or e)uations
containin !oth values of the dependent varia!le and its lo%er3order derivatives"
Some simple ordinary differential equations
From previous courses, you should !e familiar %ith the follo%in differential e)uations and their
solutions" $f you are not sure a!out the solutions, +ust su!stitute them into the oriinal differential
e)uation"
) (
0 0 0
0
t t k
e y y t t at y y with ky
dt
dy

4,5
) cos( ) sin(
2
2
2
kx B kx A y y k
dx
y d
+ 4=5
kx kx
e B e A kx B kx A y y k
dx
y d

+ + ' ' ) cosh( ) sinh(
2
2
2
475
$n e)uations 4=5 and 475 the constants . and / #or .2 and /2& are determined !y the initial or
!oundary conditions" Fote that %e have used t as the independent varia!le in e)uation 4,5 and x
as the independent varia!le in e)uations 4=5 and 475"
'here are four possi!le functions that can !e a solution to e)uation 4=5: sin#(x&, cos#(x&, e
i(x
, and
e
3i(x
, %here i
2
6 3-" 0imilarly there are four possi!le functions that can !e a solution to e)uation
475: sinh#(x&, cosh#(x&, e
(x
, and e
3(x
" $n each of these cases the four possi!le solutions are not
linearly independent"
-
'he minimum num!er of functions %ith %hich all solutions to the
differential e)uation can !e expressed is called a !asis set for the solutions" 'he solutions sho%n
a!ove for e)uations 4=5 and 475 are !asis sets for the solutions to those e)uations"
Bne final solution that is useful is the solution to eneral linear first3order differential e)uation"
'his e)uation can !e %ritten as follo%s"
) ( ) ( x g y x f
dx
dy
+ 4<5
'his e)uation has the follo%in solution, %here the constant, C, is determined from the initial
condition"
1
]
1

,
_

dx e x g C e y
dx x f dx x f ) ( ) (
) (
4G5
1
1e have the follo%in e)uations amon these various functions:
2
) sinh(
x x
e e
x

2
) cosh(
x x
e e
x

+

2
) sin(
ix ix
e e
x

2
) cos(
ix ix
e e
x

+

7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
Po!er series solutions of ordinary differential equations
'he solution to e)uation 4=5 is composed of a sine and a cosine term" $f %e consider the po%er
series for each of these, %e see that the solution is e)uivalent to the follo%in po%er series"

+
+

,
_

+ + +

,
_

+ +
0
2
0
1 2 6 4 2 7 5 3
)! 2 (
) 1 (
)! 1 2 (
) 1 (
! 6 ! 4 ! 2
1
! 7 ! 5 ! 3
n
n n
n
n n
n
x
B
n
x
A
x x x
B
x x x
x A y
4-05
1e are interested in seein if %e can o!tain such a solution directly from the differential e)uation"
. proof !eyond the level of this course can !e used to sho% that the follo%in differential
e)uation ) ( ) (
) (
) (
) (
2
2
x r y x q
dx
x dy
x p
dx
x y d
+ + has po%er series solutions, y#x& in a reion, 8,
around x 6 a, provided that p#x&, )#x& and r#x& can !e expressed in a po%er series in some reion
a!out x 6 a" Functions that can !e represented as a po%er series are called analytic functions"
2
'he po%er series solution of ) ( ) ( ) (
2
2
x r y x q
dx
dy
x p
dx
y d
+ + re)uires that the three functions
p#x&, )#x& and r#x& !e represented as po%er series" 'hen %e assume a solution of the follo%in
form"

0
0
) - ( ) (
n
n
n
x x a x y 4--5
Aere the an are un(no%n coefficients" 1e can differentiate this series t%ice to o!tain"

0
2
0
2
2
0
1
0
) - ( ) 1 ( ) - (
n
n
n
n
n
n
x x a n n
dx
y d
and x x na
dx
dy
4-25
0u!stitutin e)uations 4--5 and 4-25 into our oriinal differential e)uation ives the follo%in
result"
) ( ) - ( ) ( ) - ( ) ( ) - ( ) 1 (
0
0
0
1
0
0
2
0
x r x x a x q x x na x p x x a n n
n
n
n
n
n
n
n
n
n
+ +

4-95
1e then set the coefficients of each po%er of x on !oth sides of the e)uation to !e e)ual to each
other" 'his ives an e)uation that %e can used to solve for the un(no%n an coefficients in terms
of one or more coefficients li(e a0 and a-, %hich are used to determine the initial conditions" 'his
is !est illustrated !y usin e)uation 4=5, 0
2
2
2
+ y k
dx
y d
, as an example" Aere %e have p#x& 6 0,
)#x& 6 (
2
, and r#x& 6 0, so for this example, e)uation 4-95 !ecomes"
0 ) - ( ) - ( ) 1 (
0
0
2
0
2
0
+

n
n
n
n
n
n
x x a k x x a n n 4-45
'he only %ay to assure that e)uation 4-45 is satisfied is to have the coefficient of each po%er of #x
x0& vanish" 0o %e et the po%er series solution !y settin the coefficients of each po%er of #x
2
0ee the !rief discussion of po%er series in .ppendix . for more !asic information on po%er
series"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
x0& e)ual to Hero" 'his can !e done if the po%er of #x x0& is the same in !oth sums so that %e
can %rite the t%o suns as a sinle sum" 'o do this %e use the follo%in process:
-" 0u!stitute the index m 6 n 2 in the first sum" 'his %ill eventually ive the same po%er of #x
x0& in !oth sums" $n order to do this %e su!stitute n 6 m E 2 in the sum terms and replace
the lo%er limit of n 6 0 !y m 6 0 2 6 2" 'hese t%o steps ive the follo%in result"
0 ) - ( ) - ( ) 1 )( 2 (
0
0
2
2
0 2
+ + +

n
n
n
m
m
m
x x a k x x a m m 4-4a5
2" 'o have the same sums, %e have to have the same lo%er limit" 1e can evaluate the sums
for m 6 32 and m 6 3- separately and then have the sum over m runnin from Hero to infinity"
'his tas( is easy here since !oth the m 6 32 and m 6 3- term are Hero"
0 ) - ( ) - ( ) 1 )( 2 ( 0 0
0
0
2
0
0 2
+ + + + +

n
n
n
m
m
m
x x a k x x a m m 4-4!5
9" 0ince !oth m and n are dummy indices, %e can replace either of these !y any other letter and
have the same result for the sum" .lthouh %e used m 6 n 2 to rearrane the first sum, %e
can no% replace the m index in the first sum !y n to match the second sum" 'his ives the
follo%in result"
0 ) - ( ) - ( ) 1 )( 2 (
0
0
2
0
0 2
+ + +

n
n
n
n
n
n
x x a k x x a n n 4-4c5
4" 1e can no% com!ine the t%o sums !ecause !oth have the same upper and lo%er limits and
!oth have the same po%er of x x0" 'he result of these four steps allo%s us to re%rite
e)uation 4-45 as sho%n !elo%"
[ ] 0 ) - ( ) 1 )( 2 (
) - ( ) 1 ( ) - ( ) - ( ) 1 (
0
0
2
2
2
2
0
0
0
2
0
2
0
+ + +
+

n
n
n n
n
n
n
n
n
n
n
n
n
x x a k a n n
x x a n n x x a k x x a n n
4-,5
'he last sum in e)uation 4-,5 e)uals Hero only if the coefficient of #x x0&
n
vanishes for each n"
'his ives the follo%in relationship amon the un(no%n coefficients"
) 1 )( 2 (
0 ) 1 )( 2 (
2
2
2
2
+ +
+ + +
+ +
n n
a k
a or a k a n n
n
n n n
4-=5
'his ives us an e)uation for anE2 in terms of the coefficient previously found for an" 1e cannot
use this e)uation to find a0 or a-, so %e assume that these coefficients %ill !e determined !y the
initial conditions" Ao%ever, once %e (no% a0, %e see that %e can find all the even num!ered
coefficients as follo%s
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
2 3 4 ) 1 2 )( 2 2 (
2
) 1 2 )( 2 2 (
2 ) 1 0 )( 2 0 (
0
4
0
2
2
2
2
4
0
2
0
2
2

+ +

,
_

+ +


+ +

a k
a k
k
a k
a
a k a k
a
4-75
Continuin in this fashion %e see that the eneral pattern for a su!script %hose coefficient is an
even num!er is the follo%in"
even n
n
a k
a
n
n
n
)! (
) 1 (
0
2

4-<5
1e can verify this eneral result !y o!tainin an e)uation for anE2" 'his is done !y replacin n in
e)uation 4-<5 !y nE2 to ive
)! 2 (
) 1 (
0
2
2
) 2 (
2
+

+
+
+
n
a k
a
n
n
n
" Fext %e su!stitute this e)uation and
e)uation 4-<5 into e)uation 4-=5 to see if %e et the correct result for the ratio anE2Dan"
9
) 1 )( 2 ( )! 2 (
!
!
) 1 (
)! 2 (
) 1 (
) 1 )( 2 (
2 2
0
2
0
2
2
) 2 (
2
2
+ +

+

+ +

+
+
+
n n
k
n
n k
n
a k
n
a k
a
a
n n
k
n
n
n
n
n
n
4-G5
1e see that the ratio anE2Dan that %e computed usin our eneral e)uation for an from e)uation
4-<5 is the same as the value for this ratio that %e started %ith in e)uation 4-=5" 1e thus conclude
that e)uation 4-<5 ives us a correct solution for an %hen n is even" 1e can handle the
recurrence for an, %hen n is an odd num!er in the same %ay that %e +ust did for even n" 1e start
!y findin a9 and a, in terms of a-"
2 3 4 5 4 5
2 3
) 1 3 )( 2 3 (
2 3 ) 1 1 )( 2 1 (
1
4
1
2
2
3
2
5
1
2
1
2
3

,
_

+ +

+ +

a k
a k
k
a k
a
a k a k
a
4205
1e see that this recurrence %ill lead to a eneral e)uation of the follo%in form"
odd n
n
a k
a
n
n
n
)! (
) 1 (
1
1
2
) 1 (

42-5
3
$f you are not familiar %ith the cancellation of factorials, see the discussion in .ppendix /"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
.s !efore, %e can chec( this eneral relationship !y o!tainin an expression for anE2 and sho%in
that the ratio anE2Dan as computed from e)uation 42-5 satisfies e)uation 4-=5" 'his chec( is left as
an exercise for the reader"
Fo% that %e have expressions for an in terms of the initial values a0 and a- %e can su!stitute
these expressions #in e)uations 4-<5 and 42-5& into our proposed eneral po%er series solution for
our differential e)uation from e)uation 4--5"
( ) [ ] ( ) [ ]
( )
( ) [ ] ( ) [ ]
1
]
1

+ + 1
]
1

! 5
-
! 3
-
-
! 4
-
! 2
-
1
) - (
)! (
) 1 (
) - (
)! (
) 1 (
) - ( ) - ( ) - ( ) (
5
0
3
0
0 1
4
0
2
0
0
1
0
1
1 2
) 1 (
0
0
0
2
1
0
0
0
0
0
x x k x x k
x x k a
x x k x x k
a
x x
n
a k
x x
n
a k
x x a x x a x x a x y
n odd
n
n
n
n
n even
n
n
n
n
n odd
n
n
n
n even
n
n
n
n
n
n
4225
'hus the series multiplied !y a0 and a- are seen to !e the series for cos4(#x x0&5 and sin4(#x
x0&5, respectively" 'his is the expected solution of the differential e)uation 0
2
2
2
+ y k
dx
y d
"
.lthouh this differential e)uation has a solution in terms of sines and cosines, the !asic po%er
series methods can !e used for e)uations that do not have a conventional solution"
Summary of po!er series solutions of ordinary differential equations
1e can solve a differential e)uation li(e ) ( ) (
) (
) (
) (
2
2
x r y x q
dx
x dy
x p
dx
x y d
+ + , usin the po%er
series method, provided that p#x&, )#x& and r#x& are analytic at a point x0 %here %e %ant the
solution" 0uch a solution is o!tained in the follo%in steps"
1rite the solution for y#x& as a po%er series in un(no%n coefficients an as sho%n in
e)uation 4--5"
*ifferentiate the po%er series t%o times to et the derivatives re)uired in the differential
e)uationC see e)uation 4-25 for the results of this differentiation"
B!tain po%er series expansions for p#x&, )#x& and r#x&, if these are not constants or
simply polynomials"
0u!stitute the po%er series for y#x&, y2#x&, y@#x&, p#x&, )#x& and r#x& into the differential
e)uation for the pro!lem"
8e%rite the resultin e)uation to roup terms %ith common po%ers of x x0"
0et the coefficients of each po%er of x x0 e)ual to Hero" 'his should produce an
e)uation that relates neih!orin values of the un(no%n coefficients an"
Ise the e)uation found in the previous step to relate coefficients %ith hiher su!scripts to
those %ith lo%er su!scripts" 'he first fe% coefficients, e"", a0, a-, etc", %ill not !e (no%n"
#'hese %ill !e determined !y the initial conditions on the differential e)uation"&
7xamine the e)uation relatin the coefficients and try to o!tain a eneral e)uation for
each an in terms of the un(no%n coefficients a0, a-, etc"
0u!stitute the eneral expression for an into the oriinal po%er series for y#x&" 'his is the
final po%er series solution"
"roenius method for solution of ordinary differential equations
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
'he Fro!enius method is used to solve the follo%in differential e)uation"
0
) ( ) ( ) ( ) (
2 2
2
+ + y
x
x c
dx
x dy
x
x b
dx
x y d
4295
$n this e)uation !#x& and c#x& are analytic at x 6 0" Fote that the conventional po%er series
method cannot !e used for this e)uation !ecause the coefficients of dyDdx and y are not analytic
at x 6 0"
$n the Fro!enius method the solution is %ritten as follo%s"

0 0
) (
n
r n
n
n
n
n
r
x a x a x x y 4245
Aere the an are un(no%n coefficients and the value of r is also un(no%n" 'he value of r is
determined durin the solution procedure so that a0 0" 1e can differentiate the series for y#x& in
e)uation 4245 t%o times to o!tain"

+
+ + +
0
2
2
2
0
1
) 1 )( ( ) (
n
r n
n
n
r n
n
x a r n r n
dx
y d
and x a r n
dx
dy
42,5
1e assume that !#x& and c#x& are constants, simple polynomials or series expressions that have
the follo%in eneral forms"

0 0
) ( ) (
n
n
n
n
n
n
x c x c x b x b 42=5
1e can su!stitute e)uations 4245, 42,5, and 42=5, into e)uation 4295 and multiply the result !y x
2
to
o!tain the follo%in e)uation"
0 ) (
) 1 )( (
0 0 0
1
0
0
2 2

,
_

,
_

,
_

,
_

+ + +

+
n
r n
n
n
n
n
n
r n
n
n
n
n
n
r n
n
x a x c x a r n x b x
x a r n r n x
4275
1e can com!ine the x
2
and x terms outside the sums %ith the x terms inside the sums as follo%s"
0 ) (
) 1 )( (
0 0 0 0
0

,
_

,
_

,
_

,
_

+ + +

+
n
r n
n
n
n
n
n
r n
n
n
n
n
n
r n
n
x a x c x a r n x b
x a r n r n
42<5
1e can then %rite out the first fe% terms in each series or product of series to ive
[ ] [ ]
[ ] [ ] 0
) 2 ( ) 1 ( ) 1 (
) 1 )( 2 ( ) 1 ( ) 1 (
2
2 0 1 1 0 2
1
1 0 0 1 0 0
2
2 0 1 1 0 2
1
1 0 0 1 0 0
2
2
1
1 0
+ + + + + + +
+ + + + + + + + + +
+ + + + + +
+ +
+ +
+ +

r r r
r r r
r r r
x a c a c a c x a c a c x a c
x a r b a r b ra b x a r b ra b x ra b
x a r r x ra r x a r r
42G5
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
.s in the conventional po%er series solution, %e re)uire the coefficient of each term in the po%er
series to vanish to satisfy e)uation 42<5 or 42G5" 0tartin %ith the lo%est po%er of x, x
r
, %e re)uire
that the follo%in coefficient !e Hero"
[ ] 0 ) 1 (
0 0 0
+ + a c r b r r 4905
$n the Fro!enius method, %e %ant to (eep a0 0, thus %e re)uire that
0 ) 1 (
0 0
+ + c r b r r 49-5
'his yields a )uadratic e)uation in r %hich is called the indicial equation" 1e find t%o possi!le
values of r from the conventional solution of the )uadratic e)uation"
( )
2
4 1 1
0
2
0 0
c b b
r
t

4925
$n the Fro!enius method, the oriinal solution in e)uation 4245, %ith t%o different values of r, can
provide t%o different solutions if the t%o values of r found from e)uation 4295 are different and
their difference is not an inteer" 'hese t%o solutions form a !asis for all solutions to e)uation
4295" 'hese are

0
2
0
1
2 1
) ( ) (
n
n
n
r
n
n
n
r
x A x x y and x a x x y
4995
'he coefficients for the t%o solutions, an for y-#x& and .n for y2#x&, are different" 'hese coefficients
are found in the same %ay that the coefficients %ere found in the usual po%er series e)uation,
once the values of r are determined" $f there is a dou!le root for r or if the t%o values of r differ !y
an inteer, it is necessary to have a second solution that has a different form" For a dou!le root, r
6 r- 6 r2, %e have the t%o follo%in solutions"

+
1
1 2
0
1
) ln( ) ( ) ( ) (
n
n
n
n
n
n
r
x A x x y x y and x a x x y 4945
$f the t%o roots, r- and r2 differ !y an inteer, the t%o possi!le solutions are %ritten as follo%s"

+
0
1 2
0
1
2 1
) ln( ) ( ) ( ) (
n
n
n
r
n
n
n
r
x A x x x ky x y and x a x x y
49,5
$n the last expression the roots are defined such that r- J r2C the value of ( may !e Hero in this
case" Fote that the an coefficients and the .n coefficients are differentC also for the dou!le root,
the summation for y2#x& starts at n 6 - instead of n 6 0"
$n the next section %e examine the solution of /essel2s e)uation usin the Fro!enius method"
'his serves !oth as the derivation of the series expansions for the various /essel functions and
as an illustration of the Fro!enius method"
#pplication of "roenius Method to $essel%s equation
/essel2s e)uation arises %hen various enineerin phenomena are modeled in cylindrical
coordinates" 'his e)uation for y#x& %ith as a (no%n parameter in the e)uation has the follo%in
form"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
0
) ( 1 ) (
2
2 2
2
2

+ + y
x
x
dx
x dy
x dx
x y d
49=5
1e see that this is the eneral form iven in e)uation 4295 %ith !#x& 6 - and c#x& 6 x
2

2
" 8ecall
the !asic solution format for Fro!enius method from e)uation 4245, copied !elo%"

0 0
) (
n
r n
n
n
n
n
r
x a x a x x y 4245
'he derivatives of this solution %ere iven in e)uation 42,5, also copied !elo%"

+
+ + +
0
2
2
2
0
1
) 1 )( ( ) (
n
r n
n
n
r n
n
x a r n r n
dx
y d
and x a r n
dx
dy
42,5
$f %e multiply e)uation 49=5 !y x
2
and su!stitute e)uations 4245 and 42,5 %e o!tain the follo%in
result"
( ) 0 ) ( ) 1 )( (
0
2 2
0 0
+ + + + +

+
n
r n
n
n
r n
n
n
r n
n
x a x x a r n x a r n r n 4975
7xcept for the final sum that is multiplied !y x
2
, %e can com!ine all three summation operators
into a sinle sum"
[ ] 0 ) ( ) 1 )( (
0
2
0
2
+ + + + +

+ +

+
n
r n
n
n
r n
n
x a x a r n r n r n 49<5
1e can simplify the coefficient in the first summation"
2 2 2 2 2
) ( ) ( ) ( ) ( ) ( ) 1 )( ( + + + + + + + + + r n r n r n r n r n r n r n 49G5
1e can follo% the process outlined in the po%er series solutions of differential e)uations to
com!ine the t%o sums in e)uation 49<5" First, %e define a ne% index for the second sum, m 6
nE2" 'his allo%s us to %rite the second sum as follo%s"

+ +

2
2
0
2
m
r m
m
n
r n
n
x a x a 4405
.lthouh %e derived e)uation 4405 !y settin n 6 m 2, %e reconiHe that m is a dummy index"
'hus %e can replace m !y any other letter, includin n" $f %e do this, %e can use e)uations 49G5
and 4405 to re%rite e)uation 49<5 as follo%s"
[ ]
[ ]
( ) ( ) [ ] [ ] { } 0 ) ( 1
) (
) ( ) 1 )( (
2
2
2 2 1
1
2 2
0
2 2
2
2
0
2 2
0
2
0
2
+ + + + +
+ +
+ + + + +

+ +

+
n
r n
n n
r r
n
r n
n
n
r n
n
n
r n
n
n
r n
n
x a a r n x a r x a r
x a x a r n
x a x a r n r n r n
44-5
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
'he indicial e)uation is the e)uation that sets the coefficient of a0 e)ual to Hero" For /essel2s
e)uation the indicial e)uation is r
2

2
6 0, %hose solution is r 6 t" 'he first Fro!enius method
solution %ill !e the same reardless of the value of " Ao%ever, if 0, this is a dou!le root and
if is an inteer, the t%o roots of the indicial e)uation %ill differ !y an inteer"
4
'he first solution
%ill !e the same in all three cases" 1e %ill fist o!tain this solution and defer the consideration of
the second solution" 1e %ill ta(e the root r 6 , that ives a positive value for r" 1e %ill assume
that is positive" 0u!stitutin r 6 , into the po%er series solution in e)uation 44-5 ives
( ) [ ] [ ] { } 0 ) ( 1
2
2
2 2 1
1
2 2
+ + + +

+
n
n
n n
x a a n x a
4425
7ach coefficient in this e)uation #for each po%er of x& must e)ual Hero for series sum on the left3
hand side of e)uation 4425 to !e Hero" 'he coefficient of the x
-E
term %ill not vanish unless a- 6 0"
'hus %e conclude that a- is Hero" 'he remainin po%ers of x in e)uation 4425 have a common
e)uation for the coefficient" 1hen %e set this e)uation to Hero, %e o!tain the follo%in result"
[ ] [ ] ( ) 0 2 2 ) (
2 2
2 2 2
2
2 2
+ + + + + + +
n n n n n n
a a n n a a n n a a n 4495
1e can solve this e)uation to o!tain a recurrence relationship that ives an in terms of an32"
) 2 (
2
+

n n
a
a
n
n
4445
.ccordin to e)uation 4445, if an 6 0, then anE2 6 0" 0ince %e have sho%n that a- must !e Hero, %e
conclude that all values of an %ith an odd su!script must !e Hero" 0ince the further application of
e)uation 4445 %ill !e for even su!scripts only, %e chane the index from n to m %here n 6 2m"
'his allo%s us to re%rite e)uation 4445 as follo%s"
) ( 2 ) 2 2 ( 2
2
2 2 2 2
2
+

m m
a
m m
a
a
m m
m
44,5
1e %ill ta(e a0 as an un(no%n coefficient that is determined !y the initial conditions on the
differential e)uation" 1e can then use e)uation 44,5 to %rite the first fe% #even3num!ered&
coefficients in terms of a0" 0ettin m 6 - in e)uation 44,5 ives us a2"
) 1 ( 2 ) 1 )( 1 ( 2
2
0
2
2 ) 1 ( 2
2
+

a
a
a 44=5
Fext %e set m 6 2 in e)uation 44,5 to o!tain a4 in terms of a2, and then use e)uation 44=5 to et a4
in terms of a0"
) 1 )( 2 )( 2 ( 2 ) 2 )( 2 ( 2
) 1 ( 2
) 2 )( 2 ( 2 ) 2 )( 2 ( 2
4
0
2
2
0
2
2
2
2 ) 2 ( 2
4
+ +

,
_

a
a
a
a
a
4475
4
8emem!er that is (no%n from the oriinal differential e)uation" $t is typically a mathematical or physical
parameter from the oriinal pro!lem statement"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
Fext %e use m 6 9 in e)uation 44,5 to compute a= !y %hich point %e should !e a!le to see the
pattern in the a2m coefficients"
) 1 )( 2 )( 3 )( 2 )( 3 ( 2 ) 3 )( 3 ( 2
) 1 )( 2 )( 2 ( 2
) 3 )( 3 ( 2 ) 3 )( 3 ( 2
6
0
2
4
0
2
4
2
2 ) 3 ( 2
6
+ + +

,
_

+ +

a
a
a
a
a
44<5
1e see a pattern emerin that %e can summariHe in the follo%in eneral e)uation for the
coefficients in the po%er series solution"
) 1 )( 2 ( ) 1 )( ( ! 2
) 1 (
2
0
2
+ + + +

m m m
a
a
m
m
m
44G5
'his eneral form can !e verified !y constructin the ratio a2mDa2m32 usin e)uation 44G5 and
demonstratin that that ratio is the same as the one in e)uation 44,5" Fo% that %e have the
eneral relationship for the coefficients in the first solution, %e have to loo( at individual cases of
noninteer , inteer , and 6 0"
First Frobenius-method solution for integer 1e first consider the case of inteer , for
%hich the parameter, , is conventionally represented as n" #'he sym!ol is reserved for cases
%here this parameter is not an inteer"& 0ettin n 6 for cases %here n is an inteer ives the
follo%in version of e)uation 44G5"
) 1 )( 2 ( ) 1 )( ( ! 2
) 1 (
2
0
2
n n n m n m m
a
a
m
m
m
+ + + +

4,05
0ince a0 is an un(no%n constant, %hich has different values for different pro!lems as determined
!y the !oundary conditions for the pro!lem, %e can redefine a0 as follo%s" #'his is a convention
used to o!tain an e)uation that is used for computation and ta!ulation of /essel functions"&
! 2
1
0
n
A a
n
44G5
Fo%, . is the constant that is selected to fit the !oundary conditions" 1ith this su!stitution %e
can %rite e)uation 44<5 as follo%s"
)! ( ! 2
) 1 (
! ) 1 )( 2 ( ) 1 )( ( ! 2
) 1 (
2 2
2
n m m
A
n n n n m n m m
A
a
n m
m
n m
m
m
+

+ + + +

+ +

4,05
1e can su!stitute this expression for a2m into the po%er series e)uation 4245 proposed for the
solution %ith the value of r in that e)uation set to the solution of the indicial e)uation, r 6 , usin
n in place of for inteer values of " 1e %ill su!se)uently drop the . coefficient %ith the
understandin that any final solution can !e multiplied !y a constant to satisfy the !oundary
conditions" #'his is similar to the practice done in eneratin ta!les of sines and cosines"&

0 0
2
2
2
2
0
)! ( ! 2
) 1 (
) (
m m
n m
m m
n m
m
n
k
k
k
r
n m m
x
x x a x x a x x y 4,-5
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
'he $essel function of the first &ind of integer order n' (n)*+ is defined !y this e)uation, %ith
the ar!itrary constant, ., omitted" 'hus, %e %rite

+
+

0
2
2
)! ( ! 2
) 1 (
) (
m
n m
m m
n
n
n m m
x
x x J 4,25
;lots of these /essel functions for some lo% values of n are sho%n !elo%" Fote that J0#0& 6 -
%hile Jn#0& 6 0 for all n J 0"
$essel "unctions of the "irst ,ind for -nteger Orders
./01
./02
./03
./04
/
/04
/03
/02
/01
/05
/06
/07
/08
/09
4
/ 4 3 2 1 5 6 7 8 9 4/ 44 43 42 41 45
*
(
n
)
*
+
n : /
n : 4
n : 1
First Frobenius method solution for noninteger 'he only chane necessary %hen %e
consider noninteer , is that %e do not have the factorials used a!ove in the case of inteer 6
n" $nstead, %e use the amma function in the definition of a0" #0ee appendix C for !ac(round
on the amma function and its a!ility to eneraliHe factorial relationships to noninteer values"&
'hat is %e replace e)uation 44G5 !y the follo%in e)uation for noninteer "
) 1 ( 2
1
0
+

A a
4,95
0u!stitutin this expression for a0 into e)uation 44G5, and usin e)uation 4C395 from .ppendix C
on amma functions ives the follo%in result"
) 1 ( ! 2
) 1 (
) 1 ( ) 1 )( 2 ( ) 1 )( ( ! 2
) 1 (
2 2
2
+ +

+ + + + +

+ +
m m
A
m m m
A
a
m
m
m
m
m

4,45
1e su!stitute this e)uation for a2m #%ithout the ar!itrary constant, .& into the eneral po%er series
solution from e)uation 4245 to define the $essel function of the first &ind of )noninteger+ order
' (

)*+"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2

+ +

0
2
2
) 1 ( ! 2
) 1 (
) (
m
m
m m
m m
x
x x J 4,,5
Fote that this definition for noninteer is the same as e)uation 4,25 for inteer 6 n, since
e)uation 4C3=5 sho%s that #m E n&K 6 #m E n E -&"
Second Frobenius-method solution Fo% that %e have the first solution for inteer and
noninteer , includin 6 0, %e have to find another linearly independent solution to /essel2s
e)uation to form a !asis for all possi!le solutions to this e)uation" 'his re)uires us to consider
each case separately" 'he simplest case is %hen is not an inteer" Aere the values of J

#x&
and J3
#x& are linearly independent and %e can %rite a solution to /essel2s e)uation as follo%s"
) ( ) ( ) ( x BJ x AJ x y

+ 4,=5
Aere . and / are ar!itrary constants determined !y the initial conditions on the oriinal
differential e)uation" 'he values of J3
#x& are found from e)uation 4,,5"
1hen is an inteer #%ith the usual notation that inteer 6 n&, %e can use e)uation 4,25 to sho%
that the /essel functions for order n and order n are linearly dependent" 'hese t%o orders of
/essel functions are simply related as sho%n !elo%" 'his result confirms that J3n is not a linearly
independent solution to /essel2s e)uation"
) ( ) 1 ( ) ( x J x J
n
n
n

4,75
Second Frobenius method solution for integer = n 'o find a second solution for inteer L
%e have to separately consider the special case %here 6 n 6 0" $n this case, %here the indicial
e)uation had a dou!le root, the second solution is sho%n in e)uation 4945" 'o et the second
solution for /essel2s e)uation %ith n 6 0, %e modify e)uation 4945 !y settin the first solution to
J0#x& and the value of r in e)uation 4945 to the value of the dou!le root, r 6 0" 'his ives the
e)uation sho%n !elo% for the second solution

+
1
0 2
) ln( ) ( ) (
m
m
m
x A x x J x y 4,<5
For the more eneral case of inteer 6 n 0, the second solution is iven !y e)uation 49,5"
0ince the t%o roots of the indicial e)uation are r 6 t 6 tn for inteer , and %e used r 6 En in the
oriinal solution for Jn#x&, %e have to use the second root, r 6 3n here" Isin this value for r and
Jn#x& for the oriinal solution, y-#x&, e)uation 49,5 !ecomes"

+ +
0 0
2
) ln( ) ( ) ln( ) ( ) (
m
n m
m n
m
m
m
n
n
x A x x kJ x A x x x kJ x y 4,G5
$n e)uations 4,<5 and 4,G5 used m as the summation index in place of n, %hich %as used in
e)uations 4945 and 49,5" 'his %as done to avoid confusion %ith the use of n to represent the
order of the /essel function" 0ince the summation index is a dummy index, %e can use any
sym!ol %e %ant for this index"
1e can do the initial part of the analysis for eneral inteer 6 n, %here n may or may not !e
Hero" 'o do this %e %ill %rite /essel2s e)uation 49=5 as follo%s, su!stitutin n for n"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
( ) 0
) ( ) (
2 2
2
2
2
+ + y n x
dx
x dy
x
dx
x y d
x 4=05
1e %ill also use e)uation 4,G5 as the second solution" 1hen %e do this, %e have to remem!er to
set ( 6 - and to set the lo%er limit of the summation e)ual to - to ma(e the e)uation for eneral n
in 4,G5 correspond to e)uation 4,<5 for n 6 0" #'hese t%o chanes, plus settin n 6 0, %ill convert
the solution in 4,G5 to the special case of n 6 0 in e)uation 4,<5"&
'a(in the first derivative of e)uation 4,G5 ives

+ +
0
1 2
) (
) ln(
) ( ) ln(
) ( ) (
m
n m
m n
n
x A n m
dx
x d
x kJ x
dx
x dJ
k
dx
x dy
4=-5
0u!stitutin -Dx for dln#x&Ddx and ta(in the second derivative ives"

+ +
0
2
2 2
2
2
2
2
) 1 )( (
) ( 2 ) (
) ln(
) ( ) (
m
n m
m
n n n
x A n m n m
dx
x dJ
x
k
x
x kJ
x
dx
x dJ
k
dx
x dy
4=25
0u!stitutin e)uations 4,G5, 4=-5, and 4=25 into e)uation 4=05 ives the follo%in result"
( )
( ) 0 ) ln( ) (
) (
) ln(
) ( ) ln(
) (
) 1 )( (
) ( 2 ) (
) ln(
) (
) ( ) (
0
2 2
0
1
0
2
2 2
2
2
2
2 2
2
2
2
2
2

1
]
1

+ +
1
]
1

+ + +
1
]
1

+ +
+ +

m
n m
m n
m
n m
m n
n
m
n m
m
n n n
x A x x kJ n x
x A n m
dx
x d
x kJ x
dx
x dJ
k x
x A n m n m
dx
x dJ
x
k
x
x kJ
x
dx
x dJ
k x
y n x
dx
x dy
x
dx
x y d
x
4=95
1e can rearrane this e)uation !y multiplyin each term in !races !y the factors outside the
!races, su!stitutin -Dx for dln#x&Ddx, and collectin all the terms multiplied !y ( ln#x&"
( )
( ) 0
) ( ) 1 )( (
) (
2
) ( ) ( ) (
) ( ) (
) ln(
0
2 2
0 0
2 2
2
2
2
+
+ + +
+
1
]
1

+ +

m
n m
m
m m
n m
m
n m
m
n
n n n
n n
x A n x
x A n m x A n m n m
dx
x dJ
kx
x kJ x kJ x J n x
dx
x dJ
x
dx
x dJ
x x k
4=45
'he last t%o terms in the first ro% o!viously cancel" 'he terms multiplied !y ( ln#x& in this ro% are
+ust /essel2s e)uation #see e)uation 4=05& %ith Jn#x& as the dependent varia!le" Ao%ever, Jn#x& is
a solution to /essel2s e)uation" 'hus e)uation 4=05 tells us that the terms multiplyin ( ln#x& in the
first ro% sum to Hero"
1e can re%rite the summation terms in the second ro% of e)uation 4=45 as follo%s"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
( )
[ ]

+
+ +
+ +
0
2
0
0 0
2 2
0
2 2
0 0
) 2 (
) ( ) 1 )( (
) ( ) 1 )( (
m
n m
m
m
n m
m
m m
n m
m
n m
m
m
n m
m
m m
n m
m
n m
m
x A x A n m m
x A x A n n m n m n m
x A n x x A n m x A n m n m
4=,5
1e can use e)uation 4,25 for Jn to et a series for the derivative term, dJnDdx, in e)uation 4=45"

+
+
+
+

+
+
+
+

+
+

'

0 0
2
2
2
1 2
0
2
2
)! ( ! 2
) 2 ( ) 1 (
2
)! ( ! 2
) 2 ( ) 1 (
2
)! ( ! 2
) 1 (
2
) (
2
m m
n m
n m m
n m
n m m
m
n m
n m m
n
n m m
x n m
k
n m m
x n m
kx
n m m
x
dx
d
kx
dx
x dJ
kx
4=,5
1e can no% su!stitute e)uations 4=,5 and 4==5 into e)uation 4=45, remem!erin that %e have
previously sho%n that the first ro% of 4=45 is e)ual to Hero"
0 ) 2 (
)! ( ! 2
) 2 ( ) 1 (
2
0 0
2
0
2
2
+ +
+
+

+
+
m m
n m
m
n m
m
m
n m
n m m
x A x A n m m
n m m
x n m
k 4=75
Second Frobenius method solution for n = 0 .t this point %e %ant to return to the separate
consideration of n 6 0" 8ecall that %e had to set n 6 0 and ( 6 - in this case" 1e also had to
increase the lo%er limit on the summation of the .mx
m
terms from Hero to one" :a(in these
chanes in e)uation 4=75 ives the follo%in result for n 6 0"
( )
0
! 2
) 2 ( ) 1 (
2
1 1
2 2
0
2 2
2
+ +

m m
m
m
m
m
m
m
m m
x A x A m
m
x m
4=<5
1e can com!ine the last t%o sums if %e temporarily replace m !y + 6 m E 2 in the second sum"
#Bnce %e et the correct form for this sum, %e can replace + !y m and com!ine the t%o sums" 'o
do this %e have to explicitly %rite the terms for m 6 - and m 6 2 in the first sum"&
[ ]

+
+ + + +
+ +
3
2
2 2
2 1
3
2
1
2
1 3
2
2
1 1
2 2
4
m
m
m m
m
m
m
m
m
m
m j
j
j
m
m
m m
m
m
m
m
x A A m x A x A x A
x A m x A x A m x A x A m
4=G5
1e can use e)uation 4=G5 to replace the last t%o sums in e)uation 4=<5" $n addition, %e see that
the 2m in the numerator of the first sum in e)uation 4=<5 ma(es the first term in that sum Hero"
'hus %e can chane the lo%er limit of that sum from m 6 0 to m 6 -" :a(in these chanes
allo%s us to re%rite e)uation 4=<5 as follo%s"
( )
[ ] 0 4
! 2
) 4 ( ) 1 (
3
2
2 2
2 1
1
2 2
2
+ + + +

m
m
m m
m
m
m m
x A A m x A x A
m
x m
4705
$n the first sum, %e can com!ine the 4 in the numerator %ith the 2
2m
in the denominator and %e
can also %rite mD#mK&
2
as -D4mK#m3-&K5" 'his ives the follo%in chane to e)uation 4705"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
[ ] 0 4
)! 1 ( ! 2
) 1 (
3
2
2 2
2 1
1
2 2
2
+ + + +

m
m
m m
m
m
m m
x A A m x A x A
m m
x
47-5
0ince e)uation 47-5 is a po%er series %hose sum is e)ual to Hero, the coefficient of every po%er
in this series must vanish to for the sum to e)ual Hero for any value of x" 1e see that the lo%est
po%er of x occurs in the term .-x" 1e must have .- 6 0 for the x coefficient to vanish"
1e next consider the x
2
term" For the coefficient of this term to vanish %e must satisfy the
follo%in e)uation"
4
1
0 4 1 4
)! 1 1 ( ! 1 2
) 1 (
2 2 2
2 ) 1 ( 2
1
+ +

A A A 4725
0ince the first sum in e)uation 47-5, %hich comes from the first derivative of Jn#x&, has only even
po%ers of x, the coefficients for odd po%ers of x are iven !y the second sum only" 0ettin the
coefficient of odd po%ers of x e)ual to Hero ives the follo%in result"
m odd
m
A
A
m
m
2
2
4795
0ince %e have previously sho%n that .- 6 0, e)uation 4795 tells us that .9 6 0" $n fact %e can
apply it se)uentially to all values of .m %ith an odd su!script to sho% that all values of .m for
%hich m is odd are Hero" 'his leaves us %ith only even values of m to consider" For convenience
%e can re%rite the second sum in e)uation 47-5 to have only even po%ers" 'o do this %e
temporarily define the index ( 6 mD2 and replace m !y 2( in the second sum in e)uation 47-5"
/efore doin this %e set .2 6 0 and start the sum at m 6 4 since the m 6 9 term is Hero"
[ ] 0 ) 2 ( 4
)! 1 ( ! 2
) 1 (
2
2
2 2 2
2 2
2
1
2 2
2
+ + +

k
k
k k
m
m
m m
x A A k x A
m m
x
4745
From e)uation 4745, %e can see that the coefficients of even po%ers, x
2m
, for m J -, %ill vanish if
the follo%in e)uation is satisfied"
1 0 ) 2 (
)! 1 ( ! 2
) 1 (
2 2 2
2
2 2
> + +

m A A m
m m
m m
m
m
47,5
'his allo%s us to define ne% coefficients in terms of old ones !y the follo%in e)uation"
2
2 2
2 2 2
2
) 2 ( )! 1 ( ! ) 2 ( 2
) 1 (
m
A
m m m
A
m
m
m
m

47=5
For m 6 2, %e can compute .4 in terms of .2 6 M"
128
3
16
4
1
) 1 )( 2 )( 16 )( 4 (
1
) 2 2 ( )! 1 2 )( ! 2 ( ) 2 2 ( 2
) 1 (
2
2
2 2 ) 2 ( 2
2
4

A
A
4775
For m 6 9, %e can compute .= in terms of .4 and use the result that .4 6 39D-2< to et a value for
.="
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
284 , 13
11
284 , 13
9
284 , 13
2
608 , 4
3
912 , 6
1
36
128
3
) 2 )( 6 )( 36 )( 16 (
1
) 3 2 ( )! 1 3 )( ! 3 ( ) 3 2 ( 2
) 1 (
2
4
2 2 ) 3 ( 2
3
6
+ +

A
A
47<5
'he eneral form for .2m is not easy to see" 1e can see that the recursion e)uation 47=5 for .2m
contains one term from the series sum for the first derivative of J0 and a second term %hich
contains the previous value of .2m" /ecause this continues indefinitely, eneral expression %ill
loo( li(e the follo%in expression that %e found for .="

'

,
_

1
]
1

1
]
1

1
]
1

'

1
]
1

1
]
1

1
]
1

term J
m
term J
m
term J
m
A
term J
m
term J
m
A
term J
m
A
'
0
2 '
0
2 '
0
2
2
'
0
2 '
0
2
4
'
0
6
1
) 2 2 (
1
2
) 3 2 (
1
3
) 2 2 (
2
) 3 2 (
1
3
) 3 2 (
3
47G5
7)uation 47G5 sho%s the eneral form that the J0
2
term appears repeatedly %ith a series of factors"
.lthouh %e %ill not attempt to derive the eneral e)uation !elo%, e)uation 47G5 tells us that the
form of the e)uation proposed in 4<05 for .2m is reasona!le"

m
k
m
m
m
k m
A
1
2 2
1
2
1
) ! ( 2
) 1 (
4<05
1e can sho% that this result satisfies the difference e)uation in 47=5 as follo%s" First, %e re%rite
e)uation 4<05, replacin m !y m -"
[ ]

1
1
2 2 2
2
2 2
1
)! 1 ( 2
) 1 (
m
k
m
m
m
k
m
A 4<-5
1e then su!stitute e)uation 4<-5 into e)uation 47=5 and after some manipulation %e see that %e
o!tain the same result for .2m iven !y e)uation 4<05"
[ ]
[ ]
[ ][ ] [ ]

1
]
1

m
k
m
m m
k
m
m
m
k
m
m
m
m
m
k
m
m
m
m
m
k
m
m
m
m
m
k m k m m
k
m m m m m m m
k
m m m m m m
k
m m m m m
A
1
2 2
1 1
1
2 2
1
1
1
2 2
1
2
1
1
1
2 2 2 2 2
1
2 2 2 2
1
1
1
2 2 2
2
2 2 2 2
2
1
) ! ( 2
) 1 ( 1 1
) ! ( 2
) 1 (
1
)! 1 ( 2
) 1 (
)! 1 ( )! 1 ( 2
) 1 (
1
)! 1 ( 2 2
) 1 (
)! 1 ( )! 1 ( 2 2
) 1 (
1
)! 1 ( 2
) 1 (
) 2 (
1
)! 1 ( ! ) 2 ( 2
) 1 (
4<25
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
0ince the last expression in e)uation 4<25 is the eneral expression proposed for .2m in e)uation
4<05, %e conclude that e)uation 4<05 is a correct solution to the recursion e)uation that %e derived
for .2m in terms of .2m32 in e)uation 47=5"
$f %e su!stitute this e)uation 4<05 for .2m into the oriinal e)uation for our second solution in 4,<5,
%e et the result that

'

+
1 1
2 2
2 1
0 2
1
) ! ( 2
) 1 (
) ln( ) ( ) (
m
m
k
m
m m
k m
x
x x J x y 4<95
.lthouh this solution provides a second linearly independent solution to /essel2s e)uation for 6
0, it is conventional to ta!ulate the function N0#x& sho%n !elo% as the second linearly3independent
solution" #.ny com!ination of the t%o linearly independent solutions can !e used as one of the
linearly independent solutions" $n this e)uation is called the 7uler constant" $t has a value of
0",772-,==4G0-,92<=0=0=-,2O, and represents the limit of the follo%in sum as x approaches
infinity: - E P E -D9 E M EE-Dx ln x"
[ ]
]
1
]
1

'

,
_

'

1 1
2 2
2 1
0 0
1 1
2 2
2 1
0 0 2 0
1
) ! ( 2
) 1 (
2
ln ) (
2
) ( ) 2 ln (
1
) ! ( 2
) 1 (
) ln( ) (
2
) ( ) 2 ln ( ) (
2
) (
m
m
k
m
m m
m
m
k
m
m m
k m
x x
x J x J
k m
x
x x J x J x y x Y
4<45
Second Frobenius method solution for integer n = 0 Aavin completed the second
solution for J0, %e can return to the eneral case of findin the second solution for Jn, %here n is
nonHero" 1e return to e)uation 4=75, the last e)uation in our eneral development for any inteer
6 n, and re%rite the final t%o sums in that e)uation as follo%s"
[ ]

+
+ + + +
+
2
2
1
1
2
2
0 0 0
2
) 2 ( ) 2 1 ( 0
) 2 ( ) 2 (
j
n j
j j
n
j
n j
j
m
n m
m
m m
n m
m
n m
m
x A A n j j x A n x A
x A n m m x A x A n m m
4<,5
0u!stitutin this e)uation into e)uation 4=75 ives the follo%in result"
[ ] 0 ) 2 ( ) 2 1 (
)! ( ! 2
) 2 ( ) 1 (
2
2
2
1
1
0
2
2
+ + +
+
+

+
+
j
n j
j j
n
m
n m
n m m
x A A n j j x A n
n m m
x n m
k
4<=5
'he minimum po%er of x in this e)uation is x
-3n
" 'he coefficient of this term is #- 2n& .-" 0ince
n is an inteer, this coefficient %ill only vanish if .- 6 0"
/ecause the first term has x raised to the 2mEn po%er %hile the other terms have x raised to the
+n po%er, the t%o different summations %ill have common po%ers of x only %hen 2m E n 6 + n"
'hat is %hen + 6 2m E 2n" 0ince the minimum value of m is Hero, the first sum %ill only come into
play %hen + 2n" For + Q 2n, then, the coefficients of the x terms %ill vanish only if the coefficients
of x
+3n
in the second summation are Hero" 'his re)uires
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
) 2 (
0 ) 2 (
2
2
n j j
A
A A A n j j
j
j j j

4<75
0ince .- 6 0, this e)uation tells us that all values of .+, %ith an odd num!er for +, %ith + Q 2n are
Hero" For even + Q 2n, the coefficients are !e expressed in terms of .0, %hich %ill remain an
un(no%n" For + 6 2, 4, and =, %e have the follo%in results from e)uation 4<75"
) 2 6 )( 2 4 )( 2 2 ( 6 4 2 ) 2 4 ( 4
) 2 4 )( 2 2 ( 4 2
) 2 6 ( 6
4
) 2 4 )( 2 2 ( 4 2 ) 2 4 ( 4
) 2 2 ( 2
) 2 4 ( 4 ) 2 2 ( 2
0
0
6
6
0
0
2
4
0
2
n n n
A
n
n n
A
n
A
A
n n
A
n
n
A
n
A
A
n
A
A

,
_

,
_

4<<5
1e can continue in this fashion until %e compute .2n32C the next value of + %ill !e + 6 2n" .t this
point, the exponent of x
+3n
in the second sum of e)uation 4<=5 !ecomes e)ual to 2n n 6 n" 'his
is the same po%er of x that occurs at the lo%er limit of m 6 0 in the first sum of e)uation 4<=5" 0o
for + n, %e have to consider !oth sums in e)uation 4<=5" 'he follo%in e)uation, o!tained !y
settin m 6 0 in the first sum and + 6 2n in the second sum, is re)uired to set the coefficient of x
n

in e)uation 4<=5 e)ual to Hero"
0 ) 2 2 ( 2
)! 0 ( ! 0 2
) 0 2 ( ) 1 (
2
2 2 2
0 2
0
+ +
+
+

+
n n
n
A A n n n
n
n
k 4<G5
1e can solve this e)uation for (, the coefficient of the Jn#x&ln#x& term in the proposed second
solution
0
2 2
2
! 2
A
n
A n
k
n
n


4G05
1e %ill develop an e)uation for ( later in the derivation" Fote that the value of ( may !e Hero in
other solutions !y Fro!enius method %here the roots of the indicial e)uation differ !y an inteer"
'he remainin po%ers of x in the series have contri!utions from !oth sums" 'he po%er of x in the
t%o sums %ill !e the same if + 6 2m E 2n" #'his covers only even po%ers of x, %hich are the only
po%ers left in the series %hose coefficients are not already Hero"& 0ettin + 6 2m E 2n in the final
sum of e)uation 4<=5 ives the follo%in e)uation to ma(e the coefficients x
2mEn
vanish"
0 ) 2 )( 2 2 (
)! ( ! 2
) 2 ( ) 1 (
2
2 2 2 2 2
2
+ + +
+
+
+ +
+
n m n m
n m
m
A A m n m
n m m
n m
k 4G-5
$f %e set m 6 0 in this e)uation %e recover e)uation 4<G5 %hich %e used to solve for (" 'hus, this
e)uation only applies for m -" 1e can solve this e)uation for .2mE2n as follo%s
1
) 2 )( 2 2 ( )! ( ! ) 2 )( 2 2 ( 2
) 2 ( ) 1 (
2
2 2 2
2
2 2

+

+ +
+

+
+
+
m
m n m
A
n m m m n m
n m
k A
n m
n m
m
n m
4G25
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
1hen %e apply this e)uation for m 6 -, usin the result that ( 6 ( .0, %e o!tain .2nE2,
) 4 4 ( )! 1 ( ) 1 ( 2
) 2 (
) 2 )( 2 2 ( )! 1 ( ! 1 ) 2 )( 2 2 ( 2
) 2 ( ) 1 (
2
2
3
0 2
2
1
0 2 2
n
A
n n
n A
n
A
n n
n
A A
n
n
n
n
n
+

+ +
+

+ +
+

+ +
+
4G95
1e can continue to apply e)uation 4G45 to et additional values of .m" 'he eneral result for this
coefficient is tedious to o!tain and %e %ill s(ip the steps that lead to the follo%in result"

+ +

'

+

+

'

,
_

+
+

+
1
0
2 2
2
0 1 1
2 2
2 1
2
)! ( ! 2
)! 1 ( 1 1
)! ( ! 2
) 1 (
) ln( ) ( ) (
n
m
n m
m
n
m
m
k
n m
k
n m
m m
n
n
n m m
x m n
x
k k n m m
x
x
x x J x y
4G45
.lthouh this solution provides a second linearly independent solution to /essel2s e)uation for 6
n 0, it is conventional to ta!ulate the function Nn#x& sho%n !elo% as the second solution" 'he
definition of this second solution is similar to the one used in o!tainin e)uation 4<45 for N0#x&"

+

+
1
]
1

,
_

+
+

'

,
_

1
]
1

+ +

0
1
0
2 2
2
1 1
2 2
2 1
2
)! ( ! 2
)! 1 ( 1 1
)! ( ! 2
) 1 (
2
ln ) (
2
) ( ) 2 ln ( ) (
2
) (
m
n
m
n m
m
n
m
k
n m
k
n m
m m
n
n n n
n m m
x m n
x
k k n m m
x
x
x
x J x J x y x Y
4G,5
$f %e set n 6 0 in e)uation 4G,5, %e %ill o!tain e)uation 4<45 for N0#x& if the final sum in e)uation
4G,5 is omitted %hen n 6 0"
'here is one final step that is ta(en to o!tain consistency !et%een inteer3order and noninteer3
order /essel functions" .lthouh J

and J3
are linearly independent solutions for noninteer ,
%e define an alternative second solution for noninteer n !y the follo%in e)uation"
( )

sin
) ( ) ( cos
) (
x J x J
x Y
4G=5
$n the limit as approaches an inteer, the numerator and denominator of e)uation 4G=5 !oth
approach Hero" #8ecall e)uation 4,75 for /essel functions of inteer order: J3n#x& 6 #3-&
n
Jn#x&C since
cos n 6 #3-&
n
, the numerator approaches Hero as approaches an inteer"& $f %e apply
L2Aopital2s rule to e)uation 4G=5 as approaches inteer n, %e can sho% that the result %e et is
the same as that iven for Nn#x& in e)uation 4G,5"
'he !asic result of this section on /essel functions then is that the solution to /essel2s e)uation
49=5 is y 6 C-J

#x& E C2N

#x& for !oth inteer and noninteer " Aere C- and C2 are constants that
are determined !y the !oundary conditions on the differential e)uation" $n the same %ay that %e
reconiHe that y 6 C-sin#(x& E C2cos#(x& is the solution to d
2
yDdx
2
E (
2
y 6 0, %e can no% reconiHe
that y 6 C-J

#x& E C2N

#x& is the solution to x

2
d
2
yDdx
2
E xdyDdx E #x
2
3
2
&y 6 0" Bf course, /essel
function ta!les are not as common as ta!les of sines and cosines and you pro!a!ly do not have a
/essel function !utton on your calculator" Ao%ever, on an 7xcel spreadsheet you can et /essel
functions Jn#x&, for inteer n, !y the function !essel+#x,n&" 0imilarly, you can et Nn#x&, for !oth
inteer n, !y the function !essely#x,n&" 'hese functions %ere used to prepare the plots of Jn#x&
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
sho%n a!ove the the plots of Nn#x& sho%n !elo%" 'he :.'L./ functions !essel+#,x& and
!essely#,x& evaluate J

#x& and N

#x& for !oth inteer and noninteer "

$essel "unctions of the Second ,ind of -nteger Order
.4
./09
./08
./07
./06
./05
./01
./02
./03
./04
/
/04
/03
/02
/01
/05
/06
/ 4 3 2 1 5 6 7 8 9 4/ 44 43 42 41 45
*
;
n
)
*
+
n : /
n : 4
n:1
'his plot sho%s that the values of Nn#x& approach minus infinity as x approaches Hero due to the
ln#x& term in Nn#x&" /ecause the common application of /essel functions is to pro!lems %ith
radial eometries, %e usually have to ta(e C2 6 0 in the eneral solution, y 6 C-J

#x& E C2N

#x&, to
have a solution that remains finite at x 6 0"
Summary of $essel functions
'he differential e)uation of the form x
2
d
2
yDdx
2
E xdyDdx E #x
2
3
2
&y 6 0 are (no%n as
/essel2s e)uation" $ts main enineerin application is to pro!lems in radial eometries"
'he eneral solution to /essel2s e)uation is y 6 C-J

#x& E C2N

#x& %here C- and C2 are

constants that are determined !y the !oundary conditions on the differential e)uation"
'he functions J

#x& and N

#x& are (no%n as the /essel functions of order of the first and
second (ind, respectively" Ralues of these functions, %hich have oscillatory !ehavior,
may !e found in various ta!les and computer li!rary solutions"
For inteer n, at x 6 0, J0#x& 6 - and Jn#x& 6 0" .s x approaches Hero, Nn#x& approaches
minus infinity"
.lthouh not discussed in these notes, $t is possi!le to transform some e)uations into the
form of /essel2s e)uation"
1e %ill consider /essel functions in su!se)uent portions of the :7 ,0-./ course"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
Summary of "roenius method
Fro!enius method is used to determine series solutions to the differential e)uation 'he differential
e)uation of the form x
2
d
2
yDdx
2
E x!#x&dyDdx E c#x&y 6 0" $n these notes %e have applied this
method to the solution of /essel2s e)uation"
'he eneral form of the Fro!enius method solution is the infinite series y#x& 6 x
r
#a0 E a-x
E a2x
2
E O"&
'he eneral solution is differentiated and su!stituted into the oriinal differential e)uation"
0ettin the coefficients of each po%er of x
n
e)ual to Hero ives e)uations that can !e
solved for r and the ai coefficients"
0ettin the coefficient of x
r
6 0 ives a )uadratic e)uation for r (no%n as the indicial
e)uation"
'here are three possi!le cases for the t%o roots of the indicial e)uation: #-& the t%o roots
are the sameC #2& the t%o roots differ !y an inteer #other than Hero&C #9& the t%o roots are
different and their difference is not an inteer"
$n all three cases, the first solution is y-#x& 6 x
r
#a0 E a-x E a2x
2
E O"&, %here r is one root to
the indicial e)uation" 'his must !e ta(en as the reater of the t%o roots if the roots differ
!y an inteer"
'he values of the ai coefficients in the solution are determined in the same %ay as in the
eneral po%er series solutionC the coefficients of each po%er of x
n
in the solution must !e
Hero"
$f the t%o roots of the indicial e)uation are different, and the difference is not an inteer,
the second solution is is y2#x& 6 x
8
#.0 E .-x E .2x
2
E O"&, %here 8 is the second root to
the indicial e)uation"
$f the t%o roots are the same, the second solution is is y2#x& 6 y-#x& ln#x& E #.-x E .2x
2
E
.9x
9
E O"&"
$f the t%o roots differ !y an inteer, the second solution is is y2#x& 6 ( y-#x& ln#x& E #.0 E
.-x E .2x
2
E .9x
9
E O"&, %here ( may !e Hero"
'he coefficients .i #and the value of (& are found !y ma(in the coefficients of all po%ers
of x e)ual to Hero"
Nou should understand ho% Fro!enius method %or(s, and !e a!le to apply this method"
Ao%ever, it does not play a role in the solution of most practical enineerin pro!lems"
<he Sturm.=iouville prolem
'he 0turm3Liouville pro!lem forms the !asis for several pro!lems in enineerin analysis" $t
provides a lin( amon different functions includin /essel functions, sines and cosines, and other
special functions" 0ets of these functions, %hich are solutions to the 0turm3Liouville pro!lem,
have a common a!ility to represent any function over a certain reion in space"
'he 0turm3Liouville pro!lem is defined !y the follo%in differential e)uation over a reion a x
!"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
[ ] 0 ) ( ) ( ) ( + +

,
_

y x p x q
dx
dy
x r
dx
d
4G75
%ith the follo%in !oundary conditions at x 6 a and x 6 !" $n these !oundary conditions, at least
one of the t%o constants (- and (2 is not Hero" 0imilarly, at least one of the t%o constants S- and S2
is not Hero"
0 ) ( 0 ) (
2 1 2 1
+ +
b x a x
dx
dy
b y and
dx
dy
k a y k
4G<5
'he functions r#x&, dr#x&Ddr, )#x& and p#x& must !e continuous in the reion a x ! and %e must
have p#x& J 0" 'he 0turm3Liouville pro!lem is one of a eneral class of pro!lems involvin linear
operators" 1e can define a eneral linear operator, L, %hich may !e a derivative operator such
as dDdx or d
2
Ddx
2
, or a com!ination of operations such as #d
2
Ddx
2
E -&" $f our linear operator, L,
operates on a function and returns the same function times a constant, l, %e fall the function an
eienfunction of the operator and the constant an eienvalue"
$f Lf#x& 6 f#x&, then f#x& is an eienfunction of L and is an eienvalue 4GG5
Fote the similarity to the matrix eienvalue pro!lem, #* 6 *" For example if L is the simple first
derivative operator, e)uation 4GG5 !ecomes
x
e f f
dx
df

4-005
'hus %e say that e
x
is the eienfunction of the operator dDdx" 1e can define a eneraliHed
eienfunction pro!lem !y the follo%in e)uation"
Lf#x& 6 p#x&f#x& 4-0-5
Aere p#x& is called the %eiht function" 1e see that the 0turm3Liouville pro!lem is an
eienfunction pro!lem of %ith the form of e)uation 4-0-5C the definition of the operator for this
pro!lem can !e found !y examinin e)uation 4G75"
) (
) (
) (
2
2
x q
dx
d
dx
x r
dx
d
x r L 4-025
For functions, such as those in the solutions to the 0turm3Liouville pro!lem, the inner product is
defined as follo%s"
( )

b
a
j i j i
dx x p x f x f f f ) ( ) ( ) ( ,
*
4-095
$n this definition the function p(x&, %hich appears in the 0turm3Liouville pro!lem is called the
%eihtin function" $n many cases, p#x& 6- and is not considered in the definition of the inner
product of functions" For the 0turm3Liouville pro!lem, p#x& is defined to !e al%ays reater than
Hero"
'he 0trum3Liouville operator is one of a class of operators (no%n as Aermetian or self3ad+oint
operators" 'he linear operator, L, has an ad+oint operator, LT, %hich is also linear, defined !y the
follo%in inner product e)uation" Aere fi and f+ are functions on %hich the operators L and LT act"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
) * , ( ) , (
j i j i
f L f f Lf
4-045
$t is possi!le to have an operator that is self3ad+oint or Aermetian operatorC that is an operator for
%hich LT 6 L" 1e can sho% that the operator in the 0turm3Liouville e)uation, as defined in
e)uation 4-025, satisfies the definition of a self3ad+oint usin the inner3product definition in
e)uation 4-095" For a self3ad+oint operator, e)uation 4-045 tells us that #Lfi,f+& 6 #fi,Lf+&"
dx x fp x q
dx
df
x r
dx
f d
f dx x p f x q
dx
df
x r
dx
f d
b
a
j j
i j
b
a
i i
) ( ) ( ) ( ) ( ) ( ) (
2
2
2
2

,
_

,
_

4-0,5
1e no% state t%o important results for Aermetian #or self3ad+oint& operators in eneral and for the
0turm3Liouville operator in particular:
-" 'he eienvalues of any self3ad+oint or Aermetian operator are real"
2. 'he eienfunctions of any self3ad+oint or Aermetian operator defined over a reion a x
! form an orthoonal set over that reion"
9" 'he eienfunctions form a complete set if the vector space has a finite num!er of
dimensions as in an n x n matrix"
4" 'he 0turm3Liouville operator has a complete set of einefunctions over an infinite3
dimensional vector space"
'he notion of orthoonal functions is an extension of the notions of inner products and
orthoonality for vectors" For functions, the inner product is defined in terms of the interal in
e)uation 4-095" $f %e have a set of functions, fi#x& defined on an interval a x !, are orthoonal
over that interval if the follo%in relationship holds"
( )
ij i
b
a
j i j i
a dx x p x f x f f f

) ( ) ( ) ( ,
*
4-0=5
1e can define a set of orthonormal functions !y the follo%in e)uation"
( )
ij
b
a
j i j i
dx x p x f x f f f

) ( ) ( ) ( ,
*
4-075
.ny set of orthoonal functions i#x& can !e converted to a set of orthonormal functions fi#x& !y
dividin !y the s)uare root of the inner product, #i,i&, %hich %e can also %rite as the t%o norm, UU
iUU2" $f %e understand that %e are al%ays usin the t%o3norm, %e can drop the su!script and
%rite this more simply as UUiUU"
( )

b
a
j i
i
i i
i
i
i
i
dx x p x g x g
x g
g g
x g
g
x g
x f
) ( ) ( ) (
) (
,
) ( ) (
) (
*
4-0<5
Eigenfunction e*pansions
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
$f %e have a complete set of eienfunctions in a reion a x ! %e can expand any other
function , f#x&, in that reion !y the follo%in e)uation"

0
) ( ) (
m
m m
x y a x f 4-0G5
1e can derive a simple expression for the coefficients in this e)uation if the eienfunctions are
orthoonal" $f %e multiply !oth sides of this e)uation !y #x&yn#x& and interate !et%een x 6 a and
x 6 !, %e o!tain the follo%in result"

0 0
) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) (
m
b
a
m n m
b
a
m
m m n
b
a
n
dx x y x y x p a dx x y a x y x p dx x f x y x p
4--05
1e can re%rite these interals usin the inner product notation"
( ) ( )

0
, ,
m
m n m n
y y a f y 4---5
0ince the ym form an orthoonal set, the only nonHero term in the summation on the riht is the
one for %hich n 6 m" .ll other inner products are Hero !ecause of orthonality" 'hus %e have a
simple e)uation to solve for am after settin n 6 m every%here in the e)uation"
( )
( )
( )

b
a
m m
b
a
m
m
m
m m
m
m
dx x y x y x p
dx x f x y x p
y
f y
y y
f y
a
) ( ) ( ) (
) ( ) ( ) (
,
,
,
2
4--25
$f %e have an orthonormal set, the denominator of the am e)uation is one"
'he most common eienfunction expansions are the Fourier series of sines and cosines" 'hese
functions are solutions of a 0turm Liouville differential e)uation defined !y e)uations 4G75 and
4G<5" 'o see this %e consider the differential e)uation d
2
yDdx
2
E
2
y defined over 0 x -, %ith
y#0& 6 0 and y#-& 6 0" 'his is a 0turm3Liouville e)uation %ith r#x& 6 p#x& 6 - , 6
2
, and )#x& 6 0"
'he !oundary conditions satisfy e)uation 4G<5 %ith (- 6 S- 6 -and (2 6 S2 6 0" 'he set of functions
ym 6 sin#mx& 6 sin#mx&, %here m is any inteer, satisfied !oth the differential e)uation and the
!oundary conditions" #Nou can sho% this !y su!stitution into the differential e)uation" Nou
should !e a!le to say %hy the cosine %ill not !e a solution to this pro!lem"&
'he solutions are orthoonal, !ut %e cannot tell if they are orthonormal unless find the norm of a
function" 'his re)uires the follo%in interal for the norm" #Fote that p#x& %as e)ual to one in our
pro!lem definition so it does not appear in the inner product"& 'he resultin interal is evaluated
usin interal ta!les for indefinite interal of sin
2
#ax&"
( )
( )
2
1
4
2 sin
2
1
) ( sin ,
1
0
1
0
2
2

1
]
1

x m x m
m
dx x m y y y
m m m
4--95
'hus, for this set of functions, the coefficients in an eienfunction expansion are computed from
the follo%in modification of e)uation 4--25"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
( ) ( )

,
_

1
0
) ( ) sin( 2
2
1
, ,
dx x f x m
f y
y
f y
a
m
m
m
m

4--45
For example, %e can expand the simple function f#x& 6 c, a constant, usin the follo%in
coefficients"
[ ] [ ] ) cos( 1
2
) cos(
2
) sin( 2
1
0

m
m
c
x m
m
c
cdx x m a
b
a
m
4--,5
'he term - cos#m& 6 0 if m is even and 2 if m is odd" 1e can thus %rite the am coefficients for
this expansion as follo%s"

'

odd m
m
c
even m
a
m
4
0
4--=5
$n this case, our eienfunction expansion in e)uation 4-0G5 !ecomes

5 , 3 , 1 0
) sin(
4
) ( ) (
m m
m m
x m
m
c
x y a c x f
4--75
1e see that %e can divide !y c and o!tain the follo%in expression"

, 5 , 3 , 1
) sin( 4
1
m
m
x m
4--<5
'he partial sums of this series for a small num!er of terms are sho%n in the fiure !elo%"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
Orthogonal Sine Series for f)*+ : 4
/
/03
/01
/06
/08
4
403
401
/ /03 /01 /06 /08 4
*
S
e
r
i
e
s

s
u
m
4 term
3 terms
2 terms
5 terms
4/ terms
1e see that, as %e ta(e additional terms, the series converes to the value of f#x& 6 -, !ut there
are continued oscillations a!out this point" $n addition, at the !oundaries of x 6 0 and x 6 -, the
series sum is Hero"
$essel%s equation as a Sturm.=iouville prolem and its eigenfunction e*pansions
1e can sho% that /essel2s e)uation 49=5 is a 0turm3Liouville e)uation" 1e multiply e)uation 49=5
!y x
2
and define a ne% varia!le, H 6 xD( such that dyDdH 6 #-D(&dyDdH and #-D(
2
&d
2
yDdH
2
" 1e can
the re%rite e)uation 49=5 in terms of H" $f %e divide the result !y H, %e o!tain"
( ) 0
1 1
) (
1
2
2
2
2
2 2 2
2
2
2
2

,
_

+ +
1
]
1

+ + y z k
z dz
dy
dz
y d
z y z k
dz
dy
k
kz
dz
y d
k
kz
z
4--G5
Ao%ever, since %e started %ith an e)uation for y as a function of x in e)uation 49=5 and %e have
not chaned this, %e could formally %rite y 6 y#x& 6 y#(H& in e)uation 4--G5 and !elo% in e)uation
4-205" 0ince Hd
2
yDdH
2
E dyDdH can !e %ritten
,
as d4H dyDdH5DdH, %e can reduce e)uation 4--G5 to
the 0turm Liouville form of e)uation 4G75 as follo%s"
( ) 0
2
2
2
2
2
2
+ +

,
_

,
_

,
_

,
_

+ + y z q
dz
dy
r
dz
d
y z k
z dz
dy
z
dz
d
y z k
z dz
dy
dz
y d
z 4-205
1e see that /essel2s e)uation has the follo%in definitions of the functions r, ), and p, and the
eienvalue , in the 0turm3Liouville e)uation: r 6 H, ) 6 3n
2
DH, p 6 H, and 6 (
2
"
5
Nou can sho% that this is correct !y applyin the rule for differentiation of products to d4H dyDdH5DdH"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
$f %e consider the case of inteer 6 n over a reion 0 H 8 the eneral solution of /essel2s
e)uation, remem!erin that %e no% have y#(H& as our dependent varia!le, is a linear com!ination
of Jn#(H& and Nn#(H&" Ao%ever, Nn#(H& approaches infinity as H approaches Hero" 'hus, Nn#(H&
cannot !e part of our eneral solution %hen %e include H 6 0" $f the !oundary condition at H 6 8
is y#(8& 6 0, %e must select ( to satisfy this e)uation" $t turns out that there are an infinite num!er
of values that %e can select for %hich (8 6 0" 'here are (no%n as the Heros of the /essel
function" #0ee the plots of the /essel function a!ove that sho% the initial locations at %hich the
first fe% /essel functions are Hero"&
For convenience %e define mn and (mn as follo%s"
R
k R k J J
mn
mn mn n mn n

0 ) ( ) ( 4-2-5
'he values of mn, called the Heros of Jn, are the values of the arument of Jn for %hich Jn is Hero"
'here are an infinite of values, mn, for %hich Jn#mn& 6 0" Fote the t%o su!scripts for mnC m is an
eienfunction3countin index that ranes from one to infinity" *o not confuse this eienfunction
index %ith the index, n, for Jn" 'he index for Jn is set !y the appearance of n in the oriinal
differential e)uation" 'hus our set of eienfunctions, for fixed n, %ill !e Jn#(-nH&, Jn#(2nH&, Jn#(9nH&,
Jn#(4nH&, etc"
'his set of functions, Jn#(mnH& is orthoonal, since it is a solution to a 0turm3Liouville pro!lem"
'he orthoonality condition satisfied !y these functions is defined from e)uation 4-0=5 %ith a
%eihtin function, p#H& 6 H, as discussed in the pararaph follo%in e)uation 4-205" .pplyin
e)uation 4-0=5 to this pro!lem %e have the interval from 0 to 8 #in place of a to !& and %e have a
real function so %e do not have to consider the complex con+uate" 'his ives the follo%in
orthoonality condition"
( ) ( )
mo m
R
onz n mnz n onz n mnz n o m
a zdz z k J z k J r k J z k J f f

0
) ( ) ( ) ( ), ( , 4-225
1e can apply the usual e)uation for an eienfunction expansion from 4-0G5 to this set of /essel
functions"

1 1
) ( ) ( ) (
m
mn n m
m
m m
z k J a z y a z f 4-295
$f %e multiply !oth sides of this e)uation !y HJn#(onH&dH and interate from 0 to 8 %e o!tain the
follo%in result"

R
on o o
m
R
on n mn n m
R
on n
zdz z k J a zdz z k J z k J a zdz z k J z f
0
2
1
0 0
) ( ) ( ) ( ) ( ) ( 4-295
$n the final step %e use the orthoonality relationship %hich ma(es all terms in the sum, except
for the term in %hich m 6 o, Hero" 1e can use an interal ta!le to evaluate the normaliHation
interal"
=
6
'he eneral form of this interal, %hich is found !y interation !y parts follo%ed !y su!stitution of /essel2s
e)uation, is sho%n !elo%"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
2
) (
) (
2
1
2
0
2
z k J R
zdz z k J
on
R
on o

4-245
Com!inin e)uations 4-295 and 4-245 and usin m in place of o as the coefficient su!script ives
the follo%in result for the eienfunction expansion coefficients"

R
mn n
mn
m
zdz z k J z f
z k J R
a
0
2
1
2
) ( ) (
) (
2
4-295
.lthouh this seems li(e an unli(ely choice of eienfunctions %ith %hich to expand an ar!itrary
f#x&, %e %ill see that such expansions !ecome important in the consideration of partial differential
e)uations in cylindrical eometries"
Summary of these notes
1e have developed the !ac(round to solve the 0turm3Liouville pro!lem form several eneral
cases !y developin the tools of po%er series solutions and Fro!enius method" 1e sho%ed the
details of ho% one applies Fro!enius method to the solution of /essel2s e)uation" 'he text
covers other applications of this method"
1e discussed the 0turm3Liouville pro!lem" 1e noted that this pro!lem %as defined !y the
differential e)uation and !oundary conditions in e)uations 4G75 and 4G<5 that are copied !elo%"
[ ] 0 ) ( ) ( ) ( + +

,
_

y x p x q
dx
dy
x r
dx
d
4G75
$n the !oundary conditions !elo%, at least one of the t%o constants (- and (2 is not Hero"
0imilarly, at least one of the t%o constants S- and S2 is not Hero" $t is also possi!le to have periodic
!oundary conditions" 'hese are discussed !y VreysHi"
0 ) ( 0 ) (
2 1 2 1
+ +
b x a x
dx
dy
b y and
dx
dy
k a y k
4G<5
1e noted that the 0turm3Liouville pro!lem %as an example of a Aermetian or self3ad+oint
operator" 'his operator for functions had properties similar to a Aermetian matrix, %hich could
also !e rearded as a Aermetian operator" 1e listed the follo%in important results for
Aermetian #or self3ad+oint& operators in eneral and for the 0turm3Liouville operator in particular:
-" 'he eienvalues of any self3ad+oint or Aermetian operator are real"
2. 'he eienfunctions of any self3ad+oint or Aermetian operator defined over a reion a x
! form an orthoonal set over that reion"
9" 'he eienfunctions form a complete set if the vector space has a finite num!er of
dimensions as in an n x n matrix"
4" 'he 0turm3Liouville operator has a complete set of eienfunctions over an infinite3
dimensional vector space"
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
For functions, the inner product is defined in terms of the interal in e)uation 4-095" 'his
definition includes a %eiht function, p#x&, %hich may !e -" 'he orthoonality condition for
functions %as iven in e)uation 4-0=5, %hich is copied !elo%"
( )
ij i
b
a
j i j i
a dx x p x f x f f f

) ( ) ( ) ( ,
*
4-0=5
$f %e had an orthonormal set of functions, then the value of ai in the previous e)uation %ould !e
one"
/ecause the eienfunctions of the 0turm3Liouville pro!lem form a complete orthoonal set, %e
can expand any function in the reion in %hich the 0turm3Liouville pro!lem is defined in terms of
these eienfunctions" 7)uation 4-0G5 ives this eneral eienfunction expansion !y the follo%in
e)uation

0
) ( ) (
m
m m
x y a x f
, %here the coefficients am are iven !y e)uation 4--25 copied
!elo%"
( )
( )
( )

b
a
m m
b
a
m
m
m
m m
m
m
dx x y x y x p
dx x f x y x p
y
f y
y y
f y
a
) ( ) ( ) (
) ( ) ( ) (
,
,
,
2
4--25
7nineerin /uildin 8oom -999 :ail Code ;hone: <-<"=77"=44<
73mail: lcaretto>csun"edu <94< Fax: <-<"=77"70=2
$ntroduction to orthoonal eienfunction expansions July 7, 2004, L" 0" Caretto ;ae 92
#ppendi*
'he various sections in this appendix contain !ac(round material on the topics covered in these
notes"
#ppendi* # > Po!er Series )<aylor Series+
;o%er series or 'aylor series are important tools !oth in theoretical analysis and in numerical
analysis" 8ecall that the 'aylor series for a function of one varia!le, f#x&, expanded a!out some
point, x 6 a, is iven !y the infinite series,
.... ) - (
! 3
1
) - (
! 2
1
) ( ) ( ) (
3
3
3
2
2
2
+ + + +

a x
dx
f d
a x
dx
f d
a x
dx
df
a f x f
a x a x
a x
4.3-5
'he ?x 6 a@ su!script on the derivatives reinforces the fact that these derivatives are evaluated at
the expansion point, x 6 a" 1e can %rite the infinite series usin a summation notation as
follo%s:

0
) - (
!
1
) (
n
n
a x
n
n
a x
dx
f d
n
x f
4.325
$n the e)uation a!ove, %e use the definitions of 0K 6 -K 6 - and the definition of the Heroth
derivative as the function itself" I.e", d
0
fDdx
0
Ux6a 6 f#a&"
$f the series is truncated after some finite num!er of terms, say m terms, the omitted terms are
called the remainder in mathematical analysis and the truncation error in numerical analysis"
'hese omitted terms are also an infinite series" 'his is illustrated !elo%"
error Trncation !ed Term!
a x
dx
f d
n
a x
dx
f d
n
x f
m n
n
a x
n
n m
n
n
a x
n
n
+
+


+

1 0
) - (
!
1
) - (
!
1
) (
4.395
$n this e)uation the second sum represents the remainder or truncation error, Wm, %hen only m
terms are used in the partial sum for the series"

+


1
) - (
!
1
m n
n
a x
n
n
m
a x
dx
f d
n
4.345
'he theorem of the mean can !e used to sho% that the infinite3series truncation error can !e
expressed in terms of the first term in the truncation error, that is
1
1
1
) - (
! ) 1 (
1
+

+
+
+

m
x
m
m
m
a x
dx
f d
m
4.3,5
Aere the su!script, ?x 6 X@, on the derivative indicates that this derivative is no loner evaluated at
the (no%n point x 6 a, !ut is to !e evaluated at x 6 X, an un(no%n point !et%een x and a" 'hus,
b x
a x
b
a
x d
x dJ
x x J x xdx x J

1
1
]
1

,
_

,
_

2
2 2
2
2
2
) (
) (
) (
2
1
) (
$ntroduction to orthoonal eienfunction expansions July 7, 2004, L" 0" Caretto ;ae 99
the price %e pay for reducin the infinite series for the truncation error to a sinle term is that %e
lose the certainty a!out the point %here the derivative is evaluated" $n principle, this %ould allo%
us to compute a !ound on the error !y findin the value of X, !et%een x and a, that made the
error computed !y e)uation 4.3,5 a maximum" $n practice, %e do not usually (no% the exact
functional form, f#x&, let alone its #mE-&
th
derivative" 'he main result provided !y e)uation 4.3,5
for use in numerical analysis is that the remainder #or truncation error& depends on the step siHe,
x a, raised to the po%er mE-, %here m is the num!er of terms in the partial sum"
$n the a!ove discussion %e have not considered %hether the series %ill actually represent the
function" . series is said to !e convergent in a certain reion, a t 8, called the convergence
interval, if %e can find a certain num!er of terms, m, such that the remainder is less than any
positive value of a small )uantity " 'he half3%idth of the reion, 8, is called the radius of
convergence" . real function, f#x&, is called analytic, at a point x 6 a, if it can !e represented !y
a po%er series in x a, %ith a radius of converence, 8 J 0"
For simplicity %e can represent the coefficient in a 'aylor series as a sinle coefficient, !n" 'his
definition and the resultin form for the po%er series are sho%n !elo%"

0
) - ( ) (
!
1
n
n
n
a x
n
n
n
a x b x f that !o
dx
f d
n
b
4.3=5
1e have the follo%in results for operations on po%er series" Consider t%o series,

0
) - ( ) (
n
n
n
a x b x f
and

0
) - ( ) (
n
n
n
a x c x g
" 'he sum of the t%o series is

+ +
0
) - )( ( ) ( ) (
n
n
n n
a x c b x g x f 4.375
'he product of the t%o series re)uires a dou!le summation"
+ + + + + +

,
_

,
_

2
0 2 1 1 2 0 0 1 1 0 0 0
0 0 0 0
) - )( ( ) - )( (
) - ( ) - ( ) - ( ) - ( ) ( ) (
a x c b c b c b a x c b c b c b
a x c a x b a x c a x b x g x f
n m
m
m
n
n
m
m
m
n
n
n
4.3<5
1e can differentiate a series term !y term"

0
2
2
2
0
1
) - ( ) 1 (
) (
) - (
) (
n
n
n
n
n
n
a x b n n
dx
x f d
a x nb
dx
x df
4.3G5
#ppendi* $ > "actorials
$n the discussion a!ove, %e have introduced the concept of the factorial of a num!er" 'his is
defined as follo%s"
) 1 )( 2 )( 3 ( ) 2 )( 1 ( ! n n n n
From this definition %e can see that 2K 6 #2&#-& 6 2C 9K 6 #9&#2&#-& 6 =C 4K 6 #4&#9&#2&#-& 6 24" 1e
see that hiher factorials can !e computed in terms of smaller factorials" 'he examples a!ove
sho% that 4K 6 4#9K& and 9K 6 9#2K&" $n eneral, %e can %rite that
$ntroduction to orthoonal eienfunction expansions July 7, 2004, L" 0" Caretto ;ae 94
n
n
n or n n n
!
)! 1 ( )! 1 ( !
From this definition, %e can see %hat the values for -K and 0K are"
1
1
! 1
)! 1 1 ( ! 0 1
2
! 2
)! 1 2 ( ! 1
!
)! 1 ( and
n
n
n
1e also notice that %hen %e have ratios of similar factorials %e can reduce them as follo%s"
) 1 )( 2 (
!
! ) 1 )( 2 (
!
)! 1 )( 2 (
!
)! 2 (
+ +
+ +

+ +

+
n n
n
n n n
n
n n
n
n
#ppendi* C > Gamma "unctions
'he amma function, #x& is defined !y the follo%in interal"

0
1
) ( dt t e x
x t
4C3-5
$n this definition, t is a dummy varia!le of interation and the arument of the amma function, x,
appears only in the term t
x3-
" 'he amma function can !e sho%n to !e a eneraliHation of
factorials" 'o do this %e use e)uation 4C3-5 to evaluate #xE-&" 'o do this, %e simply replace x
!y x E - every%here in that e)uation and interate !y parts to o!tain
( ) ( )



+

+
vd v dv
t d e t e e d t dt t e x
x t x t t x x t
0
0
0 0
1 1
) ( ) 1 (
4C325
$f %e restrict the value of x to !e reater than Hero, the term e
3t
t
x
vanishes at !oth the upper and
lo%er limit" #.t the upper limit, %e have to apply l2Aopital2s rule to et this result"& 'he d#t
x
& term
in the interal is simply xt
x3-
dt and %hen %e perform this differentiation %e see that the resultin
interal is simply x#x&"
( ) ( ) ) ( 0 ) 1 (
0
1
0
0
x x dt t e x t d e t e x
x t x t x t
+

4C395
'he amma function for an arument of - is particularly simple to compute"
1 ) 1 0 ( ) 1 (
0
0 0
1 1

t t t
e dt e dt t e
4C345
1e can apply e)uation 4C395 to this result to o!tain the value of the amma function for 2, 9, and
4 as follo%s"
6 ) 3 ( 3 ) 1 3 ( ) 4 ( 2 ) 2 ( 2 ) 1 2 ( ) 3 ( 1 ) 1 ( 1 ) 1 1 ( ) 2 ( + + +
4C3,5
1e see that the eneral result of e)uation 4C3,5 is
$ntroduction to orthoonal eienfunction expansions July 7, 2004, L" 0" Caretto ;ae 9,
! ) ( ) 1 ( n n n n +
4C3=5
1e can apply e)uation 4C395 and the result that #-& 6 - to compute #0&"

+
+ ) 0 (
) 1 (
) ( ) ( ) 1 ( that !o
x
x
x x x x 4C375
7xtendin this relationship to neative inteers tells us that the value of the amma function for
neative inteers is infinite" Ao%ever, the value of the amma function for and noninteer
num!er, includin neative ones, may !e found from e)uation 4C395, once the values of the
amma function are (no%n in the interval !et%een Hero and one" Fumerical methods are
enerally re)uired to find these noninteer values of the amma function" Bne exception to this
is the value of the amma function of one3half that can !e found !y contour interation in the
complex plane" 'he result of such interation is

,
_

2
1
4C3<5
Fote that %e can use e)uation 4C395 to et other values of the amma function such as the
follo%in"

,
_

,
_

,
_

2
2
1
4
3
2
5
2 2
3
4C3G5