Lecture 25-26

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Lecture XXV-XXVI

Abstract
We introduce second order linear differential equations and explain
what regular and singular points of these equations are. In particular,
we develop a method to derive power series solutions around a regular
point of a second order linear differential equation.

A wide class of problems in engineering and physics are modelled by linear


second order differential equations of the form

d2 y dy
P (x) 2
+ Q(x) + R(x)y = 0 (1)
dx dx
where in many cases P , Q and R are polynomial functions of the independent
variable x. Famous examples are given by the Bessel1 equation

d2 y dy
x2 2
+ x + (x2 − ν 2 )y = 0
dx dx
with ν a constant and the Legendre2 equation

d2 y
2 dy
(1 − x ) 2 − 2x + α(α + 1)y = 0,
dx dx
where α is a constant. If we divide (1) by the function P (x) we obtain the
so-called standard form of (1), namely

d2 y dy
2
+ p(x) + q(x)y = 0, (2)
dx dx
where
Q(x) R(x)
p(x) = , q(x) = .
P (x) P (x)
Definition 1 Any point x = x0 at which both functions p and q in (2)
are analytic is called an ordinary point of the differential equation (2).
A singular point of (2) is any point x at which p or q, or both are not
continuous.

Notice that if p and q are both analytic at a point x0 they will admit a Taylor
expansion in some interval centered at x0 .
1
Friedrich Bessel (1784-1846) was a German mathematician and astronomer.
2
Adrien-Marie Legendre (1752-1833) was a French mathematician.

1
Example 1 Find all singular points of the differential equation
d2 y dy y
x(x − 2) 2
+ sin x + = 0.
dx dx x + 3
First we write the above equation in its standard form
d2 y sin x dy y
2
+ + =0
dx x(x − 2) dx x(x − 2)(x + 3)
with
sin x 1
p(x) = , q(x) = .
x(x − 2) x(x − 2)(x + 3)
Singular points are x = −3, 0, 2 since they are zeros of the denominators of
p and q. Notice that
sin x
lim = 1.
x→0 x
Hence, the singular point x = 0 arises from the denominator of q.

Power series solution of a second order linear differential equation


around a regular point
Let us consider the initial value problem
d2 y dy dy
2
+ p(x) + q(x)y = 0, y(0) = y0 , = v0 .
dx dx dx x=0

Suppose that x = 0 is a regular point. Why do we not need to consider the


general case x = x0 ? Because we can always change the independent variable
x according to t = x − x0 . This transformation translates the point x0 to
the point t0 = 0. Since x = 0 is a regular point, p and q are analytic in an
interval centered at x = 0 and we have
X∞ X∞
p(x) = pn xn , q(x) = qn x n .
n=0 n=0

We look for a power series solution of (2) having the form



X
y(x) = an x n . (3)
n=0

Notice that the initial conditions imply that the first two coefficients in (3)
must be a0 = y0 and a1 = v0 . Let us compute the first and second derivative
of y by differentiating the series (3). We obtain
∞ ∞
dy X d2 y X
= nan xn−1 , = n(n − 1)an xn−2 .
dx n=0 dx2 n=0

2
Substituting these series and those for p and q into (2) we get

"∞ #"∞ # "∞ #"∞ #
X X X X X
n−2 n n−1 n n
n(n−1)an x + pn x · nan x + qn x · an x = 0.
n=0 n=0 n=0 n=0 n=0
(4)
By multiplying series and combining coefficients of like powers of x in the
three resulting series, the above equation can be written as a single power
series, namely
X∞
Cn xn = 0. (5)
n=0
A great theorem in mathematics tells us that the above series will be zero for
all values of x if each coefficient Cn is zero. This condition provides us with a
recurrence relation for the coefficients an . The next step is to understand
how can we go from equation (4) to equation (5). To this purpose we need a
bit of gymnastics in the shift of indices of summation.

Shifting the index of summation


It is simply a change of the index of summation which allows to bring a series
into a certain desired form. For instance, the series

X
cn xn+k
n=0

can be brought into the series



X
cm−k xm
m=k

by making the substitution m = n + k. Notice that the sum is exactly the


same in both cases and only the index of summation and its starting value
has been changed. Indeed, we have

X ∞
X
n+k k k+1
cn x = c0 x + c1 x + ··· = cm−k xm .
n=0 m=k

Since m is a dummy index (because the result


c0 xk + c1 xk+1 + · · ·
is not dependent on it!), we can use again the index n and simply write

X ∞
X
cn xn+k = cn−k xn .
n=0 n=k

3
Example 2 Change the index of summation so that

X
3x2 an xn−1
n=0
X
is of the form cn xn . This can be achieved by means of the following
n
procedure

• Multiply each term in the series by 3x2 . Then,



X ∞
X
3x2 an xn−1 = 3an xn+1 .
n=0 n=0

• Set m = n + 1. Then, for n = 0 we have m = 1 and



X ∞
X
3an xn+1 = 3am−1 xm .
n=0 m=1

• Since m is a dummy variable we can go back to the original index n


and write ∞ ∞
X X
3x2 an xn−1 = 3 an−1 xn .
n=0 n=1

Notice that both sums give exactly the same terms.

Now, we are ready to bring equation (4) into the form (5). To this purpose
notice that

• setting m = n − 2 we have

X ∞
X
n−2
n(n − 1)an x = (m + 2)(m + 1)am+2 xm
n=2 m=0

X
= (n + 2)(n + 1)an+2 xn .
n=0

• If m = n − 1, then

X ∞
X ∞
X
nan xn−1 = (m + 1)am+1 xm = (n + 1)an+1 xn .
n=1 m=0 n=0

4

" ∞
# " ∞
# " ∞
# " ∞
#
X X X X
pn xn · nan xn−1 = pn xn · (n + 1)an+1 xn
n=0 n=0 n=0 n=0
∞ n
!
X X
= pk (n + 1 − k)an+1−k xn .
n=0 k=0

• " ∞
# " ∞
# ∞ n
!
X X X X
qn x n · nan xn = qk an−k xn .
n=0 n=0 n=0 k=0

Putting things together we get


∞ ∞ n
!
X X X
n
(n + 2)(n + 1)an+2 x + pk (n + 1 − k)an+1−k xn +
n=0 n=0 k=0
∞ n
!
X X
qk an−k xn = 0.
n=0 k=0
n
If we add the series and collect x we obtain
∞ n
!
X X
(n + 2)(n + 1)an+2 + [pk (n + 1 − k)an+1−k + qk an−k ] xn = 0.
n=0 k=0

Let
n
X
Cn = (n + 2)(n + 1)an+2 + [pk (n + 1 − k)an+1−k + qk an−k ] .
k=0

The series will be zero if Cn = 0 for every x. Hence, we get the following
equation for the coefficients an
n
X
(n + 2)(n + 1)an+2 + [pk (n + 1 − k)an+1−k + qk an−k ] = 0 n = 0, 1, 2, · · · .
k=0
(6)
Remember that as an implication of the initial conditions a0 = y0 and a1 =
v0 . Then, for n = 0 we find that the coefficient a2 is given by
1
a2 = − (p0 v0 + q0 y0 ) .
2
For n = 1 we have
1 2 
a3 = p0 v0 + (p0 q0 + q1 )y0 + (q0 + p1 )v0
6
and so on.

5
Example 3 Solve the initial value problem
d2 y dy dy
+ (1 + x) + 2y = 0, y(0) = 1, =0
dx2 dx dx x=0

by finding a power series solution for y at x = 0. The equation is already


in the standard form with p(x) = 1 + x and q(x) = 2. These functions are
polynomials and hence analytic at x = 0. We assume a solution of the form

X
y(x) = an x n (7)
n=0

Notice that at this point we could also use formula (6) in order to derive a
recurrence relation for the coefficients an . This procedure is left as an exercise
in the problem section. Let us substitute (7) into the differential equation.
We obtain

X ∞
X ∞
X
n−2 n−1
n(n − 1)an x + (1 + x) nan x +2 an xn = 0. (8)
n=0 n=0 n=0

We first multiply the polynomial 1 + x times the series in the second term

X ∞
X ∞
X
(1 + x) nan xn−1 = nan xn−1 + nan xn .
n=0 n=0 n=0

Then, equation (8) becomes



X ∞
X ∞
X ∞
X
n(n − 1)an xn−2 + nan xn−1 + nan xn + 2 an xn = 0. (9)
n=0 n=0 n=0 n=0

To put this intoX


the form of equation (5) we need to make each of the series
be of the form Cn xn by an appropriate shift of indices. The index in the
n
first term can start at n = 2, since the terms for n = 0 and n = 1 are both
zero. Then, letting m = n − 2 we see that m can start at zero and the series
can be written as

X ∞
X ∞
X
n−2 n−2
n(n − 1)an x = n(n − 1)an x = (m + 2)(m + 1)am+2 xm .
n=0 n=2 m=0

Since m is a dummy index we have



X ∞
X
n−2
n(n − 1)an x = (n + 2)(n + 1)an+2 xn .
n=0 n=0

6
Similarly, in the second series of (9) the term for n = 0 is zero so we can
start at n = 1. Making the change of index m = n − 1 we get
X∞ X∞ X∞ X∞
nan xn−1 = nan xn−1 = (m + 1)am+1 xm = (n + 1)an+1 xn .
n=0 n=1 m=0 n=0

The last two series in (9) are already in the desired form. Now, equation (9)
becomes
X∞ X∞ X∞ ∞
X
n n n
(n + 2)(n + 1)an+2 x + (n + 1)an+1 x + nan x + 2 an x n = 0
n=0 n=0 n=0 n=0

from which we get


X∞
[(n + 2)(n + 1)an+2 + (n + 1)an+1 + (n + 2)an ] xn = 0. (10)
n=0

At this point we recognize that the coefficient Cn in equation (5) is given by


Cn = (n + 2)(n + 1)an+2 + (n + 1)an+1 + (n + 2)an .
The series (10) will be zero for all values of x if Cn = 0 for each n =
0, 1, 2, · · · , that is
(n + 2)(n + 1)an+2 + (n + 1)an+1 + (n + 2)an = 0. (11)
This equation can be solved for an+2 , the coefficient with the largest index
and we have
(n + 1)an+1 + (n + 2)an
an+2 = − .
(n + 1)(n + 2)
The above expression or equivalently equation (11) gives us the recurrence
relation for the coefficients an . This relation takes its name from the fact
that it allows to find the value of an+2 once the values of a0 , a1 , · · · , an+1 are
known. Taking into account that the initial conditions
dy
y(0) = 1, =0
dx x=0
fix a0 = 1 and a1 = 0, the recurrence relation allows us to compute as many
coefficients as desired. For example, the next three coefficients are found by
letting n to have the values 0, 1 and 2. Hence,
a1 + 2a0
n=0 =⇒ a2 = − = −1,
2
2a2 + 3a1 1
n=1 =⇒ a3 = − = ,
6 3
3a3 + 4a2 1
n=2 =⇒ a4 = − = ,
12 4

7
Therefore, the first few terms in the series for y(x)are

x3 x4
y(x) = 1 − x2 + + + ···
3 4
Soon we will see a result ensuring that the series representation we have
derived is convergent.

You might also like