Chapter 2

Second-Order Linear ODEs

2.1 Homogeneous Linear ODEs of Second Order (二階齊性線性常微分方程)

A second-order ODE is called linear if it can be written as

y ′′ + p ( x) y ′ + q ( x) y = r ( x)

and nonlinear if it cannot be written in this form.

If r ( x) = 0 , y ′′ + p ( x) y ′ + q ( x) y = r ( x) reduces to

y ′′ + p ( x) y ′ + q ( x) y = 0

and is called homogeneous. If r ( x) ≠ 0 , it is called nonhomogeneous.

※ THEOREM 1 Fundamental Theorem for the Homogeneous Linear ODE

For a homogeneous linear ODE, any linear combination of two solutions on an open
interval I is again a solution on I. In particular, for such an equation, sums and
constant multiples of solutions are again solutions.

Proof:

Let y1 and y 2 be solutions of y ′′ + p ( x) y ′ + q ( x) y = 0 on I. Then by

substituting y = c1 y1 + c 2 y 2 and its derivatives we get

y ′′ + py ′ + qy = (c1 y1 + c 2 y 2 ) ′′ + p (c1 y1 + c 2 y 2 ) + q(c1 y1 + c 2 y 2 )

= c1 y1′′ + c 2 y 2′′ + p (c1 y ′ + c 2 y 2′ ) + q(c1 y1 + c 2 y 2 )
= c1 ( y1′′ + p y1′ + q y1 ) + c 2 ( y 2′′ + p y 2′ + q y 2 ) = 0

Example:

The functions sin 2 x , cos 2 x are both solutions of the second-order linear ODE
y ′′ + 4 y = 0 . y = c1 sin 2 x + c 2 cos 2 x is also a solution.

Note that this conclusion does not hold for the non-homogeneous linear ODE.

Example: A Nonhomogeneous Linear ODE

The functions y = 1 + cos 2 x and y = 1 + sin 2 x are solutions of y ′′ + 4 y = 1 . But their

linear combination is not a solution. Neither is, for instance, 2(1 + sin 2 x) ,
3(1 + cos 2 x) + 5(1 + sin 2 x) .

1
※ Initial Value Problems. Basis. General Solution

For a second-order homogeneous linear ODE, an initial value problem consists of

the ODE and two initial conditions

y ( x0 ) = K 0 , y ′( x 0 ) = K 1

The conditions are used to determine the two arbitrary constants c1 and c 2 in a general
solution

y ( x) = c1 y1 ( x) + c 2 y 2 ( x)

A general solution of y ′′ + p ( x) y ′ + q ( x) y = 0 on an open interval I is a solution in

which y1 ( x) and y 2 ( x) are solutions on I that are not proportional, and c1 and
c 2 are arbitrary constants.

A particular solution of y ′′ + p ( x) y ′ + q ( x) y = 0 on I is obtained if we assign

specific values to c1 and c 2 .

Definition: Basis (基底)

A basis of solutions of y ′′ + p ( x) y ′ + q ( x) y = 0 on an open interval I is a pair of

linearly independent solutions on I.

※ Find a Basis if One Solution is Known

It happens quite often that one solution can be found by inspection or in some other
way. Then a second linearly independent solution can be obtained by solving
first-order ODE. This is called the method of reduction of order (降階法). We first
show this method for example and then in general.

Example:

y ′′ − 2 y ′ + y = 0

It can be shown that y1 = e x is one of the solutions. Now let y = y1u = e x u

The idea of this method is substitute

y = e x u , y ′ = e x u + e x u ′ , y ′′ = e x u + 2e x u ′ + e x u ′′

into the ODE. This gives

e x u ′′ = 0 which implies u ′′ = 0 or u = x

Hence, y 2 = e x u = xe x is another solution of the ODE.

※ General Derivation
2
 y ′′ + p ( x) y ′ + q ( x) y = 0

We assume a solution y1 on an open interval I to be known and want to find a basis.

For this we need a second linearly independent solution y 2 on I. To get y 2 , we
substitute

y = y 2 = y1u , y ′ = y1′u + y1u ′ , y ′′ = y1′′u + 2 y1′u ′ + y1u ′′

into y ′′ + p ( x) y ′ + q ( x) y = 0 . This gives

y1u′′ + 2 y1′u′ + y1′′u + p ( y1u′ + y1′u ) + qy1u = 0

Collecting terms in u′′ , u′ and u, we have

y1u′′ + u′(2 y1′ + py1 ) + u ( y1′′ + py1′ + qy1 ) = 0

This will reduce to y1u′′ + u′(2 y1′ + py1 ) = 0 since y1 is one solution. Now divide
this remaining ODE by y1 and set u′ = U , u′′ = U ′ . This leads to

 2 y′ 
U ′ +  1 + p U = 0
 y1 

This is the desired first-order ODE, the reduced ODE. Separation of variables and
integration gives

= −(2 y1′ y1 + p ) dx
dU
and ln U = −2 ln y1 − ∫ pdx
U

By taking exponents we obtain

1 − ∫ pdx
U= e
y12

Here U = u′ , so that u = ∫ U dx . Hence the desired second solution is

y2 = y1u = y1 ∫ U dx

Example:
(1 − x 2 ) y ′′ − 2 xy ′ + 2 y = 0, y1 = x
Solution:

Let y 2 = y = y1u = xu , then y ′ = u + xu ′ , y ′′ = 2u ′ + xu ′′

Substituting into the ODE gives

3
 2(1 − 2 x 2 )
x(1 − x 2 )u ′′ + 2(1 − 2 x 2 )u ′ = 0 , or U′+ U =0
x(1 − x 2 )

Solve first for U

−1 1 12 12
u ′ =2 2
U= =2 + −
x ( x − 1) x x +1 x −1

Integrating gives
1 1 x +1 1 x +1
u = − + ln , y 2 = ux = −1 + x ln
x 2 x −1 2 x −1

Example:
Reduce the order and solve the ODE yy ′′ = 4 ( y ′) 2
Solution:
du du dy du
Let y ′ = u , then y ′′ = = =u
dx dy dx dy
(alternative expression y ′′ = u ′ which does not work for non-linear ODEs)
Upon substation we have
du
yu = 4u 2
dy
Rearranging gives
du dy
=4
u y
dy
u = y′ = = cy 4 , y = (c1 + c2 x) −1 3
dx

4
2.2 Homogeneous Linear ODEs with Constant Coefficients

We shall now consider second-order homogeneous linear ODE whose coefficients a

and b are constants

y′′ + ay′ + by = 0

Try a solution of the form

y = e λx

Substituting y = e λx and its derivatives

y ′ = λ e λx and y ′′ = λ2 e λx

into the equation, we obtain

(λ2 + aλ + b)e λx = 0

Hence if λ is a solution of the characteristic equation (特徵方程) or auxiliary

equation (輔助方程)

λ 2 + aλ + b = 0

then the exponential function eλ x is a solution of the ODE with two possible constants

1 1
λ1 = (−a + a 2 − 4b ) , λ2 = (−a − a 2 − 4b )
2 2

The characteristic equation and its solutions will be basic because our derivation
shows that the functions

y1 = e λ1x and y 2 = e λ2 x

are solutions of the ODE. We can verify this by substituting both into the ODE.

From algebra we further know that the quadratic equation (二次方程) may have three
kinds of roots, depending on the sign of the discriminant a 2 − 4b , namely

(I) Two distinct real roots if a 2 − 4b > 0

(II) A real double roof if a 2 − 4b = 0

(III) Complex conjugate roots if a 2 − 4b < 0

Example: Two Distinct Real Roots

5
Solve the initial value problem

y ′′ − 2 y ′ − 3 y = 0 , y (0) = 2 , y ′(0) = 14

Solution:

λ2 − 2λ − 3 = 0 , λ = 3, − 1

General solution:

y ( x) = c1e 3 x + c2 e − x

Particular solution:

y ′( x) = 3 c1e 3 x − c2 e − x

 y (0) = c1 + c2 = 2 c = 4
ICs.  ⇒ 1
 y′(0) = 3c1 − c2 = 14 c2 =−2

y ( x) = 4e 3 x − 2e − x

Example: Real Double Root

Find a general solution for the ODE.

y ′′ + 2 y ′ + y = 0

Solution:

λ 2 + 2λ + 1 = 0 , λ = −1, − 1

y ( x) = (c1 + c2 x) e − x (show this using the reduction of order)

Example: Complex Conjugate Roots

Find a general solution for the ODE.

y ′′ + 2 y ′ + 5 y = 0

Solution:

λ2 + 2λ + 5 = 0 , λ = −1± 2i

General solution:

y ( x) = d1e ( −1+ 2i ) x + d 2 e ( −1−2i ) x = e − x (c1 sin 2 x + c2 cos 2 x)

6
2.5 Euler-Cauchy Equations (尤拉-柯西方程式)

Euler-Cauchy equations are ODEs of the form

x 2 y ′′ + axy ′ + by = 0

We substitute

y = xm

and its derivatives y ′ = mx m−1 , y ′′ = m (m − 1) x m−2 into the ODE. This gives

m (m − 1) x m + amx m + bx m = 0 .

We then have the auxiliary equation

m (m − 1) + am + b = 0

Hence y = x m is a solution of the ODE if and only if m is a root of

m (m − 1) + am + b = 0 . The roots are

(1 − a )
2
1− a
m= ± −b
2 4

※ Case I: Two distinct real roots

The solutions are

y1 ( x) = x m1 , y 2 ( x ) = x m2

The corresponding general solutions are

y ( x) = c1 x m1 + c2 x m2

Example:

Solve the Cauchy-Euler equation

x 2 y′′ + xy′ − y = 0

Solution:

Auxiliary equation:

m (m − 1) + m − 1 = m 2 − 1 = 0 , m = 1, − 1

General solution:

7
 c2
y ( x) = c1 x +
x

※ Case II: Double Root

m = (1 − a) 2 ± (1 − a) 2 4 − b shows that the auxiliary equation has a double root if

and only if (1 − a ) 2 − 4b = 0 . The Euler-Cauchy equation then has the form

x 2 y′′ + axy′ + (1 − a) 2 y 4 = 0

And a solution is y1 = x (1− a ) 2 . To obtain a second linearly independent solution, we

apply the method of reduction of the order. Starting from y2 = y1u , we obtain for u the
expression

1
u = ∫ U dx where U= exp (− ∫ p dx)
y12

a
Note that p = is taken from the ODE written in the standard form, in our case
x

a
y′′ + y′ + (1 − a ) 2 y 4 x 2 = 0
x

This in turn gives u = ln x . As a result, a general solution is

y ( x) = (c1 + c2 ln x) x m

Example:

Solve the Cauchy-Euler equation

x 2 y′′ − xy′ + y = 0

Solution:

The auxiliary equation:

m (m − 1) − m + 1 = 0 , m = 1, 1

A general solution is:

y = (c1 + c2 ln x) x

※ Case III: Complex Roots

Example:

8
Solve the Euler-Cauchy equation

x 2 y′′ − xy′ + 2 y = 0

Solution:

m (m − 1) − m + 2 = m 2 − 2m + 2 = 0 , m =1± i

General solution:

y ( x) = d1 x1+i + d 2 x1−i = x (d1 x i + d 2 x − i ) = x [c1 cos(ln x) + c2 sin(ln x)]

where x i = e ln x = e i ln x = cos(ln x) + i sin(ln x) .

i

9
2.6 Existence and Uniqueness of Solutions. Wronskian (朗斯基行列式)

In this section we shall discuss the general theory of homogeneous linear ODEs

y ′′ + p ( x) y ′ + q ( x) y = 0

with continuous, but otherwise arbitrary variable coefficients p and q. This will
concern the existence and form a general solution as well as the uniqueness of the
solution of initial value problem consisting of such an ODE and two initial conditions

y ( x 0 ) = K 0 , y ′( x 0 ) = K 1

with given K 0 and K1 .

※ THEOREM 1 Existence and Uniqueness Theorem for Initial Value Problems

If p (x) and q (x) are continuous functions on some interval I and x0 is in I, then
the initial value problem consisting of y ′′ + p ( x) y ′ + q ( x) y = 0 and y ( x 0 ) = K 0 ,
y ′( x 0 ) = K 1 has a unique solution y (x) on the interval I.

※ THEOREM 2 Linear Dependence and Independence of Solutions

Let the ODE has continuous coefficients p (x) and q (x) on an open interval I.
Then two solutions y1 and y2 on I are linearly dependent on I if and only if their
Wronskian

W ( y1 , y2 ) = y1 y2′ − y2 y1′

is 0 at some x0 in I. Furthermore, if W = 0 at an x = x0 in I, then W = 0 on I;

hence if there is an x1 in I at which W is not 0, then y1 and y2 are linearly
independent on I.

※ THEOREM 3 Existence of a General Solution

If p (x) and q (x) are continuous on an open interval I, then y ′′ + p ( x) y ′ + q ( x) y = 0

has a general solution on I.

※ THEOREM 4 A General Solution Includes All Solutions

If the ODE has continuous coefficients p (x) and q (x) on some open interval I, then
every solution y = Y (x) is of the form of

Y ( x) = C1 y1 ( x) + C2 y 2 ( x)

where y1 , y2 is any basis of solutions on I and C1 , C2 are suitable constants.

10
Hence y ′′ + p ( x) y ′ + q ( x) y = 0 does not have singular solutions (solutions not
obtainable from a general solution).

11
2.7 Nonhomogeneous ODEs

In this section we proceed from homogeneous to nonhomogeneous linear ODEs

y ′′ + p ( x) y ′ + q ( x) y = r ( x)

where r ( x) ≠ 0 . We shall see that a “general solution” is the sum of a general solution
of the corresponding homogeneous ODE

y′′ + p ( x) y′ + q ( x) y = 0

and a particular solution of y ′′ + p ( x) y ′ + q ( x) y = r ( x) .

※ Definition:

- Homogeneous solution yh (x) :

General solution of homogeneous ODE

yh ( x) = c1 y1 + c2 y2

- Nonhomogeneous solution y p (x) :

Associated with r (x) , can be obtained from the method of undetermined coefficients

- General solution y (x) :

Solution of ODE y ′′ + p ( x) y ′ + q ( x) y = r ( x)

y ( x) = yh ( x) + y p ( x)

- Particular solution:

A particular solution is obtained from y ( x) = yh ( x) + y p ( x) by assigning specific

values to the arbitrary constants c1 and c2 in yh (x) .

※ THEOREM 1 Relations of Solutions of NHODE to Those of HODE

(a) The sum of a solution y of NHODE on some open interval I and a solution ~
y of

HODE on I is a solution of NHODE on I. In particular, y ( x) = yh ( x) + y p ( x) is

a solution of NHODE on I.

(b) The difference of two solutions of NHODE on I is a solution of HODE on I.

12
Proof:

Let L[ y ] denote the left hand side of y ′′ + p ( x) y ′ + q( x) y = r ( x) .

(a) For any solution y of y ′′ + p ( x) y ′ + q ( x) y = r ( x) and ~

y of
y′′ + p( x) y′ + q( x) y = 0 on I

L[ y + ~y ] = L[ y ] + L[ ~y ] = r + 0 = r

(b) For any two solutions y and y * of y ′′ + p( x) y ′ + q ( x) y = r ( x) on I we have

L[ y − y*] = L[ y ] − L[ y*] = r − r = 0

※ THEOREM 2 A General Solution of a Nonhomogeneous ODE Includes All

Solutions

If the coefficients p (x) , q (x) , and the function r (x) are continuous on some open
interval I, then every solution of y ′′ + p ( x) y ′ + q ( x) y = r ( x) on I is obtained by
assigning suitable values to the arbitrary constants c1 and c2 in a general solution

y ( x) = yh ( x) + y p ( x) of y ′′ + p ( x) y ′ + q ( x) y = r ( x) on I.

※ Method of Undetermined Coefficients (待定係數法)

To solve the nonhomogeneous ODE or an initial value problem NHODE, we have to

solve the HODE and find any solution y p (x) of NHODE, so that we obtain a general

solution y ( x) = yh ( x) + y p ( x) . One of the methods to find y p (x) is called the method

of undetermined coefficients. This method is suitable for nonhomogeneous linear

ODEs with constant coefficients a and b

y′′ + ay′ + by = r (x)

when r (x) is an exponential function, a power of x, a cosine or sine, or sums or

products of such function. These functions have derivatives similar to r (x) itself.

※ Choice Rules for the Method of Undetermined Coefficients

(a) Basic Rules. If r (x) is one of the functions in the first column in Table 2.1,

choose y p (x) in the same line and determine its undetermined coefficients by

substituting y p and its derivatives.

13
(b) Modification Rules. If a term in your choice for y p happens to be a solution of the

homogeneous ODE corresponding to y′′ + ay′ + by =

0 , multiply your choice

of y p by x (or by x 2 if this solution corresponds to a double root of the

characteristic equation of the homogeneous).

(c) Sum Rule. If r (x) is a sum of functions in the first column of Table 2.1, choose

for y p the sum of the functions in the corresponding lines of the second column.

Table 2.1 Method of Undetermined Coefficients

Term in r (x) Choice for y p (x)

keλx Ce λx

kx n (n = 0, 1, … ) K n x n + K n−1 x n−1 +  + K1 x + K 0

k cos ω x
K cos ω x + M sin ω x
k sin ω x

keαx cos ω x
eαx ( K cos ω x + M sin ω x)
αx
ke sin ω x

Example:

Solve the ODE

y ′′ − 5 y ′ + 6 y = 4e x

Solution:

y = c1e 2 x + c2 e 3 x + 2e x

Example:

Solve the ODE

y ′′ − 5 y ′ + 6 y = e 3 x

14
Solution:

Let y p ( x) = Cxe 3 x , C = 1 .

Example:

Find a particular solution for the following ODEs.

(1) y ′′ − 3 y ′ + 2 y = 4 x 2

(2) y ′′ − 5 y ′ + 6 y = 2e 2 x

(3) y ′′ − 4 y ′ + 4 y = 2e 2 x

(4) y ′′ − 5 y ′ + 6 y = e 2 x cos 3 x

(5) y ′′ − 2 y ′ + 2 y = 2e x cos x

Solution:

(1) y p ( x) = 2 x 2 + 6 x + 7

(2) y p ( x ) = −2 xe 2 x

(3) y p ( x) = x 2 e 2 x

1 2x
(4) y p ( x) = − e (3 cos 3 x + sin 3 x)
30

(5) y p ( x) = xe x sin x

15
2.10 Solution by Variation of Parameters

We continue our discussion of nonhomogeneous linear ODEs

y ′′ + p ( x) y ′ + q ( x) y = r ( x)

To obtain y p (x) when r (x) is not too complicated, we can often use the method of

undetermined coefficients. However, since this method is too restricted, we shall

develop a method for more general r (x) . The method to be developed is called the
method of variation of parameters (參數變換法) and is credited to Lagrange. Here p,
q, r may be variable (given functions of x), but we assume that they are continuous on
some open interval I.

The idea is to start from a general solution

y h ( x) = c1 y1 ( x) + c2 y 2 ( x)

of the homogeneous version, i.e., y ′′ + p ( x) y ′ + q ( x) y = 0 on an open interval I and

to replace the constants (“the parameters”) c1 and c2 by functions u (x) and v(x) .
This in turn suggests the name of the method. We shall determine u and v so that the
resulting function

y p ( x) = u ( x) y1 ( x) + v ( x) y 2 ( x) 。

is a particular solution of the nonhomogeneous ODE. We determine u and v by

substituting y p ( x) = u ( x) y1 ( x) + v( x) y2 ( x) and its derivatives

y ′p = u ′y1 + u y1′ + v′y 2 + v y 2′

Now y p must satisfy y ′′ + p ( x) y ′ + q ( x) y = r ( x) . This is one condition for two

unknowns u and v. It seems plausible to impose a second condition (Lagrange).

u ′y1 + v′y 2 = 0

This reduces the first derivate y ′p to the simpler form

y ′p = u y1′ + v y 2′

16
By differentiation, we obtain

y ′p′ = u ′y1′ + uy1′′ + v′y 2′ + vy 2′′

We now substitute y p and its derivatives into y ′′ + p ( x) y ′ + q ( x) y = r ( x) . Collecting

terms in u and terms in v, we obtain

u ( y1′′ + py1′ + qy1 ) + v ( y 2′′ + py 2′ + qy 2 ) + u ′y1′ + v′y 2′ = r

Since y1 and y 2 are solutions of the homogeneous ODE, this reduces to

u ′y1′ + v′y 2′ = r

And the second condition

u ′y1 + v′y 2 = 0

Solving the above two equations simultaneously for u ′ and v′

y2 r yr yr
u′ = − = − 2 , v′ = 1
y1 y 2′ − y 2 y1′ W W

By integration

y2 r y1r
u = −∫ dx , v=∫ dx .
W W

Example:

Solve the following ODE

ex
y ′′ − 2 y ′ + y =
1− x

Solution:

y1 = e x , y 2 = xe x , W = y1 y 2′ − y 2 y1′ = e 2 x

y2 r x 1
u′ = − =− = 1− , u = x + ln 1 − x
W 1− x 1− x

y1r 1
v′ = = , v = − ln 1 − x
W 1− x

17
Particular solution:

y p = u y1 + v y 2 = e x ( x + ln 1 − x ) − xe x ln 1 − x

Example:

Find a particular solution for the ODE

x 2 y ′′ − xy ′ + y = x ln x

Solution:

y1 = x , y 2 = x ln x , W = x

yr x ln x (ln x) 1 1
u′ =
− 2 =
− 2
=
− (ln x) 2 , u = − (ln x) 3
W x x 3

v′ =
y1r x(ln x) ln x
= = , v=
(ln x )2
W x2 x 2

1
y p ( x) = u ( x) y1 ( x) + v( x) y2 ( x) = x(ln x) 3
6

18