1

A Brief Review of Elementary Ordinary

Differential Equations

At various points in the material we will be covering, we will need to recall and use material
normally covered in an elementary course on ordinary differential equations. In these notes, we
will very briefly review the main topics that will be needed later. For more complete discussions
of these topics, see your current text on differential equations (by Boyce and DiPrima), or your
old introductory text on ordinary differential equations, or see your instructor’s treatment of these
topics.1

1.1 Basic Terminology

Recall: A differential equation (often called a “de”) is an equation involving derivatives of an
unknown function. If the unknown can be assumed to be a function of only one variable (so the
derivatives are the “ordinary” derivatives from Calc. I), then we say the differential equation is
an ordinary differential equation (ode). Otherwise, the equation is a partial differential equation
(pde). Our interest will just be in odes. In these notes, the variable will usually be denoted by x
and the unknown function by y or y(x) .
Recall, also, for any given ordinary differential equation:

1. The order is the order of the highest order derivative of the unknown function explicitly
appearing in the equation.

2. A solution is any function (or formula for a function) that satisfies the equation.

3. A general solution is a formula that describes all solutions to the equation. Typically, the
general solution to a k th order ode contains k arbitrary/undetermined constants.

4. Typically, a “differential equation problem” consists of a differential equation along with

some auxiliary conditions the solution must also satisfy (e.g., “initial values” for the
solution). In practice you usually find the general solution first, and then choose values
for the “undetermined constants” so that the auxiliary conditions are satisfied.

1 available online at www.math.uah.edu/howell/DEtext.

1/3/2013 1
Review of Elementary ODEs 2

1.2 Some “Analytic” Methods for Solving First-Order

ODEs
(Warning: Here, the word “analytic” just means that the method leads to exact formulas for
solutions, as opposed to, say, a numerical algorithm that gives good approximations to particular
solutions at fixed points. Later in this course, the word “analytic” will mean something else.)

Separable Equations∗
A first-order ode is separable if it can be written as
dy
= g(x)h(y) .
dx
Such a de can be solved by the following procedure:

1. Get it into the above form (i.e., the derivative equaling the product of a function of x
(the g(x) above), with a function of y (the above h(y) ).
2. Divide through by h(y) (but also consider the possibility that h(y) = 0 ).
3. Integrate both sides with respect to x (don’t forget an arbitrary constant).
4. Solve the last equation for y(x) .

!◮Example 1.1: Consider finding the general solution to

dy
= 2x y 2 + 1


.
dx
Going through the above steps:
1 dy
= 2x
y2 + 1 dx
Z Z
1 dy
H⇒ dx = 2x d x
y 2 + 1 dx

H⇒ arctan(y) = x 2 + c

y = tan x 2 + c


H⇒ .

Linear Equations†
A first-order ode is said to be linear if it can be written in the form
dy
+ p(x)y = q(x)
dx
where p(x) and q(x) are known functions of x . Such a differential equation can be solved by
the following procedure:
∗ see, also, chapter 4 of the online text
† see, also, chapter 5 of the online text

version: 1/3/2013
Review of Elementary ODEs 3

1. Get it into the above form.

2. Compute the integrating factor
R
p(x) dx
µ(x) = e
(don’t worry about arbitrary constants here).
3. (a) Multiply the equation from the first step by the integrating factor.
(b) Observe that, by the product rule, the left side of the resulting equation can be
d
rewritten as dx [µy] , thus giving you the equation
d
[µ(x) y(x)] = µ(x) q(x) .
dx
4. Integrate both sides of your last equation with respect to x , and solve for y(x) . Don’t
forget the arbitrary constant.

!◮Example 1.2: Consider finding the general solution to

dy
x + 4y = 21x 3 .
dx
Dividing through by x gives
dy 4
+ y = 21x 2 .
dx x
So the integrating factor is
R R
4/x dx
µ(x) = e p(x) dx
= e = e4 ln x = x 4 .
Multiplying the last differential equation above by this integrating factor and then continuing
as described in the procedure:
 
4 dy 4
+ y = x 4 21x 2
 
x
dx x

dy
H⇒ x4 + 4x 3 y = 21x 6 .
dx
But, by the product rule,
d  4 dy
x y(x) = x 4 + 4x 3 y

,
dx dx
and so we can rewrite our last differential equation as
d  4
x y(x) = 21x 6

.
dx
This can be easily integrated and solved:
Z Z
d  4
21x 6 d x

x y(x) d x =
dx

H⇒ x 4 y(x) = 3x 7 + c

3x 7 + c
H⇒ y(x) =
x4
H⇒ y(x) = 3x 3 + cx −4 .

version: 1/3/2013
Review of Elementary ODEs 4

Two notes on this method:

1. The formula for the integrating factor µ(x) is actually derived from the requirement that
d dy dµ dy
[µ(x) y(x)] = µ + y = µ + µpy ,
dx dx dx dx
which is the “observation” made in step 3b of the procedure. This means that µ must
satisfy the simple differential equation
dµ
= µp .
dx
2. Many texts state a formula for y(x) in terms of p(x) and q(x) . The better texts also
state that memorizing and using this formula is stupid.

Other Methods
Other methods for solving first-order ordinary differential equations include the integration of
exact equations, and the use of either clever substitutions or more general integrating factors to
reduce “difficult” equations to either separable, linear or exact equations. See a good de text if
you are interested.

1.3 Higher-Order Linear Differential Equations

Basics‡
An N th order differential equation is said to be linear if it can be written in the form

a0 y (N ) + a1 y (N −1) + · · · + a N −2 y ′′ + a N −1 y ′ + a N y = f

where f and the ak ’s are known functions of x (with a0 (x) not being the zero function). The
equation is said to be homogeneous if and only if f is the zero function (i.e., is always 0 ).
Recall that, if the equation is homogeneous, then we have “linearity”, that is, whenever y1
and y2 are two solutions to a homogeneous linear differential equation, and a and b are any two
constants, then y = ay1 + by2 is another solution to the differential equation. In other words,
the set of solutions to a homogeneous linear differential equations is a vector space of functions.
(Isn’t it nice to see vector spaces again?)
Recall further, that

1. The general solution to an N th order linear homogeneous ordinary differential equation

is given by
y(x) = c1 y1 (x) + c2 y2 (x) + · · · + c N y N (x)
where the ck ’s are arbitrary constants and

{y1 , y2 , . . . , y N }
‡ see, also, chapter 12, sections 1 – 3, and chapter 14 of the online text

version: 1/3/2013
Review of Elementary ODEs 5

is a linearly independent set of solutions to the homogeneous de. (i.e., {y1 , y2 , . . . , y N }

is a basis for the N -dimensional space of solutions to the homogeneous differential
equation.)

2. A general solution to an N th order linear nonhomogeneous ordinary differential equation

is given by
y(x) = y p (x) + yh (x)
where y p is any particular solution to the nonhomogeneous ordinary differential equation
and yh is a general solution to the corresponding homogeneous ode.

In “real” applications, N is usually 1 or 2 . On rare occasions, it may be 4 , and, even

more rarely, it is 3 . Higher order differential equations can arise, but usually only in courses on
differential equations. Do note that if N = 1 , then the differential equation can be solved using
the method describe for first order linear equations (see page 2).

Notes About Linear Independence

Recall that a set of functions
{y1 (x), y2 (x), . . . , y N (x)}
is linearly independent if and only if none of the yk ’s can be written as a linear combination of
the other yk ’s . There are several ways to test for linear independence. The one usually discussed
in de texts involves the corresponding Wronskian W (x) , given by

y1 y2 y3 ··· yn
y1 ′ y2 ′ y3 ′ ··· yn ′
W = y1 ′′ y2 ′′ y3 ′′ ··· yn ′′ .
.. .. .. .. ..
. . . . .
y1 (n−1) y2 (n−1) y3 (n−1) · · · yn (n−1)

The test is that the set of N solutions

{y1 (x), y2 (x), . . . , y N (x)}

to some given N th -order homogeneous linear differential equation is linearly independent if and
only if
W (x0 ) 6= 0
for any point in the interval over which these yk ’s are solutions.
This is a highly recommended test when N > 2 , but, frankly, it is silly to use it when
N = 2 . Then, we just have a pair of solutions

{y1 (x), y2 (x)}

and any such pair is linearly independent if and only if neither function is a constant multiple of
each other, and THAT is usually obvious upon inspection of the two functions.

version: 1/3/2013
Review of Elementary ODEs 6

Second-Order Linear Homogeneous Equations with

§
Constant Coefficients
Consider a differential equation of the form

ay ′′ + by ′ + cy = 0

where a , b , and c are (real) constants. To solve such an equation, assume a solution of the
form
y(x) = er x
(where r is a constant to be determined), and then plug this formula for y into the differential
equation. You will then get the corresponding characteristic equation for the de,

ar 2 + br + c = 0 .

Solve the characteristic equation. You’ll get two values for r ,

√
−b ± b2 − 4ac
r = r± =
2a
(with the possibility that r+ = r− ). Then:
1. If r+ and r− are two distinct real values, then the general solution to the differential
equation is
y(x) = c1 er+ x + c2 er− x
where c1 and c2 are arbitrary constants.
2. If r+ = r− , then r+ is real and the general solution to the differential equation is

y(x) = c1 er+ x + c2 xer+ x

where c1 and c2 are arbitrary constants. (Note: The c2 xer+ x part of the solution can be
derived via the method of “reduction of order”.)
3. If r+ or r− is complex valued, then they are complex conjugates of each other,

r+ = α + iβ and r− = α − iβ

for some real constants α and β . The general solution to the differential equation can
then be written as
y(x) = c1 e(α + iβ)x + c2 e(α − iβ)x
where c1 and c2 are arbitrary constants. However, because

e(α ± iβ)x = eαx [cos(β x) ± i sin(β x)] ,

the general solution to the differential equation can also be written as

y(x) = C1 eαx cos(β x) + C2 eαx sin(β x)

where C1 and C2 are arbitrary constants. In practice, the later formula for y is usually
preferred because it involves just real-valued functions.
§ see, also, chapter 16 of the online text

version: 1/3/2013
Review of Elementary ODEs 7

!◮Example 1.3: Consider

y ′′ − 4y ′ + 13y = 0 .
Plugging in y = er x , we get

d2  r x  d  rx
+ 13 er x = 0
 
2
e − e
dx dx
H⇒ r 2 er x − 4r er x + 13er x = 0

H⇒ r 2 − 4r + 13 = 0 .

Thus, p √
−(−4) ± (−4)2 − 4 · 13 4 ± −36
r = = = 2 ± 3i .
2 2
So the general solution to the differential equation can be written as

y(x) = c1 e(2+3i )x + c2 e(2−3i )x

or as
y(x) = C1 e2x cos(3x) + C2 e2x sin(3x) ,

with the later formula usually being preferred.

¶
Second-Order Euler Equations
A second-order Euler equation2 is a differential equation that can be written as

ax 2 y ′′ + bx y ′ + cy = 0

where a , b , and c are (real) constants. To solve such an equation, assume a solution of the
form
y(x) = x r
(where r is a constant to be determined), and then plug this formula for y into the differential
equation, and solve for r .
With luck, you will get two distinct real values for r , r1 and r2 , in which case, the general
solution to the differential equation is

y(x) = c1 x r1 + c2 x r2

where c1 and c2 are arbitrary constants.

With less luck, you only complex values for r , or only one value for r . See chapter 16 of
the online text to see what to do in these cases.

!◮Example 1.4: Consider

x 2 y ′′ + x y ′ − 9y = 0 .

¶ see, also, chapter 19 of the online text

2 also called a Cauchy-Euler equation

version: 1/3/2013
Review of Elementary ODEs 8

Plugging in y = x r , we get

d2  r  d  r
x2 x − 9 xr = 0
 
2
x + x
dx dx
x 2 r (r − 1)x r −2 + x r x r −1 − 9 x r = 0
     
H⇒

H⇒ r 2 x r − r x r + r x r − 9x r = 0

H⇒ r2 − 9 = 0

H⇒ r = ±3 .

So the general solution to the differential equation is

y(x) = c1 x 3 + c2 x −3 .

Other Methods
For solving more involved homogeneous second-order odes, there is still the method of Frobenius
(which we will later discuss in some detail). You may also want to look up the method of
reduction of order in your old differential equation text or chapter 13 of the online text (or see
the last problem in the next homework list).
For solving nonhomogeneous second-order odes, you may want to recall the methods of
“undetermined coefficients” (aka the “method of guess”) and “variation of parameters”, described
in chapters 21 and 23 of the online text.

version: 1/3/2013
Review of Elementary ODEs 9

Additional Exercises

1.1. In this set, all the differential equations are first-order and separable.
a. Find the general solution for each of the following:
dy dy
i. = x y − 4x ii. = 3y 2 − y 2 sin(x)
dx dx
dy dy y
iii. = x y − 3x − 2y + 6 iv. =
dx dx x
b. Solve each of the following initial-value problems.
dy
i. − 2y = −10 with y(0) = 8
dx
dy
ii. y = sin(x) with y(0) = −4
dx
dy
iii. x = y2 − y with y(1) = 2
dx
1.2. In this set, all the differential equations are linear first-order equations.
a. Find the general solution for each of the following:
dy dy
i. + 2y = 6 ii. + 2y = 20e3x
dx dx
dy dy
iii. = 4y + 16x iv. − 2x y = x
dx dx
b. Solve each of the following initial-value problems:
dy
i. + 5y = e−3x with y(0) = 0
dx
dy
ii. x + 3y = 20x 2 with y(1) = 10
dx
dy
 
π
iii. x = y + x 2 cos(x) with y =0
dx 2

1.3. Find the general solution to each of the following second-order linear equations with
constant coefficients. Express your solution in terms of real-valued functions only.
a. y ′′ − 9y = 0 b. y ′′ + 9y = 0
c. y ′′ + 6y ′ + 9y = 0 d. y ′′ + 6y ′ − 9y = 0
e. y ′′ − 6y ′ + 9y = 0 f. y ′′ + 6y ′ + 10y = 0
g. y ′′ − 4y ′ + 40y = 0 h. 2y ′′ − 5y ′ + 2y = 0

1.4. Solve the following initial-value problems:

a. y ′′ − 7y ′ + 10y = 0 with y(0) = 5 and y ′ (0) = 16
b. y ′′ − 10y ′ + 25y = 0 with y(0) = 1 and y ′ (0) = 0
c. y ′′ + 25y = 0 with y(0) = 4 and y ′ (0) = −15

version: 1/3/2013
Review of Elementary ODEs 10

1.5. Find the general solution to each of the following Euler equations on (0, ∞) :
a. x 2 y ′′ − 5x y ′ + 8y = 0 b. x 2 y ′′ − 2y = 0
c. x 2 y ′′ − 2x y ′ = 0 d. 2x 2 y ′′ − x y ′ + y = 0

1.6. Solve the following initial-value problems involving Euler equations:

a. x 2 y ′′ − 6x y ′ + 10y = 0 with y(1) = −1 and y ′ (1) = 7

b. 4x 2 y ′′ + 4x y ′ − y = 0 with y(4) = 0 and y ′ (4) = 2

1.7. Find the general solution to each of the following:

a. y ′ + 2y = 3

b. x y ′ + 2y = 8
1 2
c. y ′ + y = 2e x
x
d. y ′′ + a 2 y = 0 where a is a positive constant

e. y ′′ − a 2 y = 0 where a is a positive constant

f. y ′′ + 4y ′ − 5y = 0

g. y ′′ − 6y ′ + 9y = 0

h. x 2 y ′′ − 6x y ′ + 10y = 0

i. x 2 y ′′ − 9x y ′ + 25y = 0 (See following note)

Note: For the last one, start by assuming y = x r as described for Cauchy-Euler
equations. This will lead to one solution. To find the full solution, assume

y(x) = x r v(x)

where r is the exponent just found and v(x) is a function to be determined. Plug this
into the differential equation, simplify, and you should get a relatively easy differential
equation to solve for v(x) (it may help to let u(x) = v ′ (x) at one point). Solve for
v (don’t forget any arbitrary constants) and plug the resulting formula into the above
formula for y . There, you’ve just done “reduction of order”.

version: 1/3/2013
Review of Elementary ODEs 11

Some Answers to Some of the Exercises

WARNING! Most of the following answers were prepared hastily and late at night. They
have not been properly proofread! Errors are likely!
 
1
1a i. y = 4 + A exp x 2
2
−1
1a ii. y = c − 3x − cos(x) and y = 0
1 2
1a iii. y = 3 + A exp x − 2x
2
1a iv. y = Ax
1b i. y = 5 +p3e2x
1b ii. y = − 18 − 2 cos(x)
1b iii. y = 2(2 − x)−1
2a i. y = 3 + ce−2x
2a ii. y = 4e3x + ce−2x
2a iii. y = ce4x − 4x − 1
2 1
2a iv. y = ce x −
2
1  −3x −5x

2b i. e −e
2
2b ii. 4x 2 + 6x −3
2b iii. x[sin(x) − 1]
3a. y(x) = c1 e3x + c2 e−3x
3b. y(x) = c1 cos(3x) + c2 sin(3x)
3c. y(x) = c1 e−3x +√ c2 xe−3x √
3d. y(x) = c1 e(−3+3 2)x + c2 e(−3−3 2)x
3e. y(x) = c1 e3x + c2 xe3x
3f. y(x) = c1 e−3x cos(x) + c2 e−3x sin(x)
3g. y(x) = c1 e2x cos(6x) + c2 e2x sin(6x)
3h. y(x) = c1 e2x + c2 e x/2
4a. 3e2x + 2e5x
4b. e5x − 5xe5x
4c. 4 cos(5x) − 3 sin(5x)
5a. y = c1 x 2 + c2 x 4
5b. y = c1 x 2 + c2 x −1
5c. y = c1 + c2 x 3√
5d. y = c1 x + c2 x
6a. y = 3x 5 − 4x 2
6b. y = 4x 1/2 − 16x −1/2
3
7a. + ce−2x
2
7b. 4 + cx −2
2
7c. e x + c /x
7d. c1 cos(ax) + c2 sin(ax)
7e. c1 eax + c2 e−ax
7f. c1 e x + c2 e−5x
7g. c1 e3x + c2 xe3x
7h. c1 x 2 + c2 x 5
7i. c1 x 5 + c2 x 5 ln |x|

version: 1/3/2013