Chapter 2 Second-Order Linear ODEs
Part 1. Linear Theory and Homogeneous DE
Wang Zhongjian
Division of Mathematical Sciences
School of Physical and Mathematical Sciences
Nanyang Technological University, Singapore
Office: SPMS-MAS-05-05
[Link]
Email: [Link]@[Link]
MH3110 – Ordinary Differential Equations, Last updated: January 27, 2026
Outline
1 Concepts and Definitions
2 Linear Theory for Homogeneous DE
Existence and Uniqueness of Solution to IVP
Principle of Superposition
Linear Dependence/Independence
Wronskian and Abel’s Formula
Existence of Two LI Solutions
3 Reduction of Order
4 Homogeneous DE: Constant Coefficients
2
Concepts and Definitions
General form of a second-order equation:
y ′′ = F (x, y, y ′ )
where F is a given function. It can be linear or nonlinear.
We are concerned with the second-order linear DE:
y ′′ + p(x)y ′ + q(x)y = f (x), (1)
or
a(x)y ′′ + b(x)y ′ + c(x)y = g(x), a ̸= 0, (2)
where p, q, a, b, c and f, g are given continuous functions.
3
Homogeneous vs Nonhomogeneous
y ′′ + p(x)y ′ + q(x)y = f (x), (3)
If f (x) ̸= 0, the linear DE (24) is nonhomogeneous.
If f (x) = 0, namely for all x ∈ I,
y ′′ + p(x)y ′ + q(x)y = 0, (4)
the linear DE is homogeneous.
We also say (4) is the associated homogeneous DE of
(24). The same concepts apply to (2).
Example: The DE: xy ′′ + ex y ′ + y = 1 is non-homogeneous, and
the associated homogeneous equation: xy ′′ + ex y ′ + y = 0.
4
IVP vs. BVP
The initial value problem (IVP):
y ′′ + p(x)y ′ + q(x)y = f (x), y(x0 ) = y0 , y ′ (x0 ) = y0′ . (5)
where x0 , y0 , y0′ are given constants.
Boundary value problem (BVP):
y ′′ + p(x)y ′ + q(x)y = f (x), y(x0 ) = y0 , y(x1 ) = y1 . (6)
where x0 , x1 , y0 , y1 are given values.
In general, one shall expect any of the above problems has
a unique solution.
5
Outline
1 Concepts and Definitions
2 Linear Theory for Homogeneous DE
Existence and Uniqueness of Solution to IVP
Principle of Superposition
Linear Dependence/Independence
Wronskian and Abel’s Formula
Existence of Two LI Solutions
3 Reduction of Order
4 Homogeneous DE: Constant Coefficients
6
Preview of Main Result
Theorem
Let y1 (x) and y2 (x) be two linearly independent (LI) solutions
to the homogenous DE
y ′′ + p(x)y ′ + q(x)y = 0, (7)
where p, q are continuous on I. Then its general solution is
y(x) = c1 y1 (x) + c2 y2 (x), (8)
where c1 and c2 are two arbitrary constants.
Then in either IVP or BVP, the boundary condition decides the
constants ci .
7
In what follows, we shall discuss . . .
Linear theory itself implies the solution method: finding two
LI solutions!
We shall take multiple steps and introduce the related tools
to establish this theory including
Existence & uniqueness theorem of IVPs
Principle of superposition
Linear dependence/independence
Wronskian & Abel’s formula
Fundamental theorem on solution
8
A. Starting Point
Theorem
(Existence & Uniqueness) Let p(x), q(x), f (x) be continuous
functions on an interval I. Then for each x0 ∈ I, the IVP
y ′′ + p(x)y ′ + q(x)y = f (x), y(x0 ) = y0 , y ′ (x0 ) = y1 (9)
has a unique solution on I.
Remark: The proof of this theorem is fairly difficult. One may refer
to, for example, E. A. Coddington, An introduction to ODEs, Dover,
1961, for the proof.
9
B. Principle of Superposition
Theorem
If y1 (x) and y2 (x) are two solutions to the homogenous DE:
y ′′ + p(x)y ′ + q(x)y = 0, x ∈ I,
then any linear combinations of y1 and y2 ,
y(x) = c1 y1 (x) + c2 y2 (x),
is also a solution to the same equation, where c1 and c2 are
arbitrary constants.
Exercise. Prove this theorem.
Question: Is it true for nonhomogeneous DE?
10
C. Linear Dependence/Independence
Definition: Let y1 (x), y2 (x) be two functions defined on an
interval I.
They are said to be linearly independent on I, if the
relation
c1 y1 (x) + c2 y2 (x) = 0, ∀x ∈ I
holds only when c1 = c2 = 0.
Otherwise, they are called linearly dependent.
Note: This definition can be applied to a finite number of functions.
11
Practical Rule for Two Functions
If the ratio is not a constant but a function of x, i.e.,
y1 (x) y2 (x)
̸= constant or ̸= constant
y2 (x) y1 (x)
then y1 (x) and y2 (x) are linearly independent.
If the ratio is a constant, then they are linearly dependent.
It only works for two functions (Caution!).
Example: x2 and 5x2 are LD, but x and x2 are LI.
12
D. Wronskian
Definition
Let y1 and y2 be differentiable functions on the interval I. The
Wronskian of y1 , y2 , denoted by W (y1 , y2 )(x), is defined by
!
y1 y2
W (y1 , y2 )(x) = det = y1 y2′ − y1′ y2 . (10)
y1′ y2′
Note: By definition,
W (y1 , y2 )(x) = −W (y2 , y1 )(x).
Exercise: Compute the Wronskian of sin x and cos x.
13
Relation: Wronskian & LI/LD
Example
Let f, g be two differentiable functions on some interval I.
If there exists a point x0 ∈ I such that the Wronskian
W (f, g)(x0 ) ̸= 0,
then f, g are linearly independent.
Equivalently, if f, g are linearly dependent, then
W (f, g)(x) ≡ 0, ∀x ∈ I.
14
Caution
The converse statement is FALSE! Namely,
LI ̸⇒ ∃x0 , s.t., W (f, g)(x0 ) ̸= 0,
or equivalently,
∀x0 , s.t., W (f, g)(x0 ) = 0, ̸⇒ LD.
Counter Example: Consider the following two functions:
( (
x2 , −1 ≤ x ≤ 0, 0, −1 ≤ x ≤ 0,
f (x) = g(x) =
0, 0 ≤ x ≤ 1. x2 , 0 ≤ x ≤ 1.
Show that the Wronskian W (f, g)(x) = 0 for all x ∈ [−1, 1].
Are f and g linearly dependent?
What conclusion can we draw from this example?
15
E. Abel’s Formula
Theorem
If y1 and y2 are solutions to the homogenous DE
y ′′ + p(x)y ′ + q(x)y = 0,
where p, q are continuous on I, then the Wronskian W (y1 , y2 )(x) is
Z
W (y1 , y2 )(x) = c exp − p(x)dx , (11)
where c is a constant that depends on the pair y1 , y2 , but not on x.
From the expression (11), we infer that:
Corollary: When y1 , y2 are solutions, W (y1 , y2 )(x) is either zero
for all x ∈ I (c = 0) or else is never zero for all x ∈ I (c ̸= 0).
16
Derivation of the Abel’s formula
Notice that y1 and y2 satisfy
y1′′ + p(x)y1′ + q(x)y1 = 0,
y2′′ + p(x)y2′ + q(x)y2 = 0.
If we multiply the first equation by −y2 , multiply the second by y1 , and
add the resulting equations, we obtain
(y1 y2′′ − y1′′ y2 ) + p(x)(y1 y2′ − y1′ y2 ) = 0.
Next, we let W (x) = W (y1 , y2 )(x) and observe that
W ′ = (y1 y2′ − y1′ y2 )′ = y1 y2′′ − y1′′ y2 .
17
Then we can write the above equation as
W ′ (x) + p(x)W (x) = 0,
which is a separable equation and admits the solution
Z
W (x) = c exp − p(x)dx ,
where c is a constant independent of x, whereas depends on which
pair of solutions y1 , y2 is involved. However, since the exponential
function is never zero, W (x) is not zero for all x unless c = 0, in which
case W (x) is zero for all x. This completes the proof.
18
Example
Find the constant c in the Abel’s formula for the pair of solutions
(e−2x , e−3x ) of y ′′ + 5y ′ + 6y = 0.
19
Example
Find the constant c in the Abel’s formula for the pair of solutions
(e−2x , e−3x ) of y ′′ + 5y ′ + 6y = 0.
Solution: By the definition,
!
e−2x e−3x
W = det = −e−5x .
−2e−2x −3e−3x
On the other hand, by the Abel’s formula,
Z
W = c exp − 5dx = ce−5x .
Hence, for the pair of solutions e−2t and e−3t is −1.
Exercise
We may check that from a different pair of solutions, say, −e−2t
and 5e−3t , the constant in the Abel’s formula is different (c = 5).
19
Two solutions: LI = Wronskian is nonzero
Theorem
Let y1 and y2 be the solutions to
y ′′ + p(x)y ′ + q(x)y = 0
where p, q are continuous on I. Then
W (y1 , y2 )(x) ≡ 0, ∀x ∈ I ⇐⇒ y1 , y2 linearly dependent,
W (y1 , y2 )(x) ̸= 0, ∀x ∈ I ⇐⇒ y1 , y2 linearly independent.
20
Proof:
Comparing with previous theorem on general function f and g,
it suffices to show that:
if W (y1 , y2 )(x0 ) = 0 for some x0 ∈ I, then y1 , y2 are linearly
dependent.
Notice that W (y1 , y2 )(x0 ) is simply the determinant of the
coefficient matrix of the linear system
c1 y1 (x0 ) + c2 y2 (x0 ) = 0,
(12)
c1 y1′ (x0 ) + c2 y2′ (x0 ) = 0,
where c1 and c2 are unknowns. Because W (y1 , y2 )(x0 ) = 0, the
system (12) has nontrivial solutions (i.e., c1 and c2 not all zero).
We just pick one nontrivial solution c1 , c2 , and denote
ϕ(x) = c1 y1 (x) + c2 y2 (x).
Then ϕ is the solution of the IVP, with the initial conditions
y(x0 ) = y ′ (x0 ) = 0.
21
Recall
ϕ(x) = c1 y1 (x) + c2 y2 (x),
is the solution of the IVP, with the initial conditions
y(x0 ) = y ′ (x0 ) = 0.
While by uniqueness theorem, this problem only has the trivial
solution ϕ ≡ 0.
Thus, we have shown that there exist c1 and c2 , which are not
all zero such that
ϕ(x) = c1 y1 (x) + c2 y2 (x) ≡ 0, ∀x ∈ I.
Therefore, y1 and y2 are linearly dependent.
22
F. Existence of fundamental solutions
Theorem
The linear homogeneous equation:
y ′′ + p(x)y ′ + q(x)y = 0 (13)
where p, q are continuous on I, always has two linearly
independent solutions.
23
Proof. Let x0 be a point in I and consider the IVP:
y ′′ + p(x)y ′ + q(x)y = 0, y(x0 ) = 1, y ′ (x0 ) = 0.
By Theorem 2, this problem has a unique solution, denoted by y1 (x).
Similarly, the following IVP
y ′′ + p(x)y ′ + q(x)y = 0, y(x0 ) = 0, y ′ (x0 ) = 1,
also has a unique solution y2 (x). Since the Wronskian
W (y0 , y1 )(x0 ) = 1, y1 and y2 are linearly independent.
24
Main Result on Solution Structure
Theorem
Let y1 and y2 be two linearly independent solutions to the
homogenous DE
y ′′ + p(x)y ′ + q(x)y = 0 (14)
where p, q are continuous on I. Then every solution to this DE
is of the form
y(x) = c1 y1 (x) + c2 y2 (x), (15)
where c1 and c2 are two arbitrary constants.
25
Proof: Let Y (x) be any solution of (14), and let x0 be a given point in
I. Denote Y (x0 ) = Y0 and Y ′ (x0 ) = Y1 . We first look for c1 and c2 s.t.
Y (x) = c1 y1 (x) + c2 y2 (x), Y (x0 ) = Y0 , Y ′ (x0 ) = Y1 .
One finds that
Y0 y2′ (x0 ) − Y1 y2 (x0 ) −Y0 y1′ (x0 ) + Y1 y1 (x0 )
c1 = , c2 = , (16)
W (y1 , y2 )(x0 ) W (y1 , y2 )(x0 )
where the denominator is nonzero, since y1 and y2 are linearly
independent.
It is essential to notice that Y (x) is the solution of the initial value
problem: (14) with the initial values: y(x0 ) = Y0 and y ′ (x0 ) = Y1 . By
the existence and uniqueness Theorem 2, the solution Y (x) is unique
and of the form Y (x) = c1 y1 (x) + c2 y2 (x).
26
Example
Determine all the values of r such that xr is a solution to
x2 y ′′ + 3xy ′ − 8y = 0, x > 0.
Hence, find the general solution to this equation on (0, ∞).
27
Example
Determine all the values of r such that xr is a solution to
x2 y ′′ + 3xy ′ − 8y = 0, x > 0.
Hence, find the general solution to this equation on (0, ∞).
Solution: Insert y = xr into the given equation:
r(r − 1)xr + 3rxr − 8xr = 0 ⇒ r2 + 2r − 8 = 0.
Therefore, the roots are r = −4, 2, which means y1 = x−4 and
y2 = x2 are two solutions of the equation. Moreover, they are linearly
independent, so the general solution is y(x) = c1 x−4 + c2 x2 .
27
Exercise
Determine all values of r s.t. erx is a solution to
y ′′ − 4y ′ + 3y = 0.
Hence, find the general solution to this equation.
Key: r = 1 and r = 3, and the general solution is y(x) = c1 ex + c2 e3x .
28
Outline
1 Concepts and Definitions
2 Linear Theory for Homogeneous DE
Existence and Uniqueness of Solution to IVP
Principle of Superposition
Linear Dependence/Independence
Wronskian and Abel’s Formula
Existence of Two LI Solutions
3 Reduction of Order
4 Homogeneous DE: Constant Coefficients
29
Reduction of Order
What is it about?
Given one nonzero solution y1 (x) of
y ′′ (x) + p(x)y ′ (x) + q(x)y(x) = 0, (17)
we can use it to find a second LI solution y2 (x).
Hence we can find the general solution to (17):
y(x) = c1 y1 (x) + c2 y2 (x).
30
Details of This Technique
Recall: in order that y1 and y2 be linear independent, the ratio
y1 /y2 must be nonconstant.
Thus, we set
y2 (x) = u(x)y1 (x), (18)
and determine u(x) from the equation (17).
Plugging y2 into (17) gives
(uy1 )′′ + p(uy1 )′ + q(uy1 ) = 0.
Group the coefficients of u′′ , u′ , u :
y1 u′′ + 2y1′ + py1 u′ + y1′′ + py1′ + q u = 0
| {z }
= 0
31
Since y1 is a solution, the equation becomes
y1 u′′ + 2y1′ + py1 u′ = 0.
We deduce that u(x) satisfies the DE:
u′′ 2y ′
1
′
=− + p(x) . (19)
u y1
It is a first-order separable DE in u′ . Integrating directly leads to
Z
ln |u′ | = −2 ln |y1 | − p(x)dx + c (20)
which implies that
Z
1
R
− p(x)dx
u(x) = c e dx. (21)
y12 (x)
Then the general solution to the original equation is
y(x) = c1 y1 (x) + c2 u(x)y1 (x).
| {z }
y2 (x)
32
Summary
Theorem (Reduction of Order): Let y1 (x) ̸= 0 be a solution of
y ′′ (x) + p(x)y ′ (x) + q(x)y(x) = 0. (22)
Then a second linearly independent solution of (22) is
1 −R
Z
p(x)dx
y2 (x) = y1 (x) e dx. (23)
y12 (x)
33
Example
Find the general solution to
xy ′′ − 2y ′ + (2 − x)y = 0, x > 0,
given that one solution is y1 (x) = ex .
34
Solution: We look for a second linearly independent solution
y2 (x) = u(x)y1 (x) = u(x)ex .
Inserting it into the given DE and collecting the terms yield
x(u′′ + 2u′ + u) − 2(u′ + u) + (2 − x)u = 0
which simplifies to
xu′′ + 2u′ (x − 1) = 0.
Separating the variables yields
u′′
= 2(x−1 − 1).
u′
By integrating, we obtain
ln |u′ | = 2(ln x − x) + c,
which can be written as u′ = c1 x2 e−2x .
35
This leads to
1
u(x) = − c1 e−2x (1 + 2x + 2x2 ).
4
Taking c1 = −4, we get
y2 (x) = u(x)ex = e−x (1 + 2x + 2x2 ).
Consequently, the general solution to the given DE is
y(x) = c1 ex + c2 e−x (1 + 2x + 2x2 ).
36
Remark
We can directly use the formula (37) to find
1 −R
Z
p(x)dx
y2 (x) = y1 (x) e dx,
y12 (x)
but we need to rewrite the equation in the standard form:
2 ′ 2−x
y ′′ − y + y = 0, x > 0.
x x
Try it yourself!
37
Exercise
Given the solution y1 (x) = x2 , find a second linearly
independent solution to the DE
x2 y ′′ − 3xy ′ + 4y = 0, x > 0.
Key: y2 (x) = x2 ln x.
38
Reduction of order: Nonhomogeneous DE
Theorem: Consider the nonhomogeneuous DE:
y ′′ + p(x)y ′ + q(x)y = f (x), (24)
where p, q and f are given continuous functions on I. If y = y1 (x) is a
solution to the associated homogeneous DE:
y ′′ + p(x)y ′ + q(x)y = 0. (25)
Then y2 (x) = u(x)y1 (x) is the (general) solution of (24), where
y′ f
u′ (x) = v(x) and v ′ + 2 1 + p v = , (26)
y1 y1
with the solution
Z
1 R
p(x)dx
v(x) = 2 y1 (x)f (x)I(x)dx + C , I(x) = e .
y1 (x)I(x)
39
Exercise
Find the general solution to
x2 y ′′ + 3xy ′ + y = 4 ln x, x>0
given that one solution to the associated homogeneous DE:
x2 y ′′ + 3xy ′ + y = 0
is y1 (x) = x−1 .
Key: y(x) = u(x)y1 (x) with u(x) = 4x(ln x − 2) + c1 ln x + c2
40
Application of Linear Theory
Exercise
Consider the Cauchy-Euler equation
x2 y ′′ − (2m − 1)xy ′ + m2 y = 0, x > 0,
where m is a constant.
(a) Determine a particular solution to this equation of the form
y1 (x) = xr .
(b) Use your solution from (a) and the method of reduction of
order to obtain a second linearly independent solution, and
hence find the general solution.
Key: (a) y1 (x) = xm (b) y2 (x) = xm ln x.
41
Outline
1 Concepts and Definitions
2 Linear Theory for Homogeneous DE
Existence and Uniqueness of Solution to IVP
Principle of Superposition
Linear Dependence/Independence
Wronskian and Abel’s Formula
Existence of Two LI Solutions
3 Reduction of Order
4 Homogeneous DE: Constant Coefficients
42
DE with Constant Coefficients
We start with the simplest case:
ay ′′ + by ′ + cy = 0, a ̸= 0, (27)
where a, b, c are given real numbers. For example,
y ′′ + y = 0, 2y ′′ + 3y ′ − y = 0
Key: In order to find two LI solutions, we try if y = erx with
suitable constant r can be a possible solution!
43
Theorem
The function y(x) = erx is a solution to
ay ′′ + by ′ + cy = 0,
if and only if r is a root of the characteristic equation (CE):
ar2 + br + c = 0.
Note the corresponding of ODE and CE:
r2 r1 r0
′′ ′
a
y +b
7
y
7+ c y
= 0 ⇒ a r2 + b r + c = 0.
44
Example
Solve the following DE:
1) y ′′ + 2y ′ − 15y = 0
2) 9y ′′ − 12y ′ + 4y = 0
45
Solution:
1) CE: r2 + 2r − 15 = 0 has the roots: r1 = −5 and r2 = 3. The GS is
y(x) = c1 e3x + c2 e−5x .
2) CE: 9r2 − 12r + 4 = 0 has repeated roots: r1 = r2 = 32 . The GS is
y(x) = (c1 + c2 x)e2x/3 .
(try to validate first)
46
Solution:
1) CE: r2 + 2r − 15 = 0 has the roots: r1 = −5 and r2 = 3. The GS is
y(x) = c1 e3x + c2 e−5x .
2) CE: 9r2 − 12r + 4 = 0 has repeated roots: r1 = r2 = 32 . The GS is
y(x) = (c1 + c2 x)e2x/3 .
(try to validate first)
b
Rule: Suppose that the CE has only one root r = − 2a from which we
rx ′′ ′
get one solution y1 = e of ay + by + cy = 0. Then the second LI
solution is y2 = xerx . Why?
46
Example
Solver the following DE:
y ′′ + 4y ′ + 13y = 0
47
Example
Solver the following DE:
y ′′ + 4y ′ + 13y = 0
Solution: CE: r2 + 4r + 13 = 0 has the roots: r1,2 = −2 ± 3i.
The GS is
y(x) = (c1 cos 3x + c2 sin 3x)e−2x .
Rule: If CE has complex roots r = α ± βi (always in conjugate pair),
then we have two complex-valued solutions:
erx = e(α±βi)x = eαx (cos βx ± i sin βx). (28)
Here, we recall the Euler formula:
eit = cos t + i sin t
In general, for z = r + it, we have
ez = er+it = er eit = er (cos t + i sin t).
47
In this case, we can rewrite the general solution for
ay ′′ + by ′ + cy = 0
with two LI solutions from (28) as
y(x) = C1 e(α+βi)x + C2 e(α−βi)x
= (C1 + C2 )eαx cos βx + i(C1 − C2 )eαx sin βx
(29)
= c1 eαx cos βx + c2 eαx sin βx
= c1 cos βx + c2 sin βx eαx .
In other words, the two (real) LI solutions are the real and imaginary
parts of the complex solution:
y1 (x) = e(α+βi)x or y2 (x) = e(α−βi)x
48
Solutions of Cauchy-Euler Equations
Exercise
Consider the Cauchy-Euler equation of the form
αt2 y ′′ (t) + βty ′ (t) + γy(t) = 0, t > 0. (30)
Show that the substitution x = ln t transforms this equation into
an equation with constant coefficients, and use this technique
to solve the following equations (t > 0):
(1) t2 y ′′ − ty ′ + y = 0, (2) t2 y ′′ + 3ty ′ + 5y = 0.
Key: (1) y = (c1 + c2 ln t)t, (2) y = (c1 cos(2 ln t) + c2 sin(2 ln t))/t.
——– End of Chapter 2 (Part 1) ———–
49