Beruflich Dokumente
Kultur Dokumente
by
by
Dwijendra
Dwijendra Narain Pandey
Department of Mathematics
Indian Institute of Technology Roorkee
Roorkee 247667, Uttarakhand.
E-mail: dwij.iitk@gmail.com
Module 4: Ordinary and Singular Points.
1 Analytic function
A function f(x) defined on an interval containing the point x = x0 is called
analytic at x0 if its Taylor series
∞
X f (n) (x0 )
(x − x0 ) (1)
n=0
n!
is analytic ev
everywhere
ever
eryw
er ywhe
yw here
he re except
exce
ex cept
ce pt at x = 7 and
an x = −3,−3, implying
im that a rational
ratio
tion
on is analytic
function anal
an alyt
al ytic
yt ic except
exce
ex cept
ce pt at th
thos
osee values
os
those v s of
of x where
wher
wh ere its denominator is zero.
er zer
Ordi
Or dina
di nary
na
Ordinaryry and
and singular
sin
sin
ingula
lar points
la poi
oint
ntss
nt
Definition
finit
fin itio
it ion
io 1. A point
n 1. poi
poi nt x = x0 is called an ordinary point of the equation
oint
x′′ (t)
(t) + P (t)x′ (t) + Q(t)x(t) = 0 (3)
1
but, x = 1 is a regular singular point.
Solution: The given differential equation can be written in the form
t x(t)
x′′ (t) +x′ (t) − 2 =0
t2 − 4 t −4
t x(t)
⇒ x′′ (t) + x′ (t) − =0 (5)
(t + 2)(t − 2) (t + 2)(t − 2)
where
t −1
P (t) = and Q(t) = (6)
(t + 2)(t − 2) (t + 2)(t − 2)
aree bot
bboth
oth
ot analytic
h ana
a naly
na lyti
ly ticc at
ti at t =2 ttherefore
=2,, the
here
he refo
re fore t = 2 is
is a regular singular point.
ampl
am ple
pl
Example 2. Sh
e 2. Show
ow that
that x = 0 and
and x = -1 are singular points of
d classify the
and hem.
hem.
them.
lution: The
Solution: ggiven
The giviven differential equation can be written in the form
iv
(t2 − 1) ′ 2x(t)
x′′ (t) + x (t) + 2 =0 (8)
t2 (t+ 1) 2 t (t + 1)2
where
(t2 − 1) 2
P (t) = 2 2
and Q(t) = 2 (9)
t (t + 1) t (t + 1)2
Since, both P(t) and Q(t) are undefined at t = 0 and t =-1, they are not analytic
at these points. Hence, t = 0 and t = -1 are singular points. Also,
(t2 − 1) (t2 − 1)
(t − 0)P (t) = (t − 0) = and
t2 (t
+ 1) 2 t(t + 1)2
(t2 − 1) (t2 − 1)
(t − 0)2 Q(t) = (t − 0)2 2 2
= , (10)
t (t + 1) (t + 1)2
2
which clearly shows that t P(t) is not analytic at t = 0, therefore t = 0 is an
irregular singular point.
Again,
(t2 − 1) t−1
(t + 1)P (t) = (t + 1) = 2 and
t2 (t + 1)2 t
2
(t − 1) (t2 − 1)
(t + 1)2 Q(t) = (t + 1)2 2 = (11)
t (t + 1)2 t2
which clearly shows that both t P(t) and t2 Q(t) are analytic at t = -1, therefore
t = -1 is a regular singular point.
d2 x dx
L[x] = P (t) + Q(t)
Q ( t) + R(t)x
R(t)x = 0. (12)
dt2 dt
3
Putting the values of x, x′ and x′′ in the equation 16, we have
∞
X ∞
X ∞
X
t2 n(n − 1)an tn−2 + (3t − 1) nan tn−1 + an tn =0
n=0 n=0 n=0
∞
X ∞
X
[n(n − 1) + 3n + 1]an tn − nan tn−1 =0
n=0 n=0
∞
X X∞
[n2 + 2n + 1]an tn − nan tn−1 =0
n=0 n=0
∞
X X∞
(n + 1)2 an tn − nan tn−1 =0. (16)
n=0 n=0
a 0 − a 1 = 0 ⇒ a0 = a1 . (17)
2 a2
2a2 = 0 ⇒ a1 =
2 a1 − 2a ⇒ a2 = 2a0 . (18)
2
a3
32 a2 − 3a3 = 0 ⇒ a2 = ⇒ a3 = 3a3a2 = 6a1 = 3!
33!a
!a 0 . (19)
23
with similar
larr argument,
simila arg
arg
rgum
umen
um ent,
en t, in general
gene
ge nera
nerall the
ra the coefficient
coeffi
co of t n
entt of
efficien
effi en
(n a)2 an − (n
(n + a) (n + 1)a n+1 = 0 ⇒ an+1
1)an+1
1) n+1 = (n + 1)an = (n + 1)!a0 .
Thus
us the
the solution
sol
sol
olut
utio
ut ion
io n will
illl take
wil
wil t the following
foll
fo llow
ll owing form
ow
∞
X ∞
X
x= n!a0 tn = a0 n!tn . (20)
n=0 n=0
Forr radiuss of
of convergence
c
1 (n + 1)!
= lim = lim (n + 1) = ∞.
R n→0 n! n→0
d2 x dx
L[x] = t2 + αt + βx = 0 (21)
dt2 dt
4
where α and β are constants is known as Euler-Cauchy’s equation. We can easily
solve the above equation by changing the independent variable t by s = et and
(21) will be reduced to a new equivalent differential equation with constant
coefficients and can be easily solved.
Here we try to solve the Euler-Cauchy’s equation (21) without changing the
independent variable. We will assume at first, for simplicity, that t > 0. Let us
observe that coefficients are polynomial and if we assume that solution is also a
polynomial say x = tr then t2 x′′ and tx′ are both multiples of tr . This suggests
that we may try x = tr as a solution of (21). Computing
d r d2 r
t = rtr−1 and t = r(r − 1)tr−2
dt dt2
we have
where
ere
F ((r)
r) = (r − 1) + αr + β
rr(r
= r2 + (α
( − 1)
1)rr + β.
ncee x = tr is a solution
nc
Hence solut
sol utio
ut ion
io n of
of (21)
(21
(21
21)) if
if and
a ononly
ly if
if,, r satisfies the following quadra
quadratic
equation
uati
tion
ti known
on kn
know
own
ow n as iindicial
as ind
ndic
nd icia
ic iall equ
ia eequation
quatioionn
r2 + (α
(α − 1)r + β = 0. (23)
Thee sol
olut
ol utio
ut ions r1 , r2 of (2
io
solutions (23)
3) ar
are
1 p
r1 = − (α − 1) + (α − 1)2 − 4β
2
1 p
r2 = − (α − 1) − (α − 1)2 − 4β .
2
Now we will consider three cases depending upon the quantity (α − 1)2 − 4β is
positive, negative and zero.
Case 1. (α − 1)2 − 4β > 0. In this case equation (23) has two real, unequal
roots and thus (21) has two solutions of the form x1 (t) = tr1 and x2 (t) = tr2 .
Since r1 6= r2 so the functions tr1 and tr2 are linearly independent. Therefore,
the general solution of (21) is given as
5
and we have only one solution x(t) = tr1 of (21). to find the second linearly
independent solution we may use the method of reduction of order. Here we
would like to discuss an alternating method to have second linearly independent
solution x2 . Since the quadratic polynomial F (r) has two equal roots that is
F (r) = (r − r1 )2 , so from (22) we have
L[tr ] = (r − r1 )2 tr . (24)
Taking partial derivatives of both the sides of (24) with respect to r gives
∂ ∂ r ∂
L[tr ] = L t = (r − r1 )2 tr .
∂r ∂r ∂r
Or
[tr1 ln tt]] = 0.
L[t
L[ 0.
which
ich implies that x2 ((t)t) = tr1 ln t sa
satisfies
sati
tisfi
ti es the
sfies
sfi the equation
equation
equ ), and the ratio xx21 (t)
((21),
n (21
21),
21 (t)
valu
va
is not a constant value.lue.
lu e. Hence
Henc
He ncee we
nc we can
can taketake x2 (t)) as
as a se
second solution of (2 (21).
Thee general
al solution
solu
so luti
lu tion
ti on of (21)
(21) in the
the case
c e of o equal
equ
qual
qu al roots
root
ro ots is given by
ot
t)tr , t > 0.
x(t) = ((cc1 + c2 ln t)t
x(t)
Case (α − 1)2 − 4β
Case 3. (α β < 0. In th
this case both the roots
r1 = λ + iµ and r2 = λ − iµ
with
1−α [4β − (α − 1)2 ]1/2
λ= , µ= (26)
2 2
are unequal complex roots. Hence,
and
6
are two real valued independent solution of (21). Hence, in the case of complex
roots, the general solution of (21) is given by
d2 x dx
L[x] = s2 + αs + βx = 0, s > 0. (27)
ds2 ds
Since s is a dummy variable so renaming s as t we have the same equation as
(21).. So without loss of generality for |t| > 0, expressions of the general solution
solut
(21) are given by: