Math 251 Notes

Math 251: Ordinary Differential Equation
and Partial Differential Equation
Pichkitti Bannangkoon
July 14, 2014

Contents
1 Introduction 3
2 First Order Differential Equations 12

2.1 Solving First Order Differential Equations . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.1 Separable Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.2 Method of Integrating Factor . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.3 Exact Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 The Existence and Uniqueness Theorem . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Stability of Equilibrium Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.4 Modeling with First Order Differential Equation: Mixing Problems . . . . . . . . . 36
3 Second Order Differential Equations 43

3.1 Fundamental Set of Solutions & Wronskian . . . . . . . . . . . . . . . . . . . . . . 43
3.2 Homogeneous Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.1 Two Distinct Real roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.2.2 Complex Conjugate Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2.3 Repeated Real Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.4 Summary & More Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.3 Reduction of Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4 The Existence and Uniqueness Theorem . . . . . . . . . . . . . . . . . . . . . . . . 62
3.5 More on Wronskian & Abel’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.6 Nonhomogeneous Equation & Method of Undetermined Coefficients . . . . . . . . 72
3.7 Modeling with Second Order Differential Equation: Mechanical Vibrations . . . . . 82
3.7.1 Undamped Free Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.7.2 Damped Free Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.7.3 Underdamped Forced Vibrations . . . . . . . . . . . . . . . . . . . . . . . . 89
4 Higher Order Linear Equation 95

4.1 General Theory of nth Order linear Equations . . . . . . . . . . . . . . . . . . . . . 95
4.2 Homogeneous Equations with Constant Coefficients . . . . . . . . . . . . . . . . . . 100
5 Laplace Transform 104

5.1 Definition of Laplace Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.2 Inverse Laplace Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.3 Solving Initial Value Problems with Laplace Transform . . . . . . . . . . . . . . . . 120
5.4 Step Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.5 Impulse Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
6 Systems of First Order Linear Equations 138

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.2 Review of Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.3 Solutions to Systems of First Order Linear Equations . . . . . . . . . . . . . . . . . 149
1
CONTENTS 2
6.3.1 Real and Distinct Eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . 150

6.3.2 Complex Conjugate Eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . 154
6.3.3 Repeated Real Eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.3.4 Summary Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
7 Nonlinear Differential Equations and Stability 173

7.1 Autonomous Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
7.2 Locally Linear Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
8 Partial Differential Equation 184

8.1 Boundary Value Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.1.1 BVP vs IVP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.1.2 Eigenvalues Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
8.2 Fourier Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
8.2.1 The Fourier Convergence Theorem . . . . . . . . . . . . . . . . . . . . . . . 191
8.2.2 Even and Odd Functions and their Fourier Series . . . . . . . . . . . . . . . 197
8.2.3 The Cosine and Sine Series Extensions . . . . . . . . . . . . . . . . . . . . . 201
8.3 Heat Equations & Separation of Variables . . . . . . . . . . . . . . . . . . . . . . . 208
8.3.1 Type I: Homogeneous Boundary Conditions & Separation of Variables . . . 209
8.3.2 Type II: Nonhomogeneous Boundary Conditions . . . . . . . . . . . . . . . 217
8.3.3 Type III: Bar with Both Ends Insulated . . . . . . . . . . . . . . . . . . . . 221
8.4 Wave Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
Chapter 1
Introduction
A differential equation is any equation which contains derivatives, either ordinary derivatives
or partial derivatives.
Example 1 (Newton’s Second Law of motion). If an object of mass m is moving with

acceleration a and being acted on with force F then we have a relation F = ma. Indeed it is
a differential equation by using
dv d2 u
a= or a= ,
dt dt2
where v is the velocity of the object and u is the position function of the object at time t. So
Newton’s Second Law can now be written as
dv d2 u
F (t, v) = m or F (t, u) = m .
dt dt2
Order:
The order of a differential equation is the “largest” derivative present in the differential equation.
• We write first, second and third derivatives of y as y 0 , y 00 and y 000 respectively.

• For fourth and higher derivatives of y we write y (n) as nth derivative of y.
• Make sure you see the difference between y n (means y · · · y, n times) and y (n) (means nth
derivative of y).
ODE vs PDE:
• A differential equation is called an ordinary differential equation, abbreviated by ODE,
di y
if it has ordinary derivatives in it. (when you see y (i) or i for natural number i)
dt
• Similarly, an equation is called a partial differential equation, abbreviated by PDE, if it
∂2U
has partial derivatives in it. (when you see or Utx for examples)
∂x∂t
Linear vs Nonlinear:
• A linear ordinary differential equation is any differential equation that can be written
in the form
an (t)y (n) (t) + an−1 (t)y (n−1) (t) + . . . + a1 (t)y 0 (t) + a0 (t)y(t) = g(t),
where all ai (t), g(t) are functions of t and y (i) is i-th derivative of y.
3
CHAPTER 1. INTRODUCTION 4
• If an ordinary differential equation can’t be written in the above form then it is called
non-linear or nonlinear.
To consider whether the equation linear or nonlinear, we determine only on y and its deriva-
tives since all ai (t) and g(t) can be any linear or nonlinear funcions. Also, according to the
definitions, if the followings appear in the differential equation, it is nonlinear:
• Products of y and its derivative (for examples: yy 0 , y 0 y (5) ).;
• Either functionpy or its derivative raises to any power other than first power (for exam-
ples: y 2 , (y 0 )4 , y (3) ).;
y
• Nonlinear function of y and its derivative (for examples: sin y, , ey ).
y+2
Example 2. Consider the following differential equations.
Differential equations Order ODE/PDE Linear/Nonlinear
y 00 − 3y 0 + 5 = 0 2 ODE linear
y 000 + 2 sin(t)e3t y 00 + cos(t)y = 3 3 ODE linear

3 5
d y 2 d d y
( ) − ety = 5 5 ODE nonlinear
dt3 dt dt
9
d y d12 y
(y 2 + 2y − 1) 9 − 12 = 18 12 ODE nonlinear
dt dt
t2 − 3 0 5
y (4) − 9 sin(t)e−t y 00 + (y ) = 1 4 ODE nonlinear
18 − 7t
2
∂ U ∂U
9 2 = 2 PDE
∂x ∂t
2Uxx = 3Utx − Utt 2 PDE
3
∂ U ∂U
=1+ 3 PDE
∂x2 ∂t ∂y
Example 3 (Spring 2008, Spring 2011, Summer 2010 Midterm Exam I).
For each of the differential equations below, state its order and whether it is linear or nonlinear.
Differential equations Order Linear/Nonlinear
2y 000 + ty 00 − y 0 + et y = 0 3 linear
y = y 2 − y 00 2 nonlinear
y 0 − cos(2t)y = t2 − e−5t 1 linear
y 000 + ty 0 + (cos2 y)t = t3 3 nonlinear
y 0 + t2 = y 1 linear
(1 + y 3 )y 00 + ty 0 + y = et 2 nonlinear
y 0 (t) = ty 2 (t) 1 nonlinear
y 00 (t) = ln(t)y(t) 2 linear
ty(t)y 0 (t)y 00 (t) − 1 = 0 2 nonlinear
Solution:
A solution to a differential equation is any function which satisfies the differential equation.
• Therefore, to check whether the certain function is a solution or not, we simply substi-
tute that function into y.
• If that function satisfies the equation, it is a solution. If not, it is not a solution.
• Differential equation does not necessary have the unique solution.
Example 4 (Fall 2009 Midterm Exam I).
Which function is a solution of the differential equation
(y 0 )2 − 5ty = 5t2 + 1 ?
1
(a) y(t) = t2 (b) y(t) = e5t (c) y(t) = −t (d) y(t) = − 5t
Solution Let’s consider each choice separately,

(a) Since ((t2 )0 )2 − 5t(t2 ) = 4t2 − 5t3 6= 5t2 + 1 ⇒ t2 is not a solution.
(b) Since ((e5t )0 )2 − 5t(e5t ) = 25e10t − 5te5t 6= 5t2 + 1 ⇒ e5t is not a solution.
(c) Since ((−t)0 )2 − 5t(−t) = 5t2 + 1 ⇒ −t is a solution.

1 0 2 1 1 1
(d) Since ((− 5t ) ) − 5t(− 5t )= (5t)4 + 1 6= 5t2 + 1 ⇒ − 5t is not a solution.
Example 5 (Summer 2003 Midterm Exam I).
Which equation has y1 = e−2t and y2 = e3t as two solutions?
(a) −2y 0 +3y = 0 (b) y 00 +y 0 −6y = 0 (c) −y 00 +y 0 +6y = 0 (d) 2y 00 +10y 0 −12y = 0
Solution Again let’s take a look each choice one by one,

(a) For y = e−2t ⇒ −2(e−2t )0 + 3(e−2t ) = 7e−2t 6= 0 ⇒ e−2t is not a solution.
(b) For y = e−2t ⇒ (e−2t )00 + (e−2t )0 − 6e−2t = −e−2t 6= 0 ⇒ e−2t is not a solution.
(c) For y = e−2t ⇒ −(e−2t )00 + (e−2t )0 + 6(e−2t ) = 0 ⇒ e−2t is a solution.

For y = e3t ⇒ −(e3t )00 + (e3t )0 + 6(e3t ) = 0 ⇒ e3t is a solution.
(d) For y = e−2t ⇒ 2(e−2t )00 + 10(e−2t )0 − 12e−2t = −30e−2t 6= 0 ⇒ e−2t is not a
solution.
Direction Field:
• Direction field is a tool to study the behavior of the solutions of the 1st order differential
equation
y 0 = f (t, y)
without solving it. To draw the direction field, do the following steps:
1. Draw ty-lane.
2. For each point (t0 , y0 ), evaluate the value of f (t0 , y0 ) = y 0 |t=t0 ,y=y0 .
3. Draw a short line at (t0 , y0 ) representing the slope of line tangent to the solution passing
(t0 , y0 ).
4. Repeat steps 2. and 3. with other points on the plane.
• Curves passing through the arrows are called integral curves.
• Direction field and integral curves is used to find information about the long term behavior
of the solution ( lim y(t)) without knowing solution of the differential equation.
t→∞
• The first order differential equation in the form

dy
= f (y), (f is free from independent variable t)
dt
is called autonomous.
• Equilibrium solutions are constant functions y satisfying y 0 = 0. (for autonomous equa-
tion it means the function y for which f (y) = 0).
Example 6. Draw the direction field of y 0 = y − t.
Next let’s consider the automous equation, in which its direction filed is easier to be drawn since
f is free from independent variable t.
Example 7. Find the equilibrium solution(s), draw the direction field and determine the
behavior of y as t → ∞.
1. y 0 = 3 − 2y
2. y 0 = 1 + 2y
Solution
1. For y 0 = 3 − 2y, the equilibrium solution is y such that y 0 = 3 − 2y = 0 ⇒ y = 23 .
3
All solutions appear to converge to the equilibrium solution y(t) = 2 as t → ∞. In other
3
words, lim y(t) = .
t→∞ 2
2. For y 0 = 1 + 2y, set y 0 = 0 ⇒ 1 + 2y = 0 ⇒ y = − 12 . So y = − 21 is the
equilibrium solution.
In this case, all solutions diverge away from the equilibrium solution y(t) = − 21 as
t → ∞. So lim y(t) does not exist.
t→∞
Exercises 1
1. Classify the following equations as linear or non-linear, and state their order
d2 y dy
(a) t + t2 + t3 y = cos t
dt2 dt
d3 y dy
(b) t 3 + t2 + t3 y = cos y
dt dt
dy 2y − 3
(c) =
dx 2x + 2
2. Which of the following is TRUE?
t
(a) y 0 = is a first order linear differential equation
y
(b) sin t y 00 + (1 − t2 )y 0 + cos t y = 0 is a second order linear differential equation
(c) y 00 + (y 0 )3 + y = 0 is a nonlinear differential equation of order 3
(d) y 00 + y 0 + y = t is a second order homogeneous differential equation
∂y
(e) + ty = 0 is an ordinary differential equation
∂t
3. Which of the following is a linear second order differential equation?
(a) y 0 + ty = 1
(b) y 00 = t2 y + et
(c) (y 0 )2 = (y + 2)(y − 3)
(d) y 00 + 3y 0 + 2y = sin y
4. Show that the functions y1 (t) = t2 , y2 (t) = t−2 are solutions of the differential equation
t2 y 00 + ty 0 − 4y = 0
5. Classify the following differential equations as linear or non-linear and state their order.
d2 y dy
(a) ln(t) 2
+ 3et − y sin t = 0
dt dt
(b) 2y 0 − y 2 = et
(c) y 000 + (t2 − 1)y + cos t = 0
(d) y 00 − sin(t + y)y 0 + (t2 + 1)y = 0
1
6. Are the functions y1 (t) = t3 and y2 (t) = solutions to the differential equation
t
t2 y 00 − ty 0 − 3y = 0?
7. Consider the equation (t − 1)y 00 − ty 0 + y = 0. Verify that the functions y1 (t) = t and
y2 (t) = et are its solutions.
8. Give an example of the following:
i) A first order, nonlinear, autonomous, ordinary differential equation.
ii) A second order, linear, homogeneous, ordinary differential equation.
9. Determine the differential equation whose direction field is given below.
(a) y 0 = y 2
(b) y 0 = t
(c) y 0 = y 2 + t
(d) y 0 = y 2 − t
10. Determine the differential equation whose direction field is give below.
(a) y 0 = y − x
(b) y 0 = x + y
(c) y 0 = y(y − 2)
(d) y 0 = x(x − 2)
11. Which of the following is a second order, linear, homogeneous differential equation?
(a) y 00 + y 2 = 0
(b) y 00 + y 0 + 2y = t ln y
(c) ty 00 + y = 0
(d) (y 0 )2 − t2 y = 1
12. For each of the differential equations below, state its order and whether it is linear or
nonlinear.
(a) y 0 + t2 y = et
(b) 2y 00 + 3y 0 − y = te−t
(c) ty 0 = y(y + 1)(y − 1)
(d) y 000 − 2y 0 + ty − y 2 = 0
13. Identify each of the following equations as linear or non-linear and also determine their order.
dy
(a) y =t
dt
d2 y dy
(b) ( 2 )3 + ( )3 + y 3 = t3
dt dt
dy d2 y
(c) sin t + t5 y = (1 − t2 ) 2
dt dt
3
d y
(d) (1 + y) sin2 t + ( 3 + y) cos2 t = 1
dt
14. All of the equations below have y(t) = 5e6t as a particular solution, EXCEPT
(a) y 00 − 12y 0 + 36y = 0
(b) y 00 − 5y 0 − 6y = 0
(c) y 00 + 4y 0 − 12y = 0
(d) y 00 − 36y = 0
Answers
1. a) second order and linear, b) third order and non-linear, c) first order and linear
2. b)
3. b)
4. -
5. a) second order and linear, b) first order and non-linear, c) third order and linear, d) second
order and non-linear
6. yes
7. -
8. There are infinitely many correct answers for each part. A few examples are given.
a) e.g., y 0 = y 2 or y 0 = ey , b) e.g., y 00 = 0 or y 00 + y 0 + y = 0 9. c)
10. a)
11. c)
12. a) first order and linear, b) second order and linear, c) first order and non-linear, d) third
order and non-linear
13. a) first order and non-linear, b) second order and non-linear, c) second order and linear, d)
third order and linear
14. c)
Chapter 2
First Order Differential Equations
2.1 Solving First Order Differential Equations

In this chapter, we will take a look closely at solving first order differential equations,
dy
= f (y, t).
dt
Unfortunately, there is no general formula for the solution to above equation. What we can do
instead is to consider at special cases. We will study three methods to solve first order differential
equation in this course.:
• Separable equation
• Integrating factors
• Exact equation
2.1.1 Separable Equation

A separable equation is any differential equation that we can write in the following form
N (y)dy = M (t)dt. (separate y’s and dy to t’s and dt)
In other words, it is the equation that we can separate one variable to one side of the equation
and another to the opposite side of the equation. If we fail to do that, we can’t use this method
to solve the differential equation. Note here that we always place dy and dt as the last term for
each side of the equation. Solving separable equation is very easy, you just integrate both sides,
Z Z
N (y)dy = M (t)dt.
Before we see the examples, let’s learn some new terms.

• The general solution is the solution with undetermined constant c.
• An explicit solution is any solution that is given in the form y = y(t).
• The initial value problem, abbreviated IVP, is a differential equation along with initial
condition(s). The initial condition is used to determine the value of constant c in the general
solution. The solution with determined value of c by using initial condition(s) is called the
particular solution.
12
CHAPTER 2. FIRST ORDER DIFFERENTIAL EQUATIONS 13

Find the general solution to the following differential equation
3y 0 = 4y 4 t3 .
Solution
dy
Replace y 0 to dt and then split y’s and dy to the left and t’s and dt to the right of the equation,
dy 3
3y 0 = 4y 4 t3 ⇒ 3 = 4y 4 t3 ⇒ dy = 4t3 dt.
dt y4
Integrating both sides,
Z Z
3 1
dy = 4t3 dt ⇒ − = t4 + c.
y4 y3
So the general solution (in implicit form) to the above equation is

1
− = t4 + c.
y3
From the above example, if we simplify further, we have

r
1 1 1
− 3 = t4 + c ⇒ 3
y =− 4 ⇒ y= 3
−
y t +c t4 + c
as the general solution in explicit form.

The solution of the initial value problem
x
y0 = − , y(0) = −2.
y
dy
Solution We use the same tactic here by first replacing y 0 to dx (notice here that, unlike the
previous example, the independent variable is x) and then splitting y’s and dy to the left and
x’s and dx to the right of the equation,
x dy x
y0 = − ⇒ =− ⇒ y dy = −x dx.
y dx y
y2 x2
Z Z
y dy = −x dx ⇒ = − + c.
2 2
Impose the initial condition y(0) = −2, which means, if x=0 then y = −2,
(−2)2 02
=− +c ⇒ c = 2.
2 2
Therefore the particular solution (in implicit form) to the above equation is
y2 x2
= − + 2.
2 2
Simplify a few more steps to get the solution in explicit form,
y2 x2 p
=− +2 ⇒ y 2 = −x2 + 4 ⇒ y=± −x2 + 4.
2 2
√
Unfortunately, one of the solution, namely y = −x2 + 4 fails to agree with the initial con-
dition y(0) = −2, so p
y = − −x2 + 4
is the particular solution to the above IVP in the explicit form.
After a few examples, we get one useful observation here:
We often use the Method of Separable Equation to solve first order differential equation which
contains fraction expressions. Also don’t forget to check the solutions to the IVP with the
initial condition to get the final particular solution in the explicit form.

Find the explicit solution of the initial value problems
cos t + 1
y0 = , y(0) = −3.
y
Solution Rewrite the differential equation,
cos t + 1 dy cos t + 1
y0 = − ⇒ =− ⇒ y dy = (cos t + 1) dt.
y dt y
y2
Z Z
y dy = (cos t + 1) dt ⇒ = sin t + t + c.
2
Now apply the given initial condition y(0) = −3, to determine the value of c,
(−3)2 9
= − sin(0) + 0 + c ⇒ c= .
2 2
Therefore the particular solution is
y2 9
= sin t + t + .
2 2
Then rewrite it in the explicit form,
√
y 2 = 2 sin t + 2t + 9 ⇒ y = ± 2 sin t + 2t + 9.
Since the negative sign is chosen to satisfy the initial condition y(0) = −3, the explicit solution
of the above IVP is √
y = − 2 sin t + 2t + 9.
2.1.2 Method of Integrating Factor

In order to use this method, the differential equation must be in the following form
dy
+ p(t)y = g(t), (2.1)
dt
where p and g are given functions of t. This form is called standard form or canonical form of
1st order linear equation. To solve it, we assume that there is some magical function µ(t) satisfying
µ(t)p(t) = µ0 (t). (2.2)
We called µ(t) an integrating factor. Now we change the form of the equation by multiplying
Equation (2.1) both sides by µ(t):
dy
µ(t) + µ(t)p(t)y = µ(t)g(t). (2.3)
dt
Then change the second term of the left hand side of Equation (2.3) by using Equation (2.2),
dy
µ(t) + µ0 (t)y = µ(t)g(t).
dt
Now the left hand side is simply (µ(t)y(t))0 , by product rule of differentiation, and so we get
d(µ(t)y(t))
(µ(t)y(t))0 = µ(t)g(t) ⇒ = µ(t)g(t) ⇒ d(µ(t)y(t)) = µ(t)g(t) dt.
dt
Integrating both sides of the equation,
Z Z
d(µ(t)y(t)) = µ(t)g(t) dt
Z
µ(t)y(t) = µ(t)g(t) dt + c
R
µ(t)g(t) dt + c
y(t) = ,
µ(t)
which is the general solution of Equation (2.1). Now let’s find out what the formula of the magical
function µ(t) is. From Equation (2.2),
µ0 (t)
µ(t)p(t) = µ0 (t) ⇒ = p(t).
µ(t)
Notice here that

µ0 (t)
= (ln µ(t))0
µ(t)
by product rule of differentiation. Therefore
d(ln µ(t))
(ln µ(t))0 = p(t) ⇒ = p(t) ⇒ d(ln µ(t)) = p(t) dt.
dt
Z Z Z R
p(t) dt
d(ln µ(t)) = p(t) dt ⇒ ln µ(t) = p(t) dt ⇒ µ(t) = e .
Hence the formula for the integrating factor µ(t) is

R
p(t) dt
µ(t) = e .
In conclusion, we have the following:
The process of using the method of integrating factor,

1. Rewrite the differential equation into the standard form to identify p(t),
dy
+ p(t)y = g(t).
dt
2. Find the integrating factor, µ(t), by using

R
p(t) dt
µ(t) = e .
(omit constant from this integration to make our life easier and still give the same
result)
3. Multiply both sides to the equation in step 1 (not the original equation) by µ(t).
4. Change the left hand side of the equation to (µ(t)y(t))0 by product rule of differentiation.
5. Integrate both sides of the equation.
6. Find y(t) as the general solution.

Solve the following initial value problem
ty 0 − 2y = t3 cos t, y(π) = 2.
Solution Rewrite the differential equation in canonical form,

2
y 0 − y = t2 cos t. (2.4)
t
2
So we have p(t) = − and hence the integrating factor is
t
−2
− 2t dt
R R
µ(t) = e p(t) dt
=e = e−2 ln t = eln t = t−2 .
Multiply both sides of Equation (2.4) (not the original one) by µ(t),
t−2 y 0 − 2t−3 y = cos t.
Rewrite the left hand side of the equation to (t−2 y)0 and then simplify,
d −2
(t−2 y)0 = cos t ⇒ (t y) = cos t ⇒ d(t−2 y) = cos t dt
dt
Z Z
−2
d(t y) = cos t dt ⇒ t−2 y = sin t + c ⇒ y = t2 sin t + ct2 .
Then apply the given initial condition y(π) = 2,

2
2 = π 2 sin π + cπ 2 ⇒ c=
π2

2 2
y = t2 sin t + t .
π2
2
R
From the above example, we notice that e − t dt = t−2 . The number −2 which appears on both
sides is not a coincidence. Let’s consider the general case,
k k
R
dt
e t = ek ln t = eln t = tk .
So we have the shortcut here,

k
R
dt
e t = tk
for any natural number k. Let’s practice on this method with a few more examples .

Use the integrating factor technique to solve the following equation
2y 0 (t) + y(t) = 3t.
(a) What is the integrating factor µ(t)?

Solution Rewrite the given equation to,
1 3
y 0 + y = t. (2.5)
2 2
Therefore the integrating factor is
1 t
R R
p(t) dt dt
µ(t) = e =e 2 = e2 .
(b) What is the general solution of the above differential equation?

t
Solution Multiply both sides of Equation (2.5) by µ(t) = e 2 ,
t 1 t 3 t t 3 t
e 2 y 0 + e 2 y = te 2 ⇒ (e 2 y)0 = te 2 .
2 2 2
Now change 0 to d
dt ,
d t 3 t t 3 t
(e 2 y) = te 2 ⇒ d(e 2 y) = te 2 dt.
dt 2 2
Integrating both sides and using integration-by-parts for the right hand side of the equa-
tion,
Z Z
t 3 t t 3 t t t t
d(e y) =
2 te 2 dt ⇒ e 2 y = (2te 2 − 4e 2 ) + c = 3te 2 − 6e 2 + c.
2 2
t
Divide both sides by e 2 to get the general solution,
t
y = 3t − 6 + ce− 2 .

Solve the initial value problem
π
sin(t)y 0 + cos(t)y = 3t sin(t), y( ) = 0.
2
Solution The standard form of the given equation is,
y 0 + cot(t)y = 3t. (2.6)
So the integrating factor is

R R
p(t) dt cot t dt
µ(t) = e =e = eln sin t = sin t.
Multiply both sides of Equation (2.6) by sin t,
sin(t)y 0 + cos(t)y = 3t sin(t),
which gives us the original differential equation in the question (this is just a coincidence).
Rewrite the left hand side of the equation to (sin(t)y)0 and then simplify,
d
(sin(t)y)0 = 3t sin(t) ⇒ (sin(t)y) = 3t sin t ⇒ d(sin(t)y) = 3t sin t dt.
dt
Integrating both sides (using integration-by-parts for the left hand side term),
Z Z
d(sin(t)y) = 3t sin t dt ⇒ sin(t)y = −3t cos t + 3 sin t + c.
Then apply the given initial condition y( π2 ) = 0,

π π π π
sin( )0 = −3 cos + 3 sin + c ⇒ c = −3.
2 2 2 2
Hence the particular solution is
1
y= (−3t cos t + 3 sin t − 3).
sin t
2.1.3 Exact Equation

We call the differential equation exact if we can rewrite the differential equation in the form
dy
M (x, y) + N (x, y) =0 (2.7)
dx
and there is a function ψ(x, y) which has continuous second partial derivatives such that
ψx = M (x, y) and ψy = N (x, y). (2.8)
You might ask yourself already that why we need these complicated conditions. Here’s why.
Combine Equations (2.7) and (2.8) together,
dy
ψx + ψy = 0.
dx
d
Using the chain rule, the left hand side is simply dx ψ(x, y(x)). Therefore,
d
ψ(x, y(x)) = 0.
dx
After we integrate both sides, we get
ψ(x, y(x)) = c,
as an implicit solution to our differential equation. Since second partials of ψ are continuous, by
Schwarz’s theorem, we get the symmetry of second derivatives which means
ψxy = ψyx .
However, we also have the following.
ψxy = (ψx )y = My
ψyx = (ψy )x = Nx .
Therefore, if a differential equation is exact and ψ(x, y) must satisfy
M y = Nx (2.9)
Likewise if Equation(2.9) is not true there is no way for the differential equation to be exact.
Therefore we will use this equation as a test for exactness of the equation. In other words, an
equation of the form
dy
M (x, y) + N (x, y) = 0 is an exact equation if and only if My = Nx .
dx
In conclusion, we have
The process of exact equation method

1. Rewrite the differential equation into the form
dy
M (x, y) + N (x, y) = 0,
dx
to identify M and N .
2. Check exactness of the equation by checking My = Nx . (if it is not, we can’t use this
method.)
3. After using these two conditions

Z
ψx = M ⇒ ψ= M dx;
Z
ψy = N ⇒ ψ= N dy,
compare these two equations to get ψ

4. Then the general solution is ψ(x, y) = c.
Example 14 (Spring 2005 Midterm Exam I).
a) Verify that the following ODE is exact:
dy
2x + ey + (xey − cos y) = 0.
dx
Solution First we identify M and N ,
M = 2x + ey , N = xey − cos y.
To check exactness, let’s compute My and Nx .
∂ ∂ ∂ y
My = (2x + ey ) = 2x + e = 0 + ey = ey ,
∂y ∂y ∂y
∂ ∂ ∂
Nx = (xey − cos y) = ey x − cos y = ey (1) − 0 = ey .
∂x ∂x ∂x
Since My = Nx , the equation is exact.
b) Find the general solution to the ODE.
Solution Recall that
Z Z Z Z
ψx = M ⇒ ψ = M dx = (2x + ey ) dx = 2x dx + ey dx = x2 + xey + c1 (y),
Z Z Z Z
ψy = N ⇒ ψ = N dy = (xey − cos y) dy = x ey dy − cos y dy = xey − sin y + c2 (x).
Comparing two results, c1 (y) = − sin y and c2 (x) = x2 . And so
ψ(x, y) = x2 + xey − sin y.

Therefore the general solution is
x2 + xey − sin y = c. (ψ(x, y) = c)
π
c) Find the solution to the ODE in part a) which satisfies y(1) = 2.
π
Solution Apply the initial condition, y(1) = 2,
π π π
12 + (1)e 2 − sin =c ⇒ c = e2.
2
π
Hence the particular solution is x2 + xey − sin y = e 2 .

Consider the initial value problem
dy
6x2 − 2xy + ex+y + (ex+y − x2 ) = 0, y(1) = −1.
dx
a) Verify that the equation is exact.
Solution In this example, we have
M = 6x2 − 2xy + ex+y , N = ex+y − x2 .
Compute My and Nx .
∂ ∂ ∂
My = (6x2 − 2xy + ex+y ) = 0 − 2x y + ex ey = −2x + ex+y ,
∂y ∂y ∂y
∂ x+y ∂ ∂
Nx = (e − x2 ) = ey ex − x2 = ex+y − 2x.
∂x ∂x ∂x
So the equation is exact because My = Nx .
b) Solve this IVP. You may leave your answer in implicit form.
Solution Since
Z Z
ψx = M ⇒ ψ= M dx = (6x2 − 2xy + ex+y ) dx = 2x3 − x2 y + ex+y + c1 (y),
Z Z
ψy = N ⇒ ψ = N dy = (ex+y − x2 ) dy = ex+y − x2 y + c2 (x).
Combine two equations together, c1 (y) = 0 and c2 (x) = 2x3 . And so
ψ(x, y) = 2x3 − x2 y + ex+y .
Hence the general solution is
2x3 − x2 y + ex+y = c.
Impose the initial condition, y(1) = −1,
2(1)3 − (1)2 (−1) + e1−1 = c ⇒ c = 4.
Thus the particular solution is
2x3 − x2 y + ex+y = 4.
Exercises 2.1
1. Solve the following equations and the initial value problem. You may leave your answers in
implicit form.
x2
(a) y 0 =
2y
(b) x − y + (y − x)y 0 = 0
π
(c) sin(t)y 0 + cos(t)y = 3t sin(t), y( ) = 0
2
2. Solve explicitly for y(t) in the following initial value problem
et − yy 0 = 0, y(0) = 1.
3. Find the general solution of the following equation:

(2y − xexy )y 0 = 2 + yexy .
You may leave your answer in implicit form.
4. Find the value for the constant b, for which given equation is exact.
(ex sin y + bx2 y 2 ) dx + (ex cos y + x3 y) dy = 0
(a) b = 0
1
(b) b =
3
(c) b = 3
3
(d) b =
2
(e) b = 1
5. Find the general solution of
dy
t2 + 3ty = et , t > 0.
dt
(2xy − 3x2 ) dx + (x2 + 2y) dy = 0
(you may keep your solution in implicit form).
7. Solve the initial value problem
4x − sin x
y0 = , y(0) = 1.
y−2
√
(a) y = 2 + 4x2 + 2 cos x − 1
√
(b) y = 2 − 4x2 + 2 cos x − 1
√
(c) y = 2 + 2x2 − cos x + 1
√
(d) y = 2 − 2x2 − cos x + 1
8. Consider the differential equation
yy 0 = 1 − x2 .
Answer the following questions. You must give a reason to justify each of your answers.
(a) Is the equation first order linear?
(b) Is the equation separable?
(c) Is the equation exact?
(d) Is the equation autonomous?
9. Consider the equation

3x5 y 4 − 2x + (2x6 y 3 + 6e3y )y 0 = 0.
(a) Verify that it is an exact equation.
(b) Find its general solution. You may leave your answer in implicit form.
(c) Find the particular solution satisfying the initial condition y(3) = 0.
10. What is a suitable integrating factor that can be used to solve the equation
(1 + t2 )y 0 − 2ty = 4e−πt ?
1
(a) µ(t) =
1 + t2
2
(b) µ(t) = et
(c) µ(t) = (1 + t2 )2
(d) µ(t) = e−2 arctan t
11. What is the general solution of the equation
4x3 + 2y + y cos x + (2x + sin x + 2y)y 0 = 0?
(a) 2 + cos x = C
(b) 12x2 − y sin x = C
(c) x4 + 2xy + y sin x = C
(d) x4 + 2xy + y sin x + y 2 = C
12. Find the explicit solution of the initial value problem
2
4tet
y0 = , y(0) = −4.
y
13. True or false: The function µ(t) = t2 e−3t is a suitable integrating factor that can be used
to solve the equation below. Justify your answer by finding the correct µ(t).
t2 y 0 − (3t2 − 2t)y = e4t
14. Solve, explicitly for y(t), the initial value problem
y 0 = e2y (4t − cos t), y(0) = 1.

1
(2λx5 y 3 − ) + (3x6 y 2 − 4λ)y 0 = 0.
x2
(a) Find the value of λ such that the equation becomes an exact equation.
(b) Find its general solution. You may leave your answer in implicit form.
2 − 3x2 + 8x3
y0 = , y(1) = −2.
2 + 2y
Give your solution in the explicit form.
17. Which of the following is an integrating factor for the differential equation
(x2 − 1)y 0 = −2xy + x, x > 1?

2
(a) µ(t) = e−x
(b) µ(t) = x2 − 1
2
(c) µ(t) = ex
(d) µ(t) = ln(x2 − 1)
1
(x3 y 2 + βxey + ln x) + ( x4 y − 6y sin y 2 + x2 ey )y 0 = 0.
2
Find the value β such that the above equation is exact.
(a) β = −1
(b) β = 1
(c) β = −2
(d) β = 2
19. Solve the initial value problem explicitly
xe−2x
y0 = , y(0) = 1.
2y − 4
20. What is the general solution of the following equation
cos(x + y) + 2x + (cos(x + y) + 4y)y 0 = 0?
Answers
3 2 2
x x y
1. a) y 2 = + C, b) − xy + = C, c) y = −3t cot t + 3 − 3 csc t
√3 2 2
2. y(t) = 2et − 1
3. −2x − exy + y 2 = C
4. d)
5. y(t) = t−2 et − t−3 et + Ct−3
6. x2 y − x3 + y 2 = C
7. b)
8. a) no, b) yes, c) yes, d) no
1 1
9. a) My = 12x5 y 3 = Nx , b) x6 y 4 − x2 + 2e3y = C, c) x6 y 4 − x2 + 2e3y = −7
2 2
10. a)
11. d) √
12. y = − 4et2 + 12
13. True
1
14. y = − ln(−4t2 + 2 sin t + e−2 )
2
1
15. a) λ = 3, b) x6 y 3 + − 12y = C
√ x
16. y = −1 − 2x − x3 + 2x4 − 2
17. b)
18. d)
1p
19. 2 − −2xe−2x − e−2x + 5
2
20. sin(x + y) + x2 + 2y 2 = c
2.2 The Existence and Uniqueness Theorem

So far, we have seen a number of initial value problems, each of which has the unique solution.
This observation raises a natural question:
Does every initial value problem has exactly one solution or at least has a solution?
This is important question. Because you might want to know that initial value problem has a
solution before spending time and effort in trying to find it. Luckily, for linear first order equations
with mild conditions, the answer to this question is positive and it is given by the Existence and
Uniqueness Theorem.
Theorem 1 (The Existence and Uniqueness Theorem). If the functions p and g are contin-
uous on an open interval I : α < t < β containing the point t = t0 , then there exists a unique
function y = φ(t) that satisfies the differential equation
y 0 + p(t)y = g(t)
for each t in I, and that also satisfies the initial condition y(t0 ) = y0 . In other words, neither
existence nor uniqueness of a solution is guaranteed at a discontinuity of either the coefficient
functions p(t) or g(t).
Before we see the examples, let’s learn a new terminology here.

• The interval of validity for an IVP of n th-order ODE with initial condition(s)
y(t0 ) = y0 and/or y (k) (t0 ) = yk , where k = 1, · · · , n − 1,
is the largest possible interval on which the solution is valid and contains t0 .

What is the largest interval on which the solution of
ty 0 + 2y = 3t, y(2) = 5,
is guaranteed to exist?
Solution Firstly, rewrite the given equation in the canonical form,
2
y 0 + y = 3, y(2) = 5,
t
2
with p(t) = and g(t) = 3. So p has one discontinuity at t = 0 and g is continuous
t
everywhere. By the Existence and Uniqueness Theorem , the solution of this initial value
problem guaranteed to exist uniquely on any interval containing t0 = 2 but not containing
discontinuity t = 0.
Hence the interval of validity is (0, ∞).


t t+1 1
(t2 − 16)y 0 + (sin )y = , y(π) = .
5 t−1 2
Without solving the equation, what is the largest interval in which a unique solution is guar-
anteed to exist?
Solution As usual, write down the equation in the canonical form together with the fact that
t2 − 16 = (t − 4)(t + 4), we have
sin 5t t+1 1
y0 + y= , y(π) = ,
(t + 4)(t − 4) (t − 1)(t + 4)(t − 4) 2
sin 5t t+1
with p(t) = and g(t) = .
(t + 4)(t − 4) (t − 1)(t + 4)(t − 4)
So p has discontinuities at t = −4, 4 and g has discontinuities at t = −4, 1, 4.
By the Existence and Uniqueness Theorem , the largest interval in which a unique solution is
guaranteed to exist is (1, 4).
Before we move on, let’s remind ourselves of the useful trig facts here:
sin(t) = 0 ⇒ t = nπ for all integer n

(2n − 1)π
cos(t) = 0 ⇒ t= for all integer n
2
Example 18 (Fall 2000 Midterm Exam I). Consider the initial value problem
π
sin(t)y 0 + cos(t)y = 3t sin(t), y( ) = 0.
2
On what interval is the solution guaranteed to exist? Justify your answers!
Solution First, rewrite the equation in the standard form,
cos(t)
y0 + y = 3t,
sin(t)
cos(t)
with p(t) = and g(t) = 3t.
sin(t)
So p has discontinuities at t = nπ, where n is an integer and g is continuous everywhere.
By the Existence and Uniqueness Theorem , the interval of validity is (0, π).
It’s worth noting that many functions we deal with so far have a domain of all real numbers. But
the natural logarithm ln t has different domain. This function is defined only when t > 0, in other
words, its domain is the set of positive real numbers or (0, ∞). Roughly speaking, number that
you are allowed to put after ln is only the positive ones. Base on this fundamental fact, we also can
talk about the domain of other types of function involving the natural logarithm. For example,
ln(t − a) is defined only when t − a > 0 or t > a. We list others below:
ln t is defined when t > 0

ln(t − a) is defined when t > a
ln |t| is defined when t 6= 0
ln |t − a| is defined when t 6= a
Example 19 (Spring 2010 Midterm Exam I). Consider the initial value problem
(t + 1)(t − π)y 0 − ln(t)y = cos(3t), y(2) = 5.
anteed to exist?
Solution The standard form of the given equation is,
ln(t) cos(3t)
y0 − y= ,
(t + 1)(t − π) (t + 1)(t − π)
ln(t) cos(3t)
with p(t) = − and g(t) = .
(t + 1)(t − π) (t + 1)(t − π)
So p has discontinuities at t = −1,
π (since t = −4 is not in the domain of ln) and g has
discontinuities at t = −1, π. Since the natural logarithm appears in p, we draw the the line
starting from 0 to infinity (not the whole real numbers).
Therefore the largest such interval is (0, π).

Example 20 (Summer 2010 Midterm Exam I). Consider the initial value problem
(t2 − 16)y 0 − ln(t − 1)y = sin(3t), y(π) = 5.
Without solving the equation, what is the largest interval in which a unique solution is
guaranteed to exist?
Solution Again write the equation in standard form,
ln(t − 1) sin(3t)
y0 − y= ,
(t + 4)(t + 4) (t + 4)(t + 4)
ln(t − 1) sin(3t)
with p(t) = and g(t) = .
(t + 4)(t + 4) (t + 4)(t + 4)
Note that the domain of ln(t − 1) is t > 1. So p has discontinuities at t =
−4,
4 (since t = −4
is not in the domain of ln(t − 1)) and g has discontinuities at t = −4, 4.
By the Existence and Uniqueness Theorem , the largest interval in which a unique solution is
guaranteed to exist is (1, 4).
The Existence and Uniqueness Theorem requires a linear differential equation. There is a also
similar theorem for non-linear first order differential equations.
Theorem 2. Let the functions f and ∂f ∂y be continuous in some rectangle α < t < β, γ <
y < δ containing the point (t0 , y0 ). Then, in some interval t0 − h < t < t0 + h contained in
α < t < β, there is a unique solution y = φ(t) of the initial value problem
y 0 = f (t, y) y(t0 ) = y0 .
Example 21. Determine all possible solutions to the following IVP,

1
y0 = y 3 , y(0) = 0.
Solution First, notice that the given nonlinear differential equation does not meet the re-
quirement of the theorem above. We have
1 ∂f 1
f (t, y) = y 3 ⇒ = 2 .
∂y 3y 3
Even though f is continuous everywhere but its derivative with respect to y is not continuous
at y = 0 and so will not be continuous at any interval containing y = 0.
Now let’s solve this problem by using the method of separable equation,
Z Z
dy 1 1 3 2
= y3 ⇒ y − 3 dy = dt ⇒ y 3 = t + c.
dt 2
Apply initial condition, y(0) = 0 gives c = 0 and so the particular solution is

3 2
y 3 = t.
2
Write it in explicit form,
r
3 2 2 2 2 2
y3 = t ⇒ y = t
3 ⇒ y = ( t)3
2
⇒ y=± ( t)3 .
2 3 3 3
Both of them also satisfy the initial condition. Also there is a third solution to this IVP,
namely y = 0, which satisfy both the given equation and the initial condition.
The previous example shows that the IVP does not always give the unique solution.
Example 22. Which of the following initial value problems has more than one solution?
(a) y 0 = 2y, y(0) = 0
1
(b) y 0 = y 3 , y(0) = 0
(c) y 00 + ty 0 + 2y = 0, y(0) = 0, y 0 (0) = 0
(d) (1 + t2 )y 0 + ty = 0, y(0) = 0
Solution Every equation has a unique solution by the Existence and Uniqueness Theorem
except the one in (b). It is nonlinear.
Exercises 2.2
1. Consider the initial value problem
1
(t2 + t)y 0 + y = e2t , y(2) = 3π.
(t − 4)
According to the Existence and Uniqueness Theorem, what is the largest interval in which
a unique solution is guaranteed to exist?
(a) (4, ∞)
(b) (−∞, 4)
(c) (0, 4)
(d) (1, 4)
(t2 − 4)y 0 + 3y = ln |5 − t|, y(3) = 0.
(a) (2, ∞)
(b) (−2, 2)
(c) (2, 5)
(d) (−∞, 5)
3. The initial value problem
(4 − t2 )y 0 + ln(t)y = sin(t), y(1.33) = 3.14159
is certain to have a unique continuous solution on the interval.

(a) (−4, 4)
(b) (−2, 2)
(c) (0, 2)
(d) (0, 3.14159)
(t2 − 9)y 0 + (2t sin t)y = ln(1 − t), y(0) = 4.
(a) (−3, 3)
(b) (3, ∞)
(c) (1, 3)
(d) (−3, 1)
Answers
1. c)
2. c)
3. c)
4. d)
2.3 Stability of Equilibrium Solutions

Recall that
• The first order differential equation in the form
dy
= f (y) (f is free from independent variable t),
dt
is called autonomous.
• Equilibrium solutions or critical points are constant functions y satisfying y 0 = f (y) = 0.
What we are about to study will tell us about the long term behavior of y without solving the
equation like direction filed. But it’s quicker and more effective.
Given an autonomous equation,
dy
= f (y).
dt
• First, let’s draw the graph of f on yf (y)−plane.
• The y-axis along with the arrows (which we’ll talk about it shortly) is called phase line.
• It is more customary to draw it in vertical orientation.
• Here’s how you draw the arrow. On any interval where f (y) > 0, y is increasing and we
denote this fact by drawing the right arrow (upward arrow for vertical version). Similarly,
on interval where f (y) < 0, y is decreasing we draw left arrow (downward arrow for vertical
version).
Take a look at this example below.
Figure 2.1: Left: Graph of f (y) and phase line, Right: Phase line in vertical orientation.
As you notice from the right figure, the number −1 and 3 appear on the axis reflecting the fact
that y = −1 and y = 3 are equilibrium solutions of the given autonomous equation. In general,
there are three types of stability of each equilibrium solutions based on the picture of phase line.
Let y = a be an equilibrium solution.
In our example, y = 3 is (asymptotically) stable while y = −1 is unstable.

For the autonomous equation y 0 = y 3 + 2y 2 − 3y.
a) Find all equilibrium solutions.
Solution To find equilibrium solutions, set y 0 = 0. Hence
0 = y 3 + 2y 2 − 3y = y(y 2 + 2y − 3) = y(y + 3)(y − 1).
So y = −3, 0, 1 are equilibrium solutions.

b) Determine the stability of each equilibrium solution you find in part a).
Solution
• Method 1: Phase line

Now f (y) = y(y + 3)(y − 1). To consider the sign for each interval, we simply pick
the sample point and evaluate f at that point.
So y = 1 is unstable, y = 0 is (asymptotically) stable and y = −3 is unstable.

• Method 2: Checking f 0 . Since f (y) = y 3 + 2y 2 − 3y, f 0 (y) = 3y 2 + 4y − 3.
f 0 (1) = 3 + 4 − 3 > 0 ⇒ y = 1 is unstable

f 0 (0) = 0 + 0 − 3 < 0 ⇒ y = 0 is stable
0
f (−3) = 27 − 12 − 3 > 0 ⇒ y = −3 is unstable
c) Let y(t) be the solution whose initial condition is y(0) = 4. What is the behavior of y(t)
as t → ∞.
Solution Since 4 lies between the interval (1, ∞),
lim y(t) = ∞,
t→∞
from phase line above.

d) Let y(t) be the solution whose initial condition is y(0) = −1. What is the behavior of
y(t) as t → ∞.
Solution Since −1 lies between the interval (−3, 0),
lim y(t) = 0,
t→∞
from phase line above.

Consider the first order autonomous equation
dy
= (y − 6)2 (y 2 − 25).
dt
a) Find all equilibrium solutions.
Solution We can find equilibrium solutions by letting y 0 = 0 which is
0 = (y − 6)2 (y 2 − 25) = (y − 6)2 (y − 5)(y + 5).
So y = −5, 5, 6 are equilibrium solutions.

b) For each equilibrium solution, classify its stability. Justify your answer.
Solution We gonna use the phase line to determine the stability of each equilibrium
solution. In this example, we have
f (y) = (y − 6)2 (y − 5)(y + 5).
Therefore y = 6 is semistable, y = 5 is unstable and y = −5 is (asymptotically) stable.

c) If y(1) = 0, what is lim y(t)?
t→∞
Solution Since 0 lies between the interval (−5, 5),
lim y(t) = −5.

t→∞
d) Suppose y(0) = y0 and lim y(t) = 5. Find the value(s) of y0 .

t→∞
Solution If y(0) = 5, since 5 is an equilibrium solution, lim y(t) = 5. If the value of
t→∞
y(0) is not 5, then from the phase line we can see that lim y(t) is never equal to 5. So
t→∞
y0 = 5 only.
Exercises 2.3
1. For the following equation,
dy
= (y + 1)(y − 2)(1 − y).
dt
Determine and classify the critical points (equilibrium solutions).
2. Consider the autonomous differential equation
y 0 = (y − 1)(y − 2)(y + 3)2 .
(a) Find all of its equilibrium solutions.

(b) Classify the stability of each equilibrium solution. Clearly explain how you have ob-
tained your answer.
(c) If y(−3) = −1, then what is lim y(t)?
t→∞
(d) If y(500) = 2, then what is y(2500)? Without solving the equation, briefly explain your
conclusion.
y 0 = 3y(y + 2)2 (4 − y).

(b) Classify the stability of each equilibrium solution. Justify your answer.
(c) If y(−2) = −1, then what is lim y(t)?
t→∞
(d) If y(10) = 0, then what is y(1000)? Without solving the equation, briefly explain your
conclusion.
4. Consider the solution y(t) of the initial value problem
y 0 − 2y = 2, y(0) = 2.
What is lim y(t)?

t→∞
(a) −∞
(b) −1
(c) 2
(d) ∞
5. Consider the following autonomous differential equation
y 0 = cos(y).
5π
The function y(t) = is
2
(a) a semi-stable equilibrium solution.
(b) an unstable equilibrium solution.
(c) an asymptotically stable equilibrium solution.

(d) not an equilibrium solution.
y 0 = −y 4 + 16y 2 .
(c) If y(t) is a solution that satisfies y(−1) = −4, what is y(0)? Without solving the
equation, briefly explain your conclusion.
(d) If y(t) is a solution that satisfies y(3) = −3, then what is lim y(t)?
t→∞
7. Consider the autonomous equation

y 0 = 9y 2 − y 4 = y 2 (3 − y)(3 + y)
Answer the following questions without solving the equation.
(a) Find all equilibrium solutions.
22
(c) If y( ) = π, what is lim y(t)?
7 t→∞
(d) If y(2π) = −3, find y(t).
(e) If y(4) = λ. For what value (or range of values) of λ would lim y(t) = 0?
t→∞
8. Consider the autonomous equation

y 0 = y 3 − 12y 2 + 20y = y(y − 2)(y − 10).
π
Suppose y( ) = 10. What is lim y(t)?
2 t→∞
(a) 0
(b) 2
(c) 10
(d) ∞
y 0 = (y − 1)2 (y 2 − 4).
(c) If y(1325) = 1.5, then what is lim y(t)?
t→∞
(d) If y(−1325) = 2, then what is lim y(t)?
t→∞
Answers
1. The equilibrium solutions are: y = −1(stable), y = 1(unstable), and y = 2(stable).
2. a) y = −3, 1, 2, b) semistable, stable, unstable, c) lim y(t) = 1, d) y(2500) = 2
t→∞
3. a) y = −2, 0, 4, b) semistable, unstable, stable, c) lim y(t) = −2, d) y(1000) = 0
t→∞
4. d)
5. c)
6. a) y = −4, 0, 4, b) unstable, semistable, stable, c) y(0) = −4, d) lim y(t) = 0
t→∞
7. a) y = −3, 0, 3, b) unstable, semistable, stable, c) 3, d) −3, e) −3 < λ ≤ 0
8. c)
9. a) y = −2, 1, 2, b) stable, semistable, unstable, c) lim y(t) = 1, d) lim y(t) = 2
t→∞ t→∞
2.4 Modeling with First Order Differential Equation: Mix-

ing Problems
Modeling is the process of writing a differential equation to describe a physical situation. One
of the modeling with first order differential equation is mixing problems. Here is the set up:
Figure 2.2: Mixing problems
• A holding tank initially contains amount of substance dissolved in certain amount of liquid.
• The same substance is also dissolved in a liquid which will be entering the tank.
• Assume that the content of the mixing tank is stirred very quickly such that the solution
within is always of uniform concentration.
• The mixed content is then leaving the tank.
• We want to develop a differential equation that can predict the amount of the substance
dissolved in the tank at any time.
Let Q(t) be the amount of the substance dissolved in the liquid in the tank at any time t. Then
we have,
Rate of change of Q(t) = Rate at which Q(t) enters the tank − Rate at which Q(t) exits the tank
Note here that

dQ
• Rate of change of Q(t) is or Q0 (t).
dt
• Rate at which Q(t) enters the tank
=(Flow rate of liquid entering)×(concentration of substance in liquid entering)
= ri ci
• Rate at which Q(t) exits the tank
=(Flow rate of liquid exiting)×(concentration of substance in liquid exiting)
Amount of substance in the tank at time t
=(Flow rate of liquid exiting)×
Volume of liquid in the tank at time t
Q(t)
= ro ,
V0 + (ri − ro )t
where V0 is the initial volume of liquid in the tank.
Combine everything altogether, we have
Modeling for Mixing Problems
Q(t)
Q0 (t) = ri ci − ro , Q(0) = Q0
V0 + (ri − ro )t
as the initial value problem for mixing problems where

• Q(t) = the amount of the substance dissolved in the liquid in the tank at time t.
• ri = flow rate of liquid entering (flow rate in).

• ci = concentration of substance in liquid entering (conc. in).
• ro = flow rate of liquid exiting (flow rate out).
• V0 = the initial volume of liquid in the tank.

• Q(0) = the initial amount of the substance dissolved in the liquid in the tank.

A tank initially contains 200 gallons of water with a salt concentration of 0.5 oz/gal. Water
containing a salt concentration of 2 + sin t oz/gal flows into the tank at a rate of 5 gal/min
and the mixture in the tank flows out at the same rate. Let Q(t) be the amount of salt in the
tank at time t. Find the initial value problem which accurately describes the situation.
Solution First, let Q(t) be the amount of salt dissolved in the water in the tank at time t
(oz). From the information above, we have
• V0 = 200 gal.
• Q(0) = 0.5 oz/gal ×200 gal = 100 oz. Here we use the fact that
amount of substance
concentration of substance =
volume of the liquid
∴ amount of substance = concentration of substance × volume of the liquid
• ci = 2 + sin t oz/gal.
• ri = 5 gal/min.
• ro = 5 gal/min.
Substitute everything back into modeling for mixing problem. Then we have the initial value
problem describing the situation above as
Q(t)
Q0 (t) = 5(2 + sin t) − 5 , Q(0) = 100
200 + (5 − 5)t
which is
Q(t)
Q0 (t) = 5(2 + sin t) − , Q(0) = 100.
40

A 1000-gallon above-ground swimming pool is initially filled with 800 gallons of rain water.
Water containing 5 g/gal of Chlorine flows into the pool at a rate of 6 gallons per minute.
The well-mixed Chlorine solution is pumped out at a rate of 3 gallons per minute. Let Q(t)
denote the amount of Chlorine in the pool at any time t > 0 and before the pool eventually
overflows. Write down an initial value problem that Q(t) must satisfy.
Solution From the problem, we get
• V0 = 800 gal (not 1000 which is the maximum capacity).

• Q(0) = 0 g (rain water).
• ci = 5 g/gal.
• ri = 6 gal/min.
• ro = 3 gal/min.
200
• t is at most min. (After that, the pool will overflow the swimming pool).
3
Therefore we have the initial value problem describing the situation above as follow
Q(t) 200
Q0 (t) = 6(5) − 3 , Q(0) = 0 where 0 < t ≤
200 + (6 − 3)t 3
which is
Q(t) 200
Q0 (t) = 30 − 3 , Q(0) = 0, where 0 < t ≤ .
200 + 3t 3

A 100-liter vat initially contains 80 liters of 2 grams/liter sodium hydroxide solution. At
t = 0, sodium hydroxide solution with a concentration of 5 grams/liter begins to flow into the
vat at the rate of 2 liters/min. The thoroughly mixed content of the vat is drawn off at the
rate of 3 liters/min. Find the initial value problem which best describe the quantity of sodium
hydroxide, Q(t), in the vat at time t, 0 < t < 80.
Solution We have
• V0 = 80 l (not 100 which is the maximum capacity).
• Q(0) = 2 g/l ×80 l = 160.
• ci = 5 g/gal.
• ri = 2 gal/min.
• ro = 3 gal/min.
• t < 80 min (After that, the vat will be drained).

Then we get the IVP describe the situation as follows,
Q(t)
Q0 (t) = 2(5) − 3 , Q(0) = 160 where 0 < t < 80
80 + (2 − 3)t
which is
Q(t)
Q0 (t) = 10 − 3 , Q(0) = 160, where 0 < t < 80.
80 − t

A tank has 100 gal of water and 100 lb of salt mixed in it. Water enters the tank at the rate of
3 gal/min with concentration of salt in it, at time t given by e−t lb/gal. A well mixed solution
leaves the tank at the same rate of 3 gal/min.
i) Find a formula for the amount of salt in the tank at any time t.
Solution Let Q(t) be the amount of salt dissolved in the water in the tank at time t
(lb). From the information, we have
• V0 = 100 gal.
• Q(0) = 100 lb.
• ci = e−t lb/gal.
• ri = 3 gal/min.
• ro = 3 gal/min.
Thus, we have the IVP for Q,
Q(t)
Q0 (t) = 3(e−t ) − 3 , Q(0) = 100
100 + (3 − 3)t
which is
Q(t)
Q0 (t) = 3e−t − 3 , Q(0) = 100.
100
Now we’re about to solve this IVP by using the method of integrating factor. So let’s
write the equation in the standard form,
Q(t)
Q0 (t) + 3 = 3e−t . (2.10)
100
And the integrating factor is
3 3t
R R
p(t) dt dt
µ(t) = e =e 100 = e 100 .
Multiply both sides of Equation 2.10 by µ and change the left hand side of the equation
to one term,
3t 3t Q(t) 3t
e 100 Q0 (t) + 3e 100 = 3e−t e 100
100
3t 97t
(e 100 Q(t))0 = 3e− 100
d 3t 97t
(e 100 Q(t)) = 3e− 100
dt
3t 97t
d(e 100 Q(t)) = 3e− 100 dt
Now integrate both sides of the equation and simplify it to get Q,

Z Z
3t 97t
d(e Q(t)) = 3 e− 100 dt
100
3t 300 − 97t
e 100 Q(t) = − e 100 + c
97
3t 300 − 97t
Q(t) = (e− 100 )(− e 100 + c)
97
300 −t 3t
∴ Q(t) = − e + ce− 100 .
97
Imposing the initial condition Q(0) = 100,

300 10, 000
100 = − +c ⇒ c= .
97 97
Hence the formula for Q(t) is
300 −t 10, 000 − 3t
Q(t) = − e + e 100 .
97 97
ii) Find also the eventual concentration of salt in the tank.

Solution Again we use the following relation,
amount of salt in the tank at time t
concentration of salt in the tank at timet =
volume of the water in the tank at time t
Q(t)
=
V0 + (ri − ro )t
Q(t)
=
100 + (3 − 3)(t)
Q(t)
=
100
Then the eventual concentration of salt in the tank is
Q(t) 1
lim = lim Q(t)
t→∞ 100 100 t→∞
1 300 −t 10, 000 − 3t
= lim − e + e 100
100 t→∞ 97 97
= 0.
Exercises 2.4
1. A tank is filled with 200 liters of a solution containing 100 grams of salt. A solution containing
a concentration of 2 g/liter salt enters the tank at the rate 4 liters/minute and the well-stirred
mixture leaves the tank at the same rate. Set up the initial value problem for the amount of
salt in the tank at time t, find the particular solution and find the limiting amount of salt
in the tank as t → ∞.
2. A tank originally contains 100 liters of pure water. Salt water with a concentration of 2
kg/L is pumped into the tank at 3 L/min, and the well-mixed solution is drained at the
same rate.
(a) Set up an initial value problem describing the situation. Be sure to explain all of your
variables.
(b) Solve the initial value problem to find the amount of salt in the tank at any time t.
(c) What is the limiting concentration of salt in the tank?
3. A 800-gallon tank initial contains 500 gallons of water and 30 pounds of salt dissolved in
it. Water enters the tank at the rate of 3 gal/min with concentration 4 lb/gal of salt in it.
The well mixed solution leaves the tank at the rate of 1 gal/min. Which of the initial value
problems below models the change of the amount of salt Q(t) inside the tank during the
time interval 0 ≤ t ≤ 150?
Q(t)
(a) Q0 (t) = 3 − , Q(0) = 30
800
Q(t)
(b) Q0 (t) = 12 − , Q(0) = 800
500 − 2t
Q(t)
(c) Q0 (t) = 12 − , Q(0) = 30
500 + 2t
Q(t)
(d) Q0 (t) = 12 − , Q(0) = 500
500 + t
4. A jar contains a sugar solution. Initially it had 2 liters of water and 15 grams of dissolved
sugar. Water containing 5 grams of sugar per liter enters the jar at a rate of 2 liters/min.
The well stirred mixture flows out at the same rate. How many grams of dissolved sugar is
present in the jar after ln 5 minutes?
5. A 400-liter tank is initially filled with 100 liters of dye solution with a dye concentration 5
g/l. Pure water flows into the tank at a rate 3 liters per minute. The well-stirred solution
is drained at a rate of 2 liters per minute. Find the concentration of dye in the tank at the
time that the tank is completely filled.
6. A 200 m3 room initially contains fresh air. At t = 0, a faulty heating system causes gas
containing 20% carbon monoxide to be pumped into the room at a rate of 3 m3 per minute.
The well-mixed air is vented out at the same rate.
(a) Write a differential equation, and give the initial condition, that describe this event.
(b) Solve the initial value problem.
(c) A carbon monoxide detector in the room is triggered when the carbon monoxide reaches
1%. Find the time when the detector will sound the alarm.
7. A swimming pool is initially filled with 500 m3 of water with a chlorine concentration of 10
g/m3 . Water containing 50 g/m3 of chlorine flows into the pool at a rate of 5 m3 per minute.
The well-mixed water in the pool is drained away at the same rate.
(a) Find the amount of chlorine in the pool at anytime t > 0.
(b) What is the concentration of chlorine in the pool as t → ∞
8. A tank initially contains 120 liters of pure water. A salt solution with a concentration of
γ grams/liter of salt enters the tank at a rate of 2 liters/min and the well-stirred mixture
leaves the tank at the same rate. Find (in terms of γ) an expression for the amount of salt
in the tank at any time t and the limiting amount of salt in the tank as t approaches infinity.
9. A swimming pool is initially filled with 400 m3 of fresh water. At t = 0 water containing
50 g/m3 of chlorine starts to flow into the swimming pool at a rate of 2 m3 per minute.
Well-mixed water is drained from the pool at the same rate.
(a) Set up an initial value problem modeling this process.
(b) Solve the initial value problem.
(c) Find the time when the chlorine concentration within the pool reaches 25 g/m3 .
(d) (As t → ∞) what is the limiting amount of chlorine that will be in the swimming pool?
10. A culinary experiment that went horribly awry has filled a 60 m3 kitchen with air that
contains 2 g/m3 of smoke and soot. At t = 0, the ventilation system is switched on so that 3
m3 /min of fresh air is pumped in. The well-mixed smokey air is drawn off at the same rate.
(a) Let Q(t) denote the amount of smoke and soot in the air at any time t > 0. Write down
an initial value problem that Q(t) must satisfy.
(b) Solve the initial value problem to find Q(t).
(c) How much time would it take for the concentration of smoke and soot in the air to go
1
down to 10 of its original level?
Answers
−t
0 Q
1. The initial value problem is Q = 8 − , Q(0) = 100. The solution is Q(t) = 400 − 300e 50 .
50
And lim Q(t) = 400.
t→∞
−3t
0 3Q
2. i) Q = 6 − , Q(0) = 0, ii) Q(t) = 200 − 200e 100 , iii) The limiting concentration is
100
Q(t) 200
lim = = 2.
t→∞ 100 100
3. c)
4. 11
5
5. The tank is filled at t = 300 then the concentration at that moment is
64
3 3t 200 20
6. a) Q0 = 0.6 − Q, Q(0) = 0, b) Q(t) = 40 − 40e− 200 , c) t = (ln )
200 1
3 19
7. a) Q = 25, 000 − 20, 000e− 100 t , b) 50
1
8. a) Q = 120γ − 120γe− 60 t , b) 120γ
1 t
9. a) Q0 = 100 − Q, Q(0) = 0, b) Q(t) = 20, 000 − 20, 000e− 200 , c) t = 200 ln 2, d) 20, 000
200
Q t
10. a) Q0 = − , Q(0) = 120, b) Q(t) = 120e− 20 , c) 20 ln 10
20
Chapter 3
Second Order Differential

Equations
In this chapter we will study about linear second order differential equations. The most general
form is
p(t)y 00 + q(t)y 0 + r(t)y = g(t),
where the coefficients p, q, r and g are functions of t.
Homogeneous vs Nonhomogeneous:
• A second order linear equation is said to be homogeneous if g(t) = 0 for all t.
• Otherwise, the equation is called nonhomogeneous.

For the nonhomogeneous equation, we will consider only when all of the coefficients p, q and r are
real constant. So we will study the following differential equation,
ay 00 + by 0 + cy = g(t),
where a 6= 0, b, c are arbitrary real constants.
3.1 Fundamental Set of Solutions & Wronskian

Let’s consider the following example.
Example 29. Determine some solutions to
y 00 − y = 0.
Solution It is easy to see that et is a solution and so is c1 et . Another solution is e−t and so
is c2 e−t . We can check further that any function in the form
y = c1 et + c2 e−t
is a solution of the above differential equation. In fact, it is a general solution which we will
show later.
43
CHAPTER 3. SECOND ORDER DIFFERENTIAL EQUATIONS 44
This example guide us to this fact:
Theorem 3 (Principle of superposition). If y1 and y2 are solutions of the homogeneous

equation
y 00 + p(t)y 0 + q(t)y = 0,
then so is c1 y1 + c2 y2 for arbitrary constant c1 and c2 .
Also it is worth noting that

• Unlike 1st order differential equation, the general solution of 2nd order equation has two
arbitrary coefficients.
• To find a particular solution, we require two initial conditions y(t0 ) = y0 and y 0 (t0 ) = y00 .
As in the example above, it suggests us that the solution of ay 00 + by 0 + cy = 0 should be in the
form y = ert . Plug this y into the the equation,
a(ert )00 + b(ert )0 + c(ert ) = 0 ⇒ a(r2 ert ) + b(rer t) + cert = 0 ⇒ ert (ar2 + br + c) = 0.
Since ert 6= 0,
ar2 + br + c = 0.
• This equation is typically called the characteristic equation for ay 00 + by 0 + cy = 0.
• This is a quadratic equation and so we should expect two roots r1 and r2 . Once we have
these two roots, we have two solutions to ay 00 + by 0 + cy = 0
y1 (t) = er1 t and y2 (t) = er2 t .
• Apply the principle of superposition, we have

y = c 1 e r1 t + c 2 e r2 t
as a solution to ay 00 + by 0 + cy = 0.
So the only remaining question is that
Does it contain every possible solutions to ay 00 + by 0 + cy = 0?
It can be answered by the theorem below
Theorem 4 (Fundamental Set of Solutions, Wronskian).

Suppose that y1 and y2 are two solutions of
y 00 + p(t)y 0 + q(t)y = 0.
Then
y = c1 y1 + c2 y2 is the general solution of
(or the solutions y1 and y2 are said to form a fundamental set of solution of )
y 00 + p(t)y 0 + q(t)y = 0
if and only if the Wronskian of y1 and y2 , written as W (y1 , y2 ), is not a zero function where

y y
1 2
W (y1 , y2 ) := = y1 y20 − y10 y2 .

y 0 y 0
1 2
3.2 Homogeneous Equation

Recall that the characteristic equation of a differential equation ay 00 + by 0 + cy = 0 is
ar2 + br + c = 0,
which will gives us two roots. We’ll take a closer look on all 3 possible cases depend on the type
of roots as follow:
1. Two distinct real roots (it happens when b2 − 4ac > 0).
2. Two complex conjugate roots (it happens when b2 − 4ac < 0).
3. Repeated real roots (it happens when b2 − 4ac = 0).
3.2.1 Two Distinct Real roots

Assume that r1 and r2 are two distinct real roots of the characteristic equation ar2 + br + c = 0.
This gives two solutions of ay 00 + by 0 + cy = 0,
y1 = er1 t and y2 (t) = er2 t .
Compute their Wronskian,

e r1 t er2 t
= r2 e(r1 +r2 )t − r1 e(r1 +r2 )t = (r2 − r1 )e(r1 +r2 )t .

W (y1 , y2 ) =
r e r1 t r2 er2 t
1
It is nonzero since r1 6= r2 and so is exponential function. By the theorem above, y1 and y2 form
the fundamental set of solutions of ay 00 + by 0 + cy = 0. It means that the general solution is
y = c1 y1 + c2 y2 ⇒ y = c1 er1 t + c2 er2 t .
If r1 , r2 are distinct real roots of characteristic equation, then the general solution is
y = c1 er1 t + c2 er2 t .
Example 30 (Spring 2002 Midterm Exam 1). Solve the IVP
y 00 − 4y = 0, y(0) = 4, y 0 (0) = 4.
Solution The characteristic equation of y 00 − 4y = 0 is
r2 − 4 = 0 ⇒ (r − 2)(r + 2) = 0 ⇒ r = 2, −2.
So the general solution is y = c1 e2t + c2 e−2t . Then y 0 = 2c1 e2t − 2c2 e−2t .
Impose the initial conditions,
y(0) = 4 ⇒ 4 = c1 + c2
0
y (0) = 4 ⇒ 4 = 2c1 − 2c2 .
Then we have c1 = 3 and c2 = 1. So the particular solution is y = 3e2t + e−2t .

Example 31 (Summer 2009 Midterm Exam 1).

Find the particular solution to
y 00 − 5y 0 + 4y = 0, y(0) = 2, y 0 (0) = −1.
What is the behavior of the solutions when t → ∞.

Solution The characteristic equation of y 00 − 5y 0 + 4y = 0 is
r2 − 5r + 4 = 0 ⇒ (r − 4)(r − 1) = 0 ⇒ r = 1, 4.
Hence the general solution is y = c1 et + c2 e4t . Note that y 0 = c1 et + 4c2 e4t .

Plug in the initial conditions,
y(0) = 2 ⇒ 2 = c1 + c2
0
y (0) = −1 ⇒ −1 = c1 + 4c2 .
Then we have c1 = 3 and c2 = −1. So the particular solution is
y = 3et − e4t .
Moreover,
lim y(t) = lim (3et − e4t ) = lim et (3 − e3t ) = −∞.
t→∞ t→∞ t→∞
Example 32 (Fall 2004 Midterm Exam 1).

Find α so that the solution to the IVP
y 00 + 3y 0 − 4y = 0, y(0) = α, y 0 (0) = 1,
converge to 0 as t → ∞.
Solution The characteristic equation of y 00 + 3y 0 − 4y = 0 is
r2 + 3r − 4 = 0 ⇒ (r + 4)(r − 1) = 0 ⇒ r = −4, 1.
Therefore the general solution is y = c1 e−4t + c2 et . Then y 0 = −4c1 e−4t + c2 et .

y(0) = α ⇒ α = c1 + c2
y 0 (0) = 1 ⇒ 1 = −4c1 + c2 .
α−1 4α + 1
Then we get c1 = and c2 = . So the particular solution is
5 5
α − 1 −4t 4α + 1 t
y= e + e.
5 5
4α + 1 t
Note that since lim e−4t = 0, lim y(t) = e.
t→∞ t→∞ 5
t
Since lim e = ∞, the only way to get lim y(t) = 0 is to have the zero coefficient. Hence
t→∞ t→∞
4α + 1 1
=0 ⇒ α=− .
5 4
3.2.2 Complex Conjugate Roots

Assume that r1,2 = λ ± µi are the roots of the characteristic equation ar2 + br + c = 0. So we get
two solutions to the differential equation ay 00 + by 0 + cy = 0 which are
y1 = e(λ+µi)t and y2 = e(λ−µi)t .
Let’s consider their Wronskian,

e(λ+µi)t e(λ−µi)t
= [(λ − µi) − (λ + µi)]e2λt = −2iµe2λt 6= 0,

W (y1 , y2 ) =
(λ + µi)e(λ+µi)t (λ − µi)e(λ−µi)t

since exponential function is nonzero and µ 6= 0.

So y1 and y2 form a fundamental set of solutions. Hence the general solution is
y(t) = c1 e(λ+µi)t + c2 e(λ−µi)t .
But we would like to express the solution as the real valued functions (without i). To do this,
we’ll need Euler’s formula,
eiθ = cos θ + i sin θ.

From this we have
e−iθ = cos(−θ) + i sin(−θ) = cos(θ) − i sin θ.
And now we are ready to rewrite the general solution,
y(t) = c1 e(λ+µi)t + c2 e(λ−µi)t

= c1 eλt · eiµt + c2 eλt · e−iµt
= c1 eλt (cos µt + i sin µt) + c2 eλ t(cos µt − i sin µt)
= (c1 + c2 )eλt cos µt + (c1 − c2 )ieλt sin µt.
After we rename the constant, we get the new form of the general solution,
y(t) = c1 eλt cos µt + c2 eλt sin µt or y(t) = eλt (c1 cos µt + c2 sin µt)
In conclusion,
If r1,2 = λ ± µi are complex roots of the characteristic equation, then the general solution is
y(t) = eλt (c1 cos µt + c2 sin µt).
Example 33 (Spring 2004 Midterm Exam 1).

Find the general solutions to the y 00 + 2y 0 + 3y = 0.
Solution The characteristic equation of y 00 + 2y 0 + 3y = 0 is
√
√
p
−2 ± 4 − 4(3) −2 ± −8
r2 + 2r + 3 = 0 ⇒ r = = = −1 ± 2i.
2 2
√ √
So the general solution is y(t) = e−t (c1 cos 2t + c2 sin 2t).

Solve the following IVP
y 00 + 4y 0 + 13y = 0, y(0) = 3, y 0 (0) = −9.
Be sure to express your solution in terms of real-valued functions only.

√
2 −4 ± 16 − 52 −4 ± 6i
r + 4r + 13 = 0 ⇒ r = = = −2 ± 3i.
2 2
y(t) = e−2t (c1 cos 3t + c2 sin 3t).
Note that
y 0 (t) = e−2t (−3c1 sin 3t + 3c2 cos 3t) − 2e−2t (c1 cos 3t + c2 sin 3t).
y(0) = 3 ⇒ 3 = c1
0
y (0) = −9 ⇒ −9 = 3c2 − 2c1 .
Thus c1 = 3 and c2 = −1. So the particular solution is
y(t) = e−2t (3 cos 3t − sin 3t).

Solve the IVP:
y 00 − 2y 0 + 5y = 0, y(0) = 2, y 0 (0) = −4.

√
2 ± 4 − 20
r2 − 2r + 5 = 0 ⇒ r = = 1 ± 2i.
2
Hence the general solution is
y(t) = et (c1 cos 2t + c2 sin 2t).
Then
y 0 (t) = et (−2c1 sin 2t + 2c2 cos 2t) + et (c1 cos 2t + c2 sin 2t).
y(0) = 2 ⇒ 2 = c1
y 0 (0) = −4 ⇒ −4 = 2c2 + c1 .
So c1 = 2 and c2 = −3. So the particular solution is
y(t) = et (2 cos 3t − 3 sin 3t).

3.2.3 Repeated Real Roots

Now assume that r1 = r2 = r are repeated real roots of characteristic equation ar2 + br + c = 0.
So two solutions of ay 00 + by 0 + cy = 0 are
y1 (t) = er1 t = ert and y2 (t) = er2 t = ert .
Compute their Wronskian,

ert rt
e
= re2rt − re2rt = 0.

W (y1 , y2 ) =
rert rert

So y1 and y2 don’t form the fundamental set of solutions. We will use y1 (t) = ert as the first
solution, but we’re going to need a new second solution. To get the second solution, we need a new
tool namely the method of reduction of order which we will study about it shortly. After using
this method, we will receive y2 (t) = tert as the second solution. Let’s check their Wronkskian,

ert te rt
= ert (rtert + ert ) − rte2rt = e2rt 6= 0.

W (y1 , y2 ) =
rert rtert + ert

Consequently, y1 and new y2 form the fundamental set of solutions which means the general
solution is
y = c1 ert + c2 tert .
In conclusion,
If r1 = r2 = r are repeated real roots of the characteristic equation, the general solution is
y(t) = ert + c2 tert .
Example 36 (Spring 2004 Midterm Exam 1).

Find the general solutions to
y 00 + 6y 0 + 9y = 0.
r2 + 6r + 9 = 0 ⇒ (r + 3)2 = 0 ⇒ r = −3, −3.
Thus y(t) = c1 e−3t + c2 te−3t is the general solution.

Consider the IVP
y 00 − 8y 0 + 16y = 0, y(0) = 3, y 0 (0) = 10.
1. Find the solution, y(t), of this IVP.

r2 − 8r + 16 = 0 ⇒ (r − 4)2 = 0 ⇒ r = 4, 4.
So the general solution is y(t) = c1 e4t + c2 te4t .

Note that
y 0 (t) = 4c1 e4t + c2 (4te4t + e4t ).
y(0) = 3 ⇒ 3 = c1
y 0 (0) = 10 ⇒ 10 = 4c1 + c2 .
Thus c1 = 3 and c2 = −2. Therefore the particular solution is
y(t) = 3e4t − 2te4t .
2. What is lim y(t)?

t→∞
Solution lim y(t) = lim (3e4t − 2te4t ) = lim e4t (3 − 2t) = −∞.
t→∞ t→∞ t→∞

What is the general solution of
9y 00 + 6y 0 + y = 0?
1 1
(a) c1 e 3 t + c2 e 3 t
(b) c1 e−3t + c2 e3t
t t
(c) c1 e− 3 + c2 te− 3
t
(d) c1 e 3 + c2 e3t
Solution The characteristic equation of 9y 00 + 6y 0 + y = 0 is

1 1
9r2 + 6r + 1 = 0 ⇒ (3r + 1)2 = 0 ⇒ r = − ,− .
3 3
1 1
Hence the general solution is y(t) = c1 e− 3 t + c2 te− 3 t . So (c) is the correct answer.
3.2.4 Summary & More Examples

After we considered all 3 cases, we have the summary
Let r1 and r2 be roots of characteristic equation ar2 + br + c = 0 of the differential equation
ay 00 + by 0 + cy = 0.
1. Two distinct real roots: r1 6= r2

The general solution is y = c1 er1 t + c2 er2 t .
2. Complex conjugate roots: r1,2 = λ ± µi

The general solution is y(t) = eλt (c1 cos µt + c2 sin µt).
3. Repeated real roots: r1 = r2 = r
The general solution is y(t) = ert + c2 tert .

What of the equations below has y(t) = c1 et + c2 e−2t as its general solution?
(a) y 0 − 2y = 0
(b) y 00 − y 0 − 2y = 0
(c) 2y 00 + 2y 0 − 4y = 0
(d) y 00 − 3y 0 − 2y = 0
Solution According to the general solution, we have r1 = 1 and r2 = −2. Both of them are
the roots of the characteristic equation. So
(r − 1)(r + 2) = 0 ⇒ r2 + r − 2 = 0 ⇒ y 00 + y 0 − 2y = 0.
Multiply both sides of the equation by 2, we have (c) as the correct answer.

Consider the equation
y 00 + 4y 0 + 4y = 0.
a) Find its general solution.
r2 + 4r + 4 = 0 ⇒ (r + 2)2 = 0 ⇒ r = −2, −2.
Thus y(t) = c1 e−2t + c2 te−2t is the general solution.
b) Find the particular solution satisfying
y(365) = 1, y 0 (365) = −6.
Solution
• Method 1: Usual way. By using the general solution from part a) ,we have
y 0 (t) = −2c1 e−2t + c2 (−2te−2t + e−2t ).
Then impose the initial conditions,
y(365) = 1 ⇒ 1 = c1 e−2(365) + c2 (365)e−2(365)

y 0 (365) = −6 ⇒ −6 = −2c1 e−2(365) + c2 (−2(365)e−2(365) + e−2(365) ).
After a long and messy calculation, we have c1 = 1461e730 and c2 = −4e730 . Hence
the particular solution for the given IVP is
y = 1461e730 e−2t − 4e730 te−2t = (1461 − 4t)e730−2t .

• Method 2: Translation. We introduce new variable T by letting T = t − 365. Then

the initial conditions change as follow:
y(t = 365) = 1 ⇒ y(T = 0) = 1

0
y (t = 365) = −6 ⇒ y 0 (T = 0) = −6.
Now apply new initial conditions to the general solution we have in part a),
y(0) = 1 ⇒ 1 = c1
0
y (0) = −6 ⇒ −6 = −2c1 + c2 .
Then we get c1 = 1 and c2 = −4. Hence the particular solution is
y(T ) = e−2T − 4T e−2T .
The final step is to change the variable from T back to t using T = t − 365,
y(t) = e−2(t−365) − 4(t − 365)e−2(t−365) .
• Comparison Method 1 to Method 2. After we simpify the solution from Method 2,

we can see that it’s actually the same as the solution we get from Method 1. Here
is the detail:
y(t) = e−2(t−365) − 4(t − 365)e−2(t−365)

= (1 − 4(t − 365))e−2(t−365)
= (1 − 4t + 4(365))e−2t+2(365)
= (1461 − 4t)e730−2t .
c) For the solution found in b), determine lim y(t).

t→∞
Solution Using L’Hospital’s rule, we have
(1461 − 4t)e730 −4e730

lim y(t) = lim (1461 − 4t)e730−2t = lim 2t
= lim = 0.
t→∞ t→∞ t→∞ e t→∞ 2e2t
a) Find the general solution of the following differential equation,
y 00 + 6y 0 + 13y = 0.

√
2 −6 ± 36 − 52 −6 ± 4i
r + 6r + 13 = 0 ⇒ r = = = −3 ± 2i.
2 2
So the general solution is y(t) = e−3t (c1 cos 2t + c2 sin 2t).
b) Solve the following IVP for the above ODE:
y(0) = 1, y 0 (0) = 3.
Solution From part a), we have
y 0 (t) = e−3t (−2c1 sin 2t + 2c2 cos 2t) − 3e−3t (c1 cos 2t + c2 sin 2t).
Then apply initial conditions,
y(0) = 1 ⇒ 1 = c1
0
y (0) = 3 ⇒ 3 = 2c2 − 3c1 .
So c1 = 1 and c2 = 3. Then the particular solution is
y(t) = e−3t (cos2t + 3 sin 2t).
c) Solve the following IVP for the above ODE:
y(999) = 1, y 0 (999) = 3.
Solution Use the idea of translation by introduce new variable T = t − 999. Then the
initial conditions change to
y(t = 999) = 1 ⇒ y(T = 0) = 1

0
y (t = 999) = 3 ⇒ y 0 (T = 0) = 3.
By part b), we then have the solution
y(T ) = e−3T (cos2T + 3 sin 2T ).
Now change it back to variable t by using T = t − 999. So we have
y(t) = e−3(t−999) (cos2(t − 999) + 3 sin 2(t − 999)),
as the particular solution of the given IVP.

Exercises 3.2
1. Consider the second order linear equation
y 00 − 8y 0 + 16y = 0.
(a) Find the general solution of the equation.

(b) Find the solution satisfying the initial conditions y(735) = 2, y 0 (735) = −1.
(c) Find lim y(t).
t→∞
2. Construct a second order homogeneous linear equation with constant coefficients, such that
1
y1 = e−t and y2 = −πe3t are two of its solutions.
π
y 00 + 6y 0 + 10y = 0.
(a) Find its general solution.

(b) Find the solution satisfying the initial conditions y(5031) = 6 and y 0 (5031) = −1.
(c) Let y(t) be the solution found in ii). Evaluate lim y(t).
t→∞
4. What is the solution of the initial value problem
y 00 − 4y 0 + 4y = 0, y(3) = −1, y 0 (3) = 2?
(a) y = −e2t−6 + 4te2t−6

(b) y = −13e2t−6 + 4te2t−6
(c) y = −e2t+6 − 4te2t+6
(d) y = 11e2t+6 + 4te2t+6
5. Let y(t) be the solution of the initial value problem
y 00 + 2y 0 − 8y = 0, y(0) = α, y 0 (0) = −2.
Suppose lim y(t) = 0, find the value of α.

t→∞
1
(a) α =
2
(b) α = −1
(c) α = −4
(d) α = 8
6. Consider all nonzero solutions of the linear equation
y 00 − 2y 0 + 5y = 0.
As t → ∞, they will
(a) oscillate between ∞ and −∞, and not approach any limit.
(b) all approach 0.
(c) some approach ∞, all the rest approach −∞.
(d) some approach 0, some approach ∞, some approach −∞.
7. Which equation below has the function y = 19e4t as a particular solution?

(a) y 00 + 16y = 0
(b) y 00 − 8y 0 + 16y = 0
(c) y 00 + 4y 0 = 0
(d) y 00 + 5y 0 + 4y = 0
y 00 + 3y 0 − 10y = 0, y(0) = α, y 0 (0) = −2.
Suppose lim y(t) = 0, find the value of α.

t→∞
(a) −1
(b) 10
2
(c)
5
4
(d) −
5
9. Which of the following second order linear equations below has y = 6e−3t −3et as a solution?
(a) y 00 + 3y 0 = 0
(b) y 00 + 2y 0 − 3y = e−3t sin t
(c) 2y 00 − 4y 0 − 6y = 0
(d) −y 00 − 2y 0 + 3y = 0
16y 00 + 8y 0 + y = 0.
t t
(a) y = c1 e 4 + c2 e 4
t t
(b) y = c1 e− 4 + c2 te− 4
(c) y = c1 e−4t + c2 e4t
t
(d) y = c1 e− 4 + c2 e−4t
y 00 − 7y 0 + 12y = 0, y(0) = 4, y 0 (0) = 11.
(a) Find the solution y(t) of this initial value problem.

(b) Determine lim y(t).
t→∞
12. (a) Find the solution to the initial value problem
y 00 + 3y 0 + 2y = 0, y(0) = 0, y 0 (0) = α
(b) For what value(s) of α is the lim y(t) = 0?

t→∞
(c) For what value(s) of α is the lim y(t) = 0?
t→−∞
13. Find the general solution for each of the following equations.
(a) y 00 − y 0 − 2y = 0
(b) y 00 + 4y 0 + 4y = 0
(c) y 00 + 4y 0 + 5y = 0
14. Find the particular solution to
y 00 − 5y 0 + 4y = 0, y(0) = 2, y 0 (0) = −1.
What is the behavior of the solution as t approaches positive infinity?

15. In each of the parts (a), (b), and (c), find the real-valued general solution of the second order
linear equation given.
(a) y 00 + 12y 0 + 36y = 0
(b) y 00 − 7y 0 − 8y = 0
(c) y 00 − 4y 0 + 13y = 0
For each of the next parts, consider the solutions of the three equations above and write
down the correct equation whose solutions behave as stated.
(a) Every solution of this equation approaches 0 as t → ∞

(b) Every nonzero solution of this equation does not approach a finite limit, nor does it
have a limit of ∞ or −∞, as t → ∞.
Answers
1. a) y = c1 e4t + c2 te4t , b) y=2e4(t−735) − 9(t − 735)e4(t−735) , c) −∞
2. y 00 − 2y 0 − 3y = 0
3. a) y = c1 e−3t cos t + c2 e−3t sin t, b) y = 6e−3(t−5031) cos(t − 5031) + 17e−3(t−5031) sin(t − 5031),
c) 0
4. b)
5. a)
6. a)
7. b)
8. c)
9. d)
10. b)
11. a) y = 5e3t − e4t , ii) −∞
12. a) y = αe−t − αe−2t , b) All real numbers, c) α = 0 only
13. a) y = c1 e2t + c2 e−t , b) y = c1 e−2t + c2 te−2t , c) y = c1 e−2t cos t + c2 e−2t sin t
14. y = 3et − e4t , −∞
15. a) y = c1 e−6t + c2 te−6t , b) y = c1 e8t + c2 e−t , c) y = c1 e2t cos 3t + c2 e2t sin 3t, d) i, e) iii
3.3 Reduction of Order

Now we look at solutions to non constant coefficient homogeneous second order differential equation
of the form
p(t)y 00 + q(t)y 0 + r(t)y = 0.
In general, finding solutions to this differential equation is difficult. However, if we already knew
one of its solution, we can use the method of reduction of order to fnd a second solution.
Together their linear combination forms a fundamental set of solutions of the above equation. The
basic idea is to assume that the second solution has similar form as the first solution by assume
that
y2 (t) = v(t)y1 (t),
where v is the function of t. Our job is to find this certain function. Along the way, we will be
forced to solve the second order differential equation which we can’t if we do it directly. Instead
we introduce new variable u by setting u = v 0 to reduce the order of the equation to first order in
which we have more tools to solve.
The process of reduction of order method

1. Assume that the second solution is
y2 (t) = v(t)y1 (t)
where y1 (t) is the given first solution.

2. Using y2 from first step to compute y20 and y200 . Then substitute them back in the
differential equation to find v(t)
3. Reduce the order of differential equation from 2nd order to 1st order by introducing
new variable u = v 0 .
4. Solve for u.
5. Solve for v by using u = v 0 .
6. Substitute v back into y2 = v(t)y1 (t) to get the second solution y2 .

7. Choose suitable constant to get y2 with no repeated term in y1 .
8. Then the general solution to p(t)y 00 + q(t)y 0 + r(t)y = 0 is
y(t) = c1 y1 + c2 y2 .
The last step is true (as long as v is not a constant function) since their Wronskian is nonzero,

y vy1
1
= y1 (vy10 + y1 v 0 ) − vy1 y10 = y12 v 0 6= 0.

W (y1 , y2 ) =

y 0 vy 0 + y v 0
1 1 1
(y1 is nonzero and so is v 0 if v is not a constant function.)

The linear homogeneous differential equation
t2 y 00 + ty 0 − 4y = 0, t>0
has a solution y1 (t) = t2 . Use the reduction of order to find the second solution y2 of this
differential equation.
Solution Suppose that y2 (t) = v(t)y1 (t) = vt2 is the second solution. Then the first and
second derivative of y2 are
y20 = (vt2 )0 = 2tv + t2 v 0

y200 = (y20 )0 = (2tv + t2 v 0 )0 = 2tv 0 + 2v + t2 v 00 + 2tv 0 = 4tv 0 + 2v + t2 v 00 .
Since y2 is a solution to t2 y 00 + ty 0 − 4y = 0,
0 = t2 y200 + ty20 − 4y2

= t2 (4tv 0 + 2v + t2 v 00 ) + t(2tv + t2 v 0 ) − 4(vt2 )
= 4t3 v 0 + v + t4 v 00 +
0
2t2 2t2
v + t3 v − 2
4vt
= t4 v 00 + 5t3 v 0 .
Reducing the order of differential equation by letting u = v 0 , the differential equation becomes
0 = t4 u0 + 5t3 u.
This is a separable first order equation.

Z Z
du du 5 du 5
t4 = −5t3 u ⇒ = − dt ⇒ = − dt ⇒ ln u = −5 ln t + c0 .
dt u t u t
(Here we use the fact that t > 0)

Hence
−5
+c0 −5 0 0
ln u = ln t−5 + c0 ⇒ u = eln t = eln t · ec = ct−5 where c = ec .
But u = v 0 , so we have
v 0 = ct−5 .
Again we use the method of separable equation to solve it,
Z Z
dv c
= ct−5 ⇒ dv = ct−5 dt ⇒ v = − t−4 + k.
dt 4
Consequently,
c c
y2 (t) = v(t)y1 (t) = (− t−4 + k)t2 = − t−2 + kt2 .
4 4
Choose c = −4 and k = 0, we have
y2 (t) = t−2 .
(The general solution is y(t) = c1 y1 + c2 y2 = c1 t2 + c2 t−2 )

Given that y1 (t) = t3 is a known solution of the second order linear differential equation
t2 y 00 − 5ty 0 + 9y = 0, t > 0.
Find the general solution of the equation.

Solution Let y2 (t) = v(t)y1 (t) = vt3 be the second solution. Then the first and second
derivative of y2 are
y20 = (vt3 )0 = 3t2 v + t3 v 0

y200 = (y20 )0 = (3t2 v + t3 v 0 )0 = 3t2 v 0 + 6tv + t3 v 00 + 3t2 v 0 = 6t2 v 0 + 6tv + t3 v 00 .
Since y2 is a solution to t2 y 00 − 5ty 0 + 9y = 0,
0 = t2 y200 − 5ty20 + 9y2

= t2 (6t2 v 0 + 6tv + t3 v 00 ) − 5t(3t2 v + t3 v 0 ) + 9(vt3 )
= 6t4 v 0 + v + t5 v 00 −
6t3 v − 5t4 v 0 +
3
15t 9vt 3
= t4 v 0 + t5 v 00 .
Divide out both sides of the equation by t5 (it is allowed since t > 0),
0 = t−1 v 0 + v 00 .
Reducing the order of differential equation by letting u = v 0 , the differential equation becomes
0 = t−1 u + u0 .

Z Z
−1 du du dt c
−t u = ⇒ = − ⇒ ln u = − ln t + c0 ⇒ u= .
dt u t t
But u = v 0 ,
c
v0 = .
t
Again we use the method of separable equation to solve it,
Z Z
dv c c
= ⇒ dv = dt ⇒ v = c ln t + k.
dt t t
Hence
y2 (t) = v(t)y1 (t) = (c ln t + k)t3 = ct3 ln t + kt3 .
Choose c = −1 and k = 0, we have y2 (t) = t3 ln t. Therefore the general solution is
y(t) = c1 y1 + c2 y2 = c1 t3 + c2 t3 ln t.

Given that y1 (t) = t4 is a known solution of the second order linear differential equation
t2 y 00 − 6ty 0 + 12y = 0, t > 0.

Solution Let y2 (t) = v(t)y1 (t) = vt4 be the second solution. Then the first and second
derivative of y2 are
y20 = (vt4 )0 = 4t3 v + t4 v 0

y200 = (y20 )0 = (4t3 v + t4 v 0 )0 = 4t3 v 0 + 12t2 v + t4 v 00 + 4t3 v 0 = 8t3 v 0 + 12t2 v + t4 v 00 .
Since y2 is a solution to t2 y 00 − 6ty 0 + 12y = 0,
0 = t2 y200 − 6ty20 + 12y2

= t2 (8t3 v 0 + 12t2 v + t4 v 00 ) − 6t(4t3 v + t4 v 0 ) + 12(vt4 )
= 8t5 v 0 + 12tv + t6 v 00 −
4
24tv − 6t5 v 0 +
4
12vt 4
= 2t5 v 0 + t6 v 00 .
Divide out both sides of the equation by t6 (it is allowed since t > 0),
0 = 2t−1 v 0 + v 00 .
Use u = v 0 to reduce the order of the differential equation. So
0 = 2t−1 u + u0 .

Z Z
−1 du du 2
−2t u = ⇒ = − dt ⇒ ln u = −2 ln t + c0 ⇒ u = ct−2 .
dt u t
But u = v 0 ,
v 0 = ct−2 .
Using the method of separable equation to solve it,
Z Z
dv −2
= ct ⇒ dv = ct−2 dt ⇒ v = −ct−1 + k.
dt
Therefore
y2 (t) = v(t)y1 (t) = (−ct−1 + k)t4 = −ct3 + kt4 .
Choose c = −1 and k = 0, we have y2 (t) = t3 . Therefore the general solution is
y(t) = c1 y1 + c2 y2 = c1 t4 + c2 t3 .
Exercises 3.3
1. Given that y1 (t) = t2 is a known solution of the second order linear equation
t2 y 00 − 4ty 0 + 6y = 0, t > 0.
Use reduction of order to find the general solution of the equation.

1
2. Given that y1 (t) = is a known solution of the second order linear equation
t2
t2 y 00 − ty 0 − 8y = 0, t > 0.
t2 y 00 − 3ty 0 + 4y = 0, t > 0.
(a) Verify that y1 (t) = t2 is a solution of this equation.

(b) Find the equation’s general solution using the method of reduction of order.
4. Given that y1 (t) = t is a known solution of the second order linear differential equation
t2 y 00 − 4ty 0 + 4y = 0, t > 0.
5. Given that y1 (t) = t3 is a solution, find the general solution of
t2 y 00 + 2ty 0 − 12y = 0, t > 0.
6. Given that y1 (t) = t−1 is a solution to the equation,
t2 y 00 + 3ty 0 + y = 0, t > 0.
(a) Use the method of reduction of order to find another solution y2 which is not a scalar
multiple of y1 .
(b) Find the general solution of the equation.
(c) Find a solution satisfying the following initial conditions: y(e) = 1e , y 0 (e) = 1
e2 .
7. Given that y1 (t) = t4 is a known solution of the linear differential equation
t2 y 00 − 7ty 0 + 16y = 0, t > 0.
Use reduction of order to find the general solution of the equation.
Answers
1. y = c1 t2 + c2 t3
2. y = c1 t−2 + c2 t4
3. a)-, b) y = c1 t2 + c2 t2 ln t
4. y = c1 t + c2 t4
5. y = c1 t3 + c2 t−4
6. a) t−1 ln t, b) y(t) = c1 t−1 + c2 t−1 ln t, c) y(t) = −t−1 + 2t−1 ln t
7. y = c1 t4 ln t + c2 t4
3.4 The Existence and Uniqueness Theorem

This theorem is analogous to the one for first order linear equation.
Theorem 5 (The Existence and Uniqueness Theorem (for second order linear equation)).
y 00 + p(t)y 0 + q(t)y = g(t), y(t0 ) = y0 , y 0 (t0 ) = y00 ,
where p, q and g are continuous on an open interval I that contains the point t0 . Then there
is exactly one solution y = φ(t) of this problem, and the solution exists throughout the interval
I.

The existence and uniqueness theorem guarantees that the solution to
1 0
sin(t)y 00 + y + et y = t3 , y(1) = 0, y 0 (1) = 1
t−3
is valid on
(a) (0, 3) (b) (0, π) (c) (−∞, 3) (d) (−∞, ∞)
Solution Rewrite the given equation in suitable form,
1 et t3
y 00 + y0 + y= .
(t − 3) sin(t) sin(t) sin(t)
Then the discontinuities of p are t = nπ for all integer n and t = 3 while the discontinuities
of q and r are t = nπ for all integer n
Since t0 = 1 lies in (0, 3), the interval of validity is (0, 3). So (a) is the correct answer.

Consider the IVP
π 4π π π
sin(t)y 00 + tan(t)y 0 + ty = et , y( ) = , y0 ( ) = − .
4 3 4 4
anteed to exist?
(a) (−∞, ∞) (b) ( π2 , π2 ) (c) (0, π2 ) (d) ( π2 , 3π

2 )
Solution Rewrite the given equation,
sint t et
y 00 + y0 +

y= .
sin t cos t
sin t sin t
(2n − 1)
So the discontinuities of p are t = π for all integer n while the discontinuities of q
2
and r are t = nπ for all integer n.
π π π
Since t0 = lies in (0, ), the interval of validity is (0, ). Then (c) is the final answer.
4 2 2

Consider the IVP
t+2 0
(t + 5)y 00 + y + ln(t)y = 0, y(1) = −10, y 0 (1) = 3.
t−π
anteed to exist?
(a) (−∞, ∞) (b) (−5, π) (c) (−2, ∞) (d) (0, π)
Solution Change the form of the equation by dividing out by t + 5,
t+2 ln(t)
y 00 + y0 + y=0
(t − π)(t + 5) t+5
Then p has discontinuities at t = π, −5, q has no discontinuity since t = −5 is not in the

domain of ln t while r is continuous everywhere.
Since t0 = 1 lies in (0, π), the largest such interval is (0, π) and so (d) is the correct answer.

What is the largest interval on which a solution of the IVP is guaranteed to exist? Do not
solve the differential equation.
1
y 00 + ln |t − 10|y 0 + y = , y(8) = 8, y 0 (8) = 0.
t2 − t
anteed to exist?
Solution Since t2 − t = t(t − 1), we have
1
y 00 + ln |t − 10|y 0 + y = .
t(t − 1)
Then p is continuous everywhere but is not defined at t = 10, q is continuous while r has
discontinuities at t = 0, 1.
Since t0 = 8 lies between (1, 10), the interval of validity is (1, 10).
Exercises 3.4
t−1 0 t
(t2 − 16)y 00 + y + y = 0, y(3) = 10, y 0 (3) = −2.
t+π 2−t
What is the largest interval in which a unique solution is guaranteed to exist?
(a) (2, ∞)
(b) (−π, 2)
(c) (−4, 4)
(d) (2, 4)
2. The Existence and Uniqueness Theorem guarantees that the solution to
ty e2t
(t + 2)y 00 − sin ty 0 + = , y(−1) = 0
t−4 t
(a) is valid on (−∞, ∞).
(b) is valid on (−π, 0).
(c) is valid on (−2, 0).
(d) does not exist.
t+2 0
(t + 5)y 00 + y + ln(t)y = 0, y(1) = −10, y 0 (1) = 2.
t−π
What is the largest interval in which a unique solution is guaranteed to exist?
(a) (−∞, −5)
(b) (−5, π)
(c) (−2, ∞)
(d) (0, π)
4. The largest interval on which the differential equation
(t2 − 1)y 00 + sin(t)y 0 + cos(t)y = 0, y(5) = 0, y 0 (5) = 1.
is certain to have a unique twice differentiable solution is
(a) (−∞, 5)
(b) (−1, 1)
(c) (5, ∞)
(d) (1, ∞)
(e) (−∞, −1)
Answers
1. d)
2. c)
3. d)
4. d)
3.5 More on Wronskian & Abel’s Theorem

Let y1 and y2 be solutions of y 00 + p(t)y 0 + q(t)y = g(t). Then y1 and y2 form a fundamental set
of solution if their Wronskian is not a zero function. And the Wronskian can be found by

y y
1 2
W (y1 , y2 ) := = y1 y20 − y10 y2 .

y 0 y 0
1 2
Let’s consider a few more examples about it here.

Are y1 (t) = t2 , y2 (t) = t−2 a fundamental set of solutions of t2 y 00 + ty 0 − 4y = 0?
Solution We completely answer this question in 2 steps as follow:
• Check y1 (t) = t2 , y2 (t) = t−2 are solutions to t2 y 00 + ty 0 − 4y = 0.
t2 y100 + ty10 − 4y1 = t2 (t2 )00 + t(t2 )0 − 4(t2 ) = 2t2 + 2t2 − 4t2 = 0
t2 y200 + ty20 − 4y2 = t2 (t−2 )00 + t(t−2 )0 − 4(t−2 ) = t2 (6t−4 ) + t(−2t−3 ) − 4t−2 = 0
• Compute their Wronskian:

t2 t−2
= −2t−1 − 2t−1 = −4t−1 6= 0.

W (y1 , y2 ) =
2t −2t−3

Therefore y1 and y2 form a fundamental set of solutions of t2 y 00 + ty 0 − 4y = 0.

Which pair of functions below can be a fundamental set of solutions?
a) 1 − 2t, 4t − 2
b) 3 cos t, −2 sin t
c) −e2t , e2t+5
d) 0, 3 sin 2t
Solution Let’s consider each choice one by one,

1 − 2t 4t − 2

(a) W (1 − 2t, 4t − 2) = = (4 − 8t) − (−8t + 4) = 0.

−2 4

3 cos t −2 sin t
= −6 cos2 t − 6 sin2 t = −6 6= 0.

(b) W (3 cos t, −2 sin t) =
−3 sin t −2 cos t

(Here we use the identity: sin2 t + cos2 t = 1.)

−e2t e2t+5
2t 2t+5

(c) W (−e , e ) = = −2e4t+5 − (−2e4t+5 ) = 0.

−2e2t 2e 2t+5

0 3 sin 2t

(d) W (0, 3 sin 2t) = = 0 − 0 = 0.

0 6 cos 2t

So b) is the correct answer since it it the only one which gives us a nonzero function Wron-
skian.
Besides the definition, there is another way to compute Wronkskian without knowing solutions y1
and y2 by using Abel’s theorem.
Theorem 6 (Abel’s Theorem).

If y1 and y2 are solutions of
y 00 + p(t)y 0 + q(t)y = 0,
then R
W (y1 , y2 )(t) = ce− p(t) dt
,
where c is certain constant depends on y1 and y2 but not on t.
a) Let y1 , y2 be two solutions to the equation
ty 00 − 2y 0 − y = 0.
Determine the Wronskian W (y1 , y2 ) of y1 , y2 .

Solution Rewrite the differential equation into the standard form,
2 y
y 00 − y 0 − = 0.
t t
So p(t) = − 2t . By Abel’s theorem,
− 2t dt
R R
W (y1 , y2 )(t) = ce− p(t) dt
= ce− = ce2 ln t = ct2 .
b) If W (y1 , y2 )(2) = 1, then determine W (y1 , y2 )(3)?

Solution If W (y1 , y2 )(2) = 1, then by part a) we have
1
1 = c(2)2 ⇒ c= .
4
t2
Hence W (y1 , y2 )(t) = and so
4
32 9
W (y1 , y2 )(3) = = .
4 4

Consider the second order linear differential equation
ty 00 − 2y 0 + y = 0.
Suppose y1 (t) and y2 (t) are two fundamental sets of solutions of the equation satisfying
y1 (1) = 2, y10 (1) = 0, y2 (1) = 2, y20 (1) = 2.
Compute their Wronskian W (y1 , y2 )(t) as a function of t.

Solution Again write down the differential equation in the standard form,
2 y
y 00 − y 0 + = 0.
t t
So p(t) = − 2t . By Abel’s theorem,
− 2t dt
R R
W (y1 , y2 )(t) = ce− p(t) dt
= ce− = ce2 ln t = ct2 . (3.1)
Note here that

W (y1 , y2 )(1) = c(1)2 = c (3.2)
On the other hand, by definition, we know

y y
1 2
W (y1 , y2 )(t) = = y1 (t)y20 (t) − y10 (t)y2 (t).

y 0 y 0
1 2
∴ W (y1 , y2 )(1) = y1 (1)y20 (1) − y10 (1)y2 (1) = 2(2) − 0(2) = 4. (3.3)
Comparing Equations (3.2) and (3.3), c = 4. Hence
W (y1 , y2 )(t) = 4t2 ,
according to Equation (3.1).

Exercises 3.5
1. Let y1 (t) and y2 (t) be any two solutions of the second order linear equation
(t + 2)y 00 + 4y 0 + (t2 + 9)e5t y = 0.
What is the general form of their Wronskian, W (y1 , y2 )(t)?
(a) ce−4t
(b) ce4t
(c) c(t + 2)4
c
(d)
(t + 2)4
2. Suppose y1 (t) = et and y2 (t) = cos t are both solutions of the second order linear equation
y 00 + p(t)y 0 + q(t)y = 0.
All of functions below must also be solutions of the same equation, EXCEPT
(a) y = 100π cos t
(b) y = 2et cos t
(c) y = 7et+1 − 3 cos t
(d) y = 0
3. Suppose y1 (t) = 4et and y2 (t) = −2t are two solutions of a certain second order homogeneous
linear equation
y 00 + p(t)y 0 + q(t)y = 0.
(a) Find the Wronskian W (y1 , y2 )(t).
(b) True or false: y1 and y2 form a set of fundamental solutions of this equation. Why
or why not?
(c) Write down a general solution of the differential equation.
(d) Find the particular solution satisfying the initial conditions y(0) = 8 and y 0 (0) = −1.
(e) True or false: y3 = 0 is also a solution of this equation. Why or why not?
(f) True or false: y4 = −8tet is also a solution of this equation. Why or why not?
4. Suppose y1 (t) = 15t3 and y2 (t) = −4t3 ln(t) are two solutions of a certain second order
homogeneous linear equation
y 00 + p(t)y 0 + q(t)y = 0.
(a) Find the Wronskian W (y1 , y2 )(t).

(b) True or false: y1 and y2 form a set of fundamental solutions of this equation. Why
or why not?
(c) Write down a general solution of the differential equation.
(d) True or false: y3 = 9t6 ln(t) is also a solution of this equation. Why or why not?
(e) True or false: y4 = 0 is also a solution of this equation. Why or why not?
y 00 + 2 tan(t)y 0 + t3 y = 0.
What is the general form of their Wronskian, W (y1 , y2 )(t)?

(a) c cos2 t
2
(b) cesec t
(c) ce2 sin t

(d) c sec2 t
6. Which pair of functions below can be a fundamental set of solutions?
(a) 0, 5te2t
(b) sin 3t, − sin 3t
−t
(c) 6e , −6et
(d) 4t − cos t, 3 cos t − 12t
7. Suppose y1 (t) = −t2 and y2 (t) = 3t2 ln(t) are both solutions of the second order linear
equation
y 00 + p(t)y 0 + q(t)y = 0, t > 0.
Which statement below is FALSE?
(a) y = 2t2 − 10t2 ln(t) is also a solution of the equation.
(b) W (y1 , y2 )(t) = 0
(c) y = 0 is also a solution of the equation.
(d) y = 1 can never be a solution of the equation.
8. Show y1 (t) = t2 and y2 (t) = t3 , t > 0 form a fundamental set of solutions for
t2 y 00 − 4ty 0 + 6y = 0.
9. Given the linear ordinary differential equation
4et
ty 00 − 4y 0 + y=0
t
and two fundamental solutions y1 (t), y2 (t) such that
y1 (1) = 1, y10 (1) = 0, y2 (1) = 2, and y20 (1) = 3,
compute their Wronskian W (y1 (t), y2 (t)) as a function of time, using Abel’s Theorem. Use
the initial condition to determine the constant of the Wronskian.
10. Let y1 and y2 be two solutions to the linear equation
2t2 y 00 − ty 0 − y = 0.
Then the Wronskian of y1 and y2 must be a constant multiple of

√
(a) t
(b) t
(c) 1
(d) t2
(t2 + 4)y 00 + 2ty 0 − t3 y = 0.
In what general form must their Wronskian, W (y1 , y2 )(t), appear?
(a) c(t2 + 4)
√
(b) c t2 + 4
c
(c) 2
(t + 4)
(d) c(t2 + 4)2
12. Given that y1 (t) = 1 and y2 (t) = arctan(t) are both solutions of the second order homoge-
neous linear equation
(t2 + 1)y 00 + 2ty 0 = 0.
Determine whether each of the following statements is true or false. State a brief reason
that justifies each answer.
(a) Wronskian W (y1 , y2 )(t) = 0.
(b) y1 and y2 from a set of fundamental solutions of this equation.
(c) y3 (t) = 7 + 5 arctan(t) is also a solution of the equation.
(d) There are additional solutions that cannot expressed as a linear combination of y1 and
y2 .
(e) Each solution is unique to its corresponding initial conditions only on the interval
−1 < t < 1.
13. y1 (t) = tet and y2 (t) = t2 are both solutions of the second order linear differential equation
y 00 + p(t)y 0 + q(t)y = 0.
(a) Compute W (y1 , y2 )(t) = 0.

(b) (TRUE or FALSE) y1 and y2 is not a fundamental set of solutions.
(c) (TRUE or FALSE) y = 0 is a solution of the equation.
(d) (TRUE or FALSE) y = (10t − 4et )t is not a solution of the equation.
Answers
1. d)
2. b)
3. a) −8et + 8tet , b) True since W (y1 , y2 ) 6= 0, c) y(t) = c1 et + c2 t, d) y(t) = 8et − 9t, e) True, f)
False
4. a) −60t5 , b) True since W (y1 , y2 ) 6= 0, c) y(t) = c1 t3 + c2 t3 ln(t), d) False, e) True
5. a)
6. c)
7. b)
8. First substitute y1 and y2 into the equation to verify that they both satisfy it. Then calculate
their Wronskian, W (y1 , y2 )(t) = t4 6= 0 when t > 0.
9. 3t4
10. a)
11. c)
12. a) F, b) T, c) T, d) F, e) F
13. a) t2 et − t3 et , b) F, c) T, d) F
3.6 Nonhomogeneous Equation & Method of Undetermined

Coefficients
For the nonhomongeneous equation,
y 00 + p(t)y 0 + q(t)y = g(t). (3.4)
The general solution to Equation (3.4) can be written as
y(t) = yc (t) + Y (t),
where yc (t) (which is called complementary solution)

= the general solution of the corresponding homogeneous equation y 00 + p(t)y 0 + q(t)y = 0.
Y (t) or yp (t) (which is called particular solution)
= some specific solution of the nonhomogeneous equation (3.4).
In this course we focus on a simpler case where the coefficients are constant:
ay 00 + by 0 + cy = g(t),
where a 6= 0, g(t) 6= 0 and a, b, c are constants. The complementary solution yc (t) (the general
solution to ay 00 +by 0 +cy = 0) can be easily found from the roots of characteristic equation that we
studied before. So the only task remaining is to find the particular solution Y (t). The method
of undetermined coefficients will give us the answer.
Let
g(t) = g1 (t) + · · · + gn (t).
This method gives us the form of
Y (t) = Y1 (t) + · · · + Yn (t),
with undetermined coefficients where the form of Yi (t) depends on the form of gi (t) by the table
below.
gi (t) Yi (t)
1. Pn (t) = a0 tn + a1 tn−1 + · · · + an ts (A0 tn + A1 tn−1 + · · · + An )
2. Pn (t)eαt ts (A0 tn + A1 tn−1 + · · · + An )eαt
3. Pn (t)eαt sin βt, Pn (t)eαt cos βt, ts (A0 tn + A1 tn−1 + · · · + An )eαt cos βt
Pn (t)eαt sin βt + Pn (t)eαt cos βt +ts (B0 tn + B1 tn−1 + · · · + Bn )eαt sin βt
Here s is the smallest nonnegative integer (s = 0, 1, or 2) that will ensure that no term in Yi (t) is
a solution of the corresponding homogeneous equation.
Example 53. Assume that the complementary solution is
yc (t) = c1 e3t cos t + c2 e3t sin t.
gi (t) Yi (t)
3 A
2t At + B
−5t4 − 3t2 + 1 At4 + Bt3 + Ct2 + Dt + E
tet (At + B)et
cos 3t A cos 3t + B sin 3t
−3 sin 2t A cos 2t + B sin 2t
2t sin 5t (At + B) cos 5t + (Ct + D) sin 5t
e3t cos 3t (Ae3t cos 3t + Be3t sin 3t)t
te3t sin 3t {(At + B)e3t cos 3t + (Ct + D)e3t sin 3t}t
t2 e2t sin 3t (At2 + Bt + C)e2t cos 3t + (Dt2 + Et + F )e2t sin 3t
Example 54. Suppose that the complementary solution is
yc (t) = c1 e2t + c2 te2t .
gi (t) Yi (t)
t2 − 5t At2 + Bt + C
2te2t (At + B)e2t t2
sin 2t A cos 2t + B sin 2t
e3t cos 2t Ae3t cos 2t + Be3t sin 2t
3e−t Ae−t
4t2 sin t (At2 + Bt + C) cos t + (Dt2 + Et + F ) sin t
−5t − 3 At + B
2e2t (Ae2t )t2


What is the form of the particular solution of
y 00 + 4y 0 + 4y = te−2t + 2t2 cos 2t − 3?
Solution First consider the corresponding homogeneous equation,
y 00 + 4y 0 + 4y = 0.
Its characteristic equation is r2 + 4r + 4 = 0 which is (r + 2)2 = 0. So r = −2, −2. Then the

complementary solution is
yc (t) = c1 e−2t + c2 te−2t .
Now let’s consider the form of the particular solution Y ,
g1 (t) = te−2t ⇒ Y1 (t) = (At + B)e−2t t2

g2 (t) = 2t2 cos 2t ⇒ Y2 (t) = (Ct2 + Dt + E) cos 2t + (F t2 + Gt + H) sin 2t
g3 (t) = −3 ⇒ Y3 (t) = I.
So the form of the particular solution is
Y (t) = (At + B)e−2t t2 + (Ct2 + Dt + E) cos 2t + (F t2 + Gt + H) sin 2t + I.

Consider the nonhomogeneous second order linear equation of the form
y 00 − 4y 0 + 8y = g(t).
a) Find its complementary solution yc (t).

Solution Consider y 00 − 4y 0 + 8y = 0. Its characteristic equation is r2 − 4r + 8 = 0.
√
4 ± 16 − 32
r= = 2 ± 2i.
2
So yc (t) = c1 e2t cos 2t + c2 e2t sin 2t.
For each of parts b) through d) write down the correct choice of the form of particular solution
that you would use to solve the given equation using the Method of Undetermined Coefficients.
b) y 00 − 4y 0 + 8y = 2e2t − 5t2 + sin 2t.
Solution From g(t) = 2e2t − 5t2 + sin 2t, then
g1 (t) = 2e2t ⇒ Y1 (t) = Ae2t

g2 (t) = −5t2 ⇒ Y2 (t) = Bt2 + Ct + D
g3 (t) = sin 2t ⇒ Y3 (t) = E cos 2t + F sin 2t.
So Y (t) = (Ae2t ) + (Bt2 + Ct + D) + (E cos 2t + F sin 2t).

c) y 00 − 4y 0 + 8y = −e2t sin 2t + 1.
Solution From g(t) = −e2t sin 2t + 1, we have
g1 (t) = −e2t sin 2t ⇒ Y1 (t) = (Ae2t cos 2t + Be2t sin 2t)t

g2 (t) = 1 ⇒ Y2 (t) = C.
So Y (t) = (Ae2t cos 2t + Be2t sin 2t)t + C.

d) y 00 − 4y 0 + 8y = t2 e−t cos 5t.
Solution Since g(t) = t2 e−t cos 5t,
Y (t) = (At2 + Bt + C)e−t cos 5t + (Dt2 + Et + F )e−t sin 5t.

y 00 + 5y 0 + 4y = g(t).
a) Find its complementary solution yc (t).

Solution Consider the corresponding homogeneous equation y 00 + 5y 0 + 4y = 0. Hence
Its characteristic equation is
r2 + 5r + 4 = 0 ⇒ (r + 4)(r + 1) = 0 ⇒ r = −4, −1.
Therefore yc (t) = c1 e−4t + c2 e−t .
For each of parts b) through d) write down the correct choice of the form of particular solution
that you would use to solve the given equation using the Method of Undetermined Coefficients.
b) y 00 + 5y 0 + 4y = 6e−4t + 2et .
Solution From g(t) = 6e−4t + 2et , then
g1 (t) = 6e−4t ⇒ Y1 (t) = Ae−4t t

g2 (t) = 2et ⇒ Y2 (t) = Bet
So Y (t) = Ae−4t t + Bet .

c) y 00 + 5y 0 + 4y = 2t3 e−t .
Solution Since g(t) = 2t3 e−t ,
Y (t) = (At3 + Bt2 + Ct + D)e−t t.
d) y 00 + 5y 0 + 4y = t2 e−4t sin t.
Solution Since g(t) = t2 e−4t sin t,
Y (t) = (At2 + Bt + C)e−4t cos t + (Dt2 + Et + F )e−4t sin t.
After we learned how to find an appropriate form of the particular solution Y (t) by using the
method of undetermined coefficients, we are ready to solve nonhomogeneous equation
ay 00 + by 0 + cy = g(t).
The process of solving nonhomogeneous equation ay 00 + by 0 + cy = g(t)

1. Find the complementary solution yc (t) by solving the corresponding homogeneous equa-
tion ay 00 + by 0 + cy = 0.
2. Find appropriate form of the particular solution Y (t) by using the method of undeter-
mined coefficients.
3. Substitute Y (t) back into ay 00 + by 0 + cy = g(t) to determine the coefficients of Y (t).
4. Then the general solution of ay 00 + by 0 + cy = g(t) is
y(t) = yc (t) + Y (t).

Find the general solution of nonhomogeneous second order linear equation,
y 00 − 4y 0 + 5y = e2t − 10t.
Solution First, find yc (t) by considering y 00 − 4y 0 + 5y = 0. Its characteristic equation is

√
4 ± 16 − 20
r2 − 4r + 5 = 0 ⇒ r = = 2 ± i.
2
Therefore yc (t) = c1 e2t cos t + c2 e2t sin t. Here g(t) = e2t − 10t which gives the form of the
particular solution Y as
Y (t) = Ae2t + Bt + C.
Then Y 0 (t) = 2Ae2t + B and Y 00 = 4Ae2t . Since Y satisfies the given equation,
e2t − 10t = Y 00 − 4Y 0 + 5Y
= (4Ae2t ) − 4(2Ae2t + B) + 5(Ae2t + Bt + C)
= (4Ae2t − 8Ae2t + 5Ae2t ) + 5Bt + (−4B + 5C)
= (4A − 8A + 5A)e2t + 5Bt + (5C − 4B)
= Ae2t + 5Bt + (5C − 4B)
Comparing the coefficients from both sides of the equation,
A = 1, 5B = −10, 5C − 4B = 0.
8 8
So A = 1, B = −2 and C = − . So Y (t) = e2t − 2t − . Thus the general solution is
5 5
8
y(t) = yc (t) + Y (t) = c1 e2t cos t + c2 e2t sin t + e2t − 2t − .
5

y 00 − y 0 − 2y = g(t).
a) Find its complementary solution, yc (t).

Solution The characteristic equation of y 00 − y 0 − 2y = 0 is
r2 − r − 2 = 0 ⇒ (r − 2)(r + 1) = 0 ⇒ r = 2, −1.
So the complementary solution is
yc (t) = c1 e2t + c2 e−t .
b) Find a particular solution Y (t) satisfying y 00 − y 0 − 2y = 3 sin 2t.

Solution Since
g(t) = 3 sin 2t ⇒ Y (t) = A cos 2t + B sin 2t.
Then Y 0 (t) = −2A sin 2t + 2B cos 2t and Y 00 (t) = −4A cos 2t − 4B sin 2t. Since Y is the
particular solution to y 00 − y 0 − 2y = 3 sin 2t, we have
3 sin 2t = Y 00 − Y 0 − 2Y
= (−4A cos 2t − 4B sin 2t) − (−2A sin 2t + 2B cos 2t) − 2(A cos 2t + B sin 2t)
= (−4A − 2B − 2A) cos 2t + (−4B + 2A − 2B) sin 2t
= (−6A − 2B) cos 2t + (2A − 6B) sin 2t.
By the comparison of the coefficients, we have
−6A − 2B = 0, and 2A − 6B = 3.
3 9
Then A = and B = − . So
20 20
3 9
Y (t) = cos 2t − sin 2t.
20 20
c) Write down the correct choice of the form of particular solution that you would use to
solve the equation below using the method of undetermined coefficients.
y 00 − y 0 − 2y = t3 e−t − 7e2t cos 6t.
Solution Since g(t) = t3 e−t − 7e2t cos 6t, we have
Y (t) = (At3 + Bt2 + Ct + D)e−t t + (Ee2t cos 6t + F e2t sin 2t).
a) Find the solution to the homogeneous differential equation
y 00 + 3y 0 − 4y = 0.
Solution The characteristic equation is
r2 + 3r − 4 = 0 ⇒ (r + 4)(r − 1) = 0 ⇒ r = −4, 1.
So the general solution is

y = c1 e−4t + c2 et .
b) Find the general solution to y 00 + 3y 0 − 4y = 50 sin 2t.

Solution From part a), yc (t) = c1 e−4t + c2 et . Since
g(t) = 50 sin 2t ⇒ Y (t) = A cos 2t + B sin 2t.
Then Y 0 (t) = −2A sin 2t + 2B cos 2t and Y 00 (t) = −4A cos 2t − 4B sin 2t. Since Y is the
particular solution to y 00 + 3y 0 − 4y = 50 sin 2t, we have
50 sin 2t = Y 00 + 3Y 0 − 4Y
= (−4A cos 2t − 4B sin 2t) + 3(−2A sin 2t + 2B cos 2t) − 4(A cos 2t + B sin 2t)
= (−4A + 6B − 4A) cos 2t + (−4B − 6A − 4B) sin 2t
= (−8A + 6B) cos 2t + (−6A − 8B) sin 2t.
Comparing the coefficients, we have
−8A + 6B = 0, and − 6A − 8B = 50.
They give A = −3 and B = −4. So Y (t) = −3 cos 2t−4 sin 2t. Thus the general solution
to y 00 + 3y 0 − 4y = 0 is
y(t) = yc (t) + Y (t) = c1 e−4t + c2 et − 3 cos 2t − 4 sin 2t.

Find the general solution of the nonhomogeneous linear equation,
y 00 − 2y 0 + 5y = 5t2 + 6t − 12.
Solution The characteristic equation of y 00 − 2y 0 − 5y = 0 is

√
2 ± 4 − 20
r2 − 2r + 5 = 0 ⇒ r = = 1 ± 2i.
2
Thus yc (t) = c1 et cos 2t + c2 et sin 2t. Since g(t) = 5t2 + 6t − 12, then
Y (t) = At2 + Bt + C.
Then Y 0 (t) = 2At + B and Y 00 = 2A. Since Y satisfies the given equation,
5t2 + 6t − 12 = Y 00 − 2Y 0 + 5Y
= (2A) − 2(2At + B) + 5(At2 + Bt + C)
= 5At2 + (−4A + 5B)t + (2A − 2B + 5C)
Comparing the coefficients,
5A = 5, −4A + 5B = 6, 2A − 2B + 5C = −12.
Then A = 1, B = 2 and C = −2. Therefore Y (t) = t2 + 2t − 2 and so the general solution is
y(t) = yc (t) + Y (t) = c1 et cos 2t + c2 et sin 2t + t2 + 2t − 2.

Exercises 3.6
1. Consider the nonhomogeneous second order linear equation of the form
y 00 − 8y 0 + 16y = g(t).
(a) Find its complementary solution, yc (t).

(b) Find the general solution of y 00 − 8y 0 + 16y = 10e2t .
(c) What is the correct form of particular solution that you would use to solve the equation
below using the Method of Undetermined coefficients? DO NOT ATTEMPT TO
SOLVE THE COEFFICIENTS.
y 00 − 8y 0 + 16y = 9 cos 4t − 5t2 e4t
2. Consider the second order nonhomogeneous linear equation
y 00 − 8y 0 − 9y = 18t + 20.
(a) Find yc (t), the solution of its corresponding homogeneous equation.

(b) Find its general solution.
(c) What is the form of particular solution Y that you would use to solve the following
equation using the Method of Undetermined coefficients? DO NOT ATTEMPT TO
y 00 − 8y 0 − 9y = te−t + 2e9t sin 4t.
y 00 − 10y 0 + 25y = 6e2t .

(b) Use the Method of Undetermined Coefficients to find its general solution.
y 00 − 10y 0 + 25y = 9te5t − 3e−5t cos t.
y 00 + 4y 0 = 2et + t.

(b) Find the general solution of the equation.
y 00 + 4y 0 = 11te−4t (sin 6t − 3 cos 6t) + 7e−4t .
5. What is the form of a particular solution of the equation:
y 00 − 6y 0 + 9y = t2 + e3t − 4 cos t?
Do not solve for the constants.

6. What is the form of the particular solution of
y 00 + 4y 0 + 4y = te−2t + 2t2 cos 2t − 3?
(a) y(x) = (At3 + Bt2 )e−2t + (Ct2 + Dt + E) cos 2t

(b) y(x) = (At2 + Bt)e−2t + (Ct2 + Dt + E) cos 2t + F
(c) y(x) = (At + B)e−2t + (Ct2 + Dt + E) cos 2t + (F t2 + Gt + H) sin 2t + I
(d) y(x) = (At3 + Bt2 )e−2t + (Ct2 + Dt + E) cos 2t + (F t2 + Gt + H) sin 2t + I
2 1
7. Given that y(t) = + 2 is a solution to the differential equation
t t
8 6
y 00 + 2y 0 + 4y = + 4,
t t
find its general solution.
c1 c2 2 1
(a) y = + 4 + + 2
t t t t
√ √ 2 1
(b) y = c1 e−t cos 3t + c2 e−t sin 3t + +
t t2
√ √ 8 6
(c) y = c1 e−t cos 3t + c2 e−t sin 3t + +
t t4
c1 c2
(d) y = + 2
t t
8. In Parts a) and b) determine the form of a particular solution yp = yp (t) having the least
number of unknown constants. DO NOT DETERMINE the unknown constants appear-
ing in your answers in Parts i) and ii).
(a) y 00 − 14y 0 + 49y = 2t2 e7t
(b) y 00 − 50y 0 + 49y = 3tet
9. Of what form will the particular solution to the following differential equation be? Do not
solve the equation.
y 00 − 4y 0 + 4y = e2t + t2 e3t − sin(2πt)
(a) Ae2t + Bt2 e3t + Cte3t + De3t + E sin(2πt) + F cos(2πt)
(b) Ate2t + Bt2 e3t + Cte3t + De3t + E sin(2πt) + F cos(2πt)
(c) At2 e2t + Bt2 e3t + Cte3t + De3t + E sin(2πt) + F cos(2πt)
(d) Ae2t + Bt2 e3t + E sin(2πt) + F cos(2πt)
10. Solve the following initial value problem:
y 00 − 5y 0 − 14y = −14t2 − 10t − 26, y(0) = 0, y 0 (0) = 13.
11. When using the method of undetermined coefficients to solve the following equation, what
is the form of the particular solution? Do not solve for the constants.
y 00 − y 0 − 2y = t2 e3t + 3e−t + te2t cos(2t)

12. Consider the nonhomogeneous second order linear equation of the form
y 00 − 4y 0 + 8y = g(t).
(a) Find its complementary solution, yc (t).
For each of parts b) through d), write down the correct choice of the form of particular
solution that you would use to solve the given equation using the Method of Undetermined
Coefficients? DO NOT ATTEMPT TO SOLVE THE COEFFICIENTS.
(a) y 00 − 4y 0 + 8y = 2e2t − 5t2 + sin 2t
(b) y 00 − 4y + 8y 0 = −e2t sin 2t + 1
(c) y 00 − 4y + 8y 0 = t2 e−t cos 5t
y 00 − 2y 0 − 3y = 3te2t .

(c) Find the solution satisfying the conditions y(0) = 1 and y 0 (0) = 1.
y 00 + 3y 0 + 2y = 20 sin(2t) + 4t.

(c) Find the solution satisfying the conditions y(0) = 0 and y 0 (0) = −2.
Answers
1. a) yc (t) = c1 e4t + c2 te4t , b) y = c1 e4t + c2 te4t + 25 e2t , c) Y = A cos 4t + B sin 4t + (Ct4 + Dt3 +
Et2 )e4t
2. a) yc = c1 e9t + c2 e−t , b) y = c1 e9t + c2 e−t − 2t − 94 , c) Y = (At2 + Bt)e−t + Ce9t cos(4t) +
De9t sin(4t)
3. a) yc = c1 e5t + c2 te5t , b) y = c1 e5t + c2 te5t + 23 e2t , c) Y = (At3 + Bt2 )e5t + Ce−5t cos(t) +
De−5t sin(t)
4. a) yc = c1 + c2 e−4t , b) y = c1 + c2 e−4t + 25 et + 18 t2 − 16 1
t,
−4t −4t −4t
c) Y = (At + B)e cos 6t + (Ct + D)e sin 6t + Ete
5. Y = At2 + Bt + C + Dt2 e3t + E cos t + F sin t
6. d)
7. b)
8. a) (At2 + Bt + c)e7t t2 , b) (At + B)et t
9. c)
10. y(t) = e7t − 3e−2t + t2 + 2
11. Y (t) = (At2 + Bt + C)e3t + Dte−t + (Et + F )e2t cos 2t + (Gt + H)e2t sin 2t
12. a) yc = c1 e2t cos 2t + c2 e2t sin 2t, b) Ae2t + Bt2 + Ct + D + E cos 2t + F sin 2t
c) Ate2t cos 2t + Bte2t sin 2t + C, d) (At2 + Bt + C)e−t cos 5t + (Dt2 + Et + F )e−t sin 5t
13. a) yc = c1 e−t + c2 e3t , b) Y (t) = c1 e−t + c2 e3t − (t + 32 )e2t , c) Y (t) = 23 e−t + e3t − (t + 23 )e2t
14. a) yc = c1 e−2t + c2 e−t , b) Y (t) = c1 e−2t + c2 e−t − sin(2t) − 3 cos(2t) + 2t − 3, c) Y (t) =
−4e−2t + 10e−t − sin(2t) − 3 cos(2t) + 2t − 3
3.7 Modeling with Second Order Differential Equation: Me-

chanical Vibrations
A mass m is hooked at the end of the spring. Its weight stretches the spring by a length L to
reach the equilibrium position. After that the mass is then set in the motion. We want to find
the displacement of the mass from its equilibrium position at any time.
Figure 3.1: A spring mass system
Let u(t), measured positive downward, denote the displacement of the mass from its equilibrium
position at time t. The differential equation governing the movement of the mass is
mu00 (t) + γu0 (t) + ku(t) = F (t), u(t0 ) = u0 , u0 (t0 ) = u00 ,
where constant m, k > 0, γ ≥ 0 and
• m = mass (m > 0) (kg) which can be found from

w
m= .
g
• w = weight of the object (lb).

• g = gravitational constant (10 m/s2 or 32 ft/s2 ).
• γ = damping constant (γ ≥ 0) which can be found from
Fd
γ= .
u0
• Fd = damping force or resistive force.
• u0 = velocity.
• k = spring constant (k > 0) which can be found from
mg w
k= = .
L L
• L = elongation of the spring caused by the weight of mass.

• F (t) = external force.
• u(t0 ) = initial displacement of the mass.

• u0 (t0 ) = initial velocity of the mass.
Modeling for Mechanical Vibrations
mu00 (t) + γu0 (t) + ku(t) = F (t), u(t0 ) = u0 , u0 (t0 ) = u00 ,
as the IVP for mechanical vibrations in which its coefficients can be found from
w Fd mg w
m= , γ= , k= = ,
g u0 L L
where
• m = mass (m > 0) (kg).
• w = weight of the object (lb).
• g = gravitational constant. (10 m/s2 or 32 ft/s2 )

• γ = damping constant (γ ≥ 0).
• Fd = damping force or resistive force.
• u0 = velocity.
• k = spring constant (k > 0).

• L = elongation of the spring caused by the weight of mass.
• F (t) = external force.
• u(t0 ) = initial displacement of the mass.

• u0 (t0 ) = initial velocity of the mass.
Example 62.
A mass weighing 4 lb stretches a spring 2 in. Suppose that the mass is displaced in additional
6 in the positive direction and then released. The mass is in a medium that exerts a vicious
resistance of 6 lb when the mass has a velocity of 3 ft/s. Under the assumption, formulate
the IVP that governs the motion of the mass.
Solution We will measure t in second and displacement u in feet. The initial conditions are
6 1
u(0) = = , u0 (0) = 0.
12 2
We also have
w 4 lb 1
m= = 2 = 8
g 32 ft/s
Fd 6 lb
γ= 0 = =2
u 3 ft/s
w 4 lb
k= = 2 = 24.
L 12 ft
Nothing is said about the external force, F (t) = 0. So the IVP describes the u(t) is
1 00 1
u + 2u0 + 24u = 0, u(0) = , u0 (0) = 0.
8 2
Now we ’re taking a closer look to the solution of the modeling of mechanical vibrations
mu00 (t) + γu0 (t) + ku(t) = F (t)
in a few special cases.
3.7.1 Undamped Free Vibrations

Suppose that the system has neither damping (γ = 0) nor external force (F (t) = 0). Then the
equation becomes
mu00 + ku = 0.
q
k
Its characteristic equation is mr2 + k = 0 which gives roots r1,2 = ± m i.
u(t) = c1 cos w0 t + c2 sin w0 t, (3.5)
q
k
where w0 = m .
• w0 is called the natural frequency of the system.
• The (natural) period of the motion T is given by

r
2π m
T = = 2π .
w0 k
To get a clear picture of what’s going on, we rewrite the solution of u(t) to the form
u = R cos(w0 t − δ), (3.6)
where R > 0 is chosen. Note that Equation (3.6) can be written as
u = R cos(w0 t) cos(δ) + R sin(w0 t) sin(δ). (3.7)
Comparing Equations (3.5) and (3.7), we have
c1 = R cos δ, c2 = R sin δ.
These two relations give,

c2 R cos δ
= = tan δ.
c1 R sin δ
Also
q
c21 + c22 = (R cos δ)2 + (R sin δ)2 = R2 (cos2 δ + sin2 δ) = R2 ⇒ R= c21 + c22 .
• The angle δ is called the phase of displacement.

• R is the amplitude of the motion.
Using the new form of the solution, it is easier to graph as it is shown below.
We summarize all the details here:
Figure 3.2: Undamped free vibration
Undamped Free Vibrations (γ = 0, F (t) = 0)

If we assume that the system has neither damping (γ = 0) nor external force (F (t) = 0),
mu00 + ku = 0
gives the general solution as
u(t) = c1 cos w0 t + c2 sin w0 t = R cos(w0 t − δ) ⇒ lim u(t) does not exist.

t→∞
where
c2
q
tan δ = , R= c21 + c22 ,
c1
• w0 = the natural frequency of the system.
r
2π m
• T = the (natural) period of the motion, T = = 2π .
w0 k
• δ = the phase of displacement.
• R = amplitude of the motion.
Example 63 (Fall 2007 Math 250 Midterm Exam I).

A mass of 2 kg stretches a spring 10 cm. The mass is pulled down 20 cm from the equilibrium
position and then released with downward initial velocity 2 m/s. Ignore air resistance and
take g=10 m/s2 .
a) Write down the differential equation governing the motion of the mass.
Solution From the information above, we have
20 1 10 1
u(0) = = m, u0 (0) = 2 m/s2 , m = 2 kg, L= = m
100 5 100 10
Since we ignore air resistance, γ = 0. There is no external force mentioned so F (t) = 0.
Also we can find the spring constant from,
mg 2(10)
k= = 1 = 200.
L 10
So the IVP governing the motion of the mass is

1
2u00 + 200u = 0, u(0) = , u0 (0) = 2.
5
b) Determine the position of the mass at any time t.

Solution From part a), we have u00 + 100u = 0. Its characteristic equation is
r2 + 100 = 0 ⇒ r = ±10i.
So we have u(t) = c1 cos 10t+c2 sin 10t. Hence u0 (t) = −10c1 sin 10t+10c2 cos 10t. Then
apply initial conditions,
1 1
u(0) = ⇒ c1 =
5 5
u0 (0) = 2 ⇒ 10c2 = 2.
1
Therefore c1 = c2 = 5 and then
1 1
u(t) = cos 10t + sin 10t.
5 5
c) Find an amplitude, frequency, period and phase of the motion.

Solution
q q
k 200
• Frequency: w0 = m = 2 = 10.
2π 2π π
• Period: T = = = .
w0 10 5
p q √
2
• Amplitude: R = c21 + c22 = ( 15 )2 + ( 15 )2 = 5 .
1
c2 π
• Phase: Since tan δ = = 5
1 =1 ⇒ δ = tan−1 1 = .
c1 5
4
3.7.2 Damped Free Vibrations

If we include the effect of damping to the system (γ > 0) without the external force (F (t) = 0),
then the differential equation of u(t) is
mu00 + γu0 + ku = 0.
The roots of corresponding characteristic equation mr2 + br + c = 0 is

p
−γ ± γ 2 − 4mk
r1,2 = .
2m
Depending on the sign of γ 2 − 4mk, the solution u has one of the following form:
1. γ 2 − 4mk > 0. This mass-spring system is called overdamped.

In this case the roots of characteristic equation are two distinct real roots, both actually are
negative. Hence
u(t) = c1 er1 t + c2 er2 t ,
where both r1 6= r2 are negative. Then lim u(t) = 0. This system does not oscillate and it
t→∞
could cross the equilibrium position at most once.
2. γ 2 − 4mk = 0. We call this mass-spring system critically damped.

γ
For this case, r = r1 = r2 = − 2m which are repeated negative real roots. So
u(t) = c1 ert + c2 tert
and hence lim u(t) = 0.

t→∞
t
Figure 3.3: Critically damped motions: u00 + u0 + 0.25u = 0; u = (A + Bt)e− 2
Like the first case, the mass could cross its equilibrium position at most one time.
3. γ 2 − 4mk < 0 (the most interesting case). A system is called underdamped. This case we
have p
−γ 4mk − γ 2
r1,2 = ± i.
2m 2m
So the solution is
u(t) = c1 eλt cos µt + c2 eλt sin µt
√
γ 4mk−γ 2
where λ = − 2m < 0 and µ = 2m > 0. Since λ < 0, we have lim u(t) = 0. Also we
t→∞
can use the same tactic that we used in undamped free vibration case to rewrite the solution
to
u(t) = Reλt cos(µt − δ),
p c2
where R = c21 + c22 and tan δ = . Here
c1
• µ is called the quasi frequency of the system.
2π
• Td is called the quasi period which can be found from Td = µ .
In this case the mass crosses its equilibrium position infinitely many times. The displacement
function u(t) is oscillating (not periodic) but the amplitude is decaying exponentially.
Note here that in all 3 cases, the displacement function u(t) tends to zero as t → ∞ reflecting the
fact that we we include the effect of damping to the system.

A mass-spring system is described by the initial value problem,
u00 + 4u0 + 20u = 0, u(0) = 2, u0 (0) = 0.
a) Find the real-valued particular solution of the initial value problem.

Figure 3.4: Underdamped motion
Solution Consider its characteristic equation,

√
−4 ± 16 − 80
0 = r2 + 4r + 20 ⇒ r= = −2 ± 4i.
2
y(t) = c1 e−2t cos 4t + c2 e−2t sin 4t.
Then apply the initial conditions to receive c1 = 2 and c2 = 1. (check!) So the particular
solution is
y(t) = 2e−2t cos 4t + e−2t sin 4t.
b) What are the quasi-frequency and quasi-period of the mass-spring system?

Solution The quasi frequency can be found from
p p √
4mk − γ 2 4(1)(20) − 42 64
µ= = = = 4.
2m 2 2
Therefore the quasi period of the system is
2π 2π π
Td = = = .
µ 4 2
c) True or False: Every nonzero solution of this mass-spring system will cross the equilib-
rium position more than once.
Solution True since this system is underdamped (γ 2 −4mk = 42 −4(1)(20) = −64 < 0).
(Indeed, it crosses the equilibrium position infinitely many times.)

A mass-spring system is described by the following equation,
3u00 + γu0 + 12u = 0.
a) If γ = 6, what is the quasi-period of this mass-spring problem?

Solution The system is underdamped since
γ 2 − 4mk = 62 − 4(3)(12) = −108 < 0.

Then its quasi frequency µ is

√ √
6 3 √
p
4mk − γ 2 108
µ= = = = 3.
2m 2(3) 6
2π 2π
So the quasi period is Td = =√ .
µ 3
b) For which values of γ, the system is overdamped.
Solution The system is over damped when
γ 2 − 4mk > 0 ⇒ γ 2 − 4(3)(12) > 0 ⇒ γ 2 − 144 > 0.
And so
(γ − 12)(γ + 12) > 0 ⇒ γ > 12 or γ < −12.
But for the mass-spring system, γ ≥ 0. Hence γ > 12 is the final answer.
c) For which values of γ, there is a solution that crosses equilibrium position exactly 5
times.
Solution None, it could not happen since a solution either crosses equilibrium position
at most once (for overdamped or critically damped system) or infinitely many times (for
underdamped system).
3.7.3 Underdamped Forced Vibrations

Now we assume that the system has no damping (γ = 0) but we allow external force to affect the
system (F (t) 6= 0). The motion equation now then becomes
mu00 + ku = F (t).
We are interested in the case where F (t) is a periodic function. Let us assume that F (t) = F0 cos wt
(or F (t) = F0 sin wt). So now we have
mu00 + ku = F0 cos wt.
This is a nonhomogeneous linear equation with the complementary solution
uc (t) = c1 cos w0 t + c2 sin w0 t,

r
k
where w0 = (It is the solution to mu00 + ku = 0, undamped free vibration). The form of the
m
particular solution depends on the value of w by the undetermined coefficients method.
1. Beat (when w 6= w0 )
In this case, the form of the particular solution corresponding to the forcing function is
Y = A cos wt + B sin wt.
After finding A and B, we will get

F0
Y = cos wt.
m(w02 − w2 )
So the general solution of the displacement function is

F0
u(t) = c1 cos w0 t + c2 sin w0 t + cos wt.
m(w02 − w2 )
If the initial conditions are u(0) = 0 and u0 (0) = 0, we get

F0
c1 = − , c2 = 0.
m(w02 − w2 )
Therefore the particular solution to the system is
F0
u(t) = (cos wt − cos w0 t).
m(w02 − w2 )
Then we use this trig identity
cos(A − B) − cos(A + B)
sin A sin B = ,
2
we will change the form of u to
2F0 (w0 − w)t (w0 + w)t
u(t) = sin sin .
m(w02 − w2 ) 2 2
Figure 3.5: Beat; sol of u00 + u = 0.5 cos 0.8t, u(0) = 0, u0 (0) = 0; u = 2.77778(sin 0.1t)(sin 0.9t).
This type of motion, possessing a periodic variation of amplitude, exhibits what is called a
beat.
2. Resonance (when w = w0 )
In this case, since w = w0 , the form of the particular solution is
Y = (A cos w0 t + B sin w0 t)t = At cos w0 t + Bt sin w0 t.

F0
Then we find that A = 0 and B = 2mw0 . Hence the general solution is
F0
u(t) = c1 cos w0 t + c2 sin w0 t + t sin w0 t.
2mw0
The first two terms can be combine to R cos(w0 t − δ) representing the steady oscillation.
While the last term will become unbounded as t → ∞ regardless of the values of c1 and c2 .
So
F0
u(t) = R cos(w0 t − δ) + t sin w0 t.
2mw0
This phenomenon is called resonance.
Figure 3.6: Resonance; solution of u00 + u = 0.5 cos t, u(0) = 0, u0 (0) = 0; u = 0.25t sin t.

True or false: a mass-spring system is described by the equation
2u00 + 2u = 5 cos t,
then the mass-spring system is undergoing resonance?

Solution Since F (t) = 5 cos t = F0 cos wt, we have w = 1. On the other hand,
r r
k 1
w0 = = = 1.
m 1
Since w = w0 , the mass-spring system is undergoing resonance.
Exercises 3.7
1. A mass weighing 0.25 kg stretches a spring 1.25 m. The mass-spring system has a damping
constant of 12 kg/s. At t = 0 the mass is displaced an additional 50 cm downward from its
equilibrium position and set in motion with an upward velocity of 2 m/s. (You may use
g = 10 m/s2 as the gravitational constant. )
(a) Write an initial value problem that describes the motion of the mass.
(b) Determine the system’s quasi frequency and quasi period.
2. A mass-spring system is described by the equation
5u00 + γu0 + ku = F (t).
(a) Suppose the mass originally stretched the spring 2 m to reach its equilibrium position.
What is the spring constant k? (Assume g = 10 m/s2 to be the gravitational constant)
(b) Suppose k = 45. For what value(s) of γ would this system be critically damped?
(c) Suppose γ = 0 and k = 400. What is the natural frequency of this system?
A second mass-spring system is described by the initial value problem
2u00 + 12u0 + 20u = 0, u(0) = 0, u0 (0) = 2.

3π
(a) Find the position of the mass when t = 2 .
(b) What is the quasi-frequency of this system?
3. A 2 kg mass is placed on a spring with k = 8. At t = 0, the system is suddenly set in motion

from its equilibrium position by an external force given by 2 cos(ωt), where ω is a positive
constant. For which value of ω will the system have resonance?
4. Consider a mass-spring system described by the equation
3u00 + 12u0 + ku = 0, k > 0.
Answer the following questions. Be sure to justify your answer. Full credit will not be given
without supporting work.
(a) For what value(s) of k would the system be overdamped?
(b) When k = 12, determine whether the system is overdamped, underdamped or critically
damped.
(c) The system would oscillate when: k = 9 or k = 15. (Circle the correct value and justify
your answer.)
(d) Find the quasi-period of the system whose k-value you found in part iii).
(e) True or false: When k = 3 the mass will never cross the system’s equilibrium position
more than once.
5. A mass of 2 kg stretches a spring 0.4 m. The system has no damping. At t = 0, the mass
is pulled down 1 m from its equilibrium position and set in motion with with an initial
downward velocity of 4 m/s. You may use g = 10 m/s2 as the gravitational constant.
(a) Find the Hooke’s constant, k, of the spring.

(b) Set up, but do not solve, an initial value problem to find the system’s displacement
function u(t).
6. When an object with mass 5 kg is attached to a spring, the object stretches the spring by 2
m. A damper with damping coefficient of 4 N/m is attached to the system. Assume there
is no external force acting on the system and that acceleration due to gravity 10 m/s2 . If
the object is released 1 m above its equilibrium position and is given an initial downward
velocity of 3 m/s, which initial value problem describes the displacement of the mass from
its equilibrium position? Take the downward direction to be positive for all displacements
and forces.
(a) 5y 00 + 4y 0 + 25y = 0, y(0) = 1, y 0 (0) = −3
(b) 5y 00 + 4y 0 + 25y = 0, y(0) = −1, y 0 (0) = 3
(c) 5y 00 + 4y 0 + 10y = 0, y(0) = 1, y 0 (0) = −3
(d) 5y 00 + 4y 0 + 10y = 0, y(0) = −1, y 0 (0) = 3
7. A spring-mass system is modeled by the initial value problem
2y 00 + γy 0 + 8y = F (t), γ ≥ 0, y(0) = 3, y 0 (0) = −4.
(a) If γ = 0 and F (t) = 0, what is the amplitude of displacement?

(b) If γ = 0 and F (t) = 3 cos(ωt), for which value(s) of ω will the system undergo resonance?
(c) If F (t) = 0, for which value(s) of γ will the system not oscillate?
8. A spring-mass system, subject to an external force of 10 cos(2t) Newtons, is equipped with
a spring with a Hooke’s constant 12 Newtons per meter. For what mass will the resonance
occur?
(a) 2 kg
(b) 3 kg
(c) 6 kg
(d) 10 kg
9. Consider a mass-spring system described by the equation
6u00 + γu0 + ku = 0, γ ≥ 0, k > 0.
Answer the following questions. Be sure to justify your answer. Full credit will not be given
without supporting work.
(a) Suppose the spring was stretched 4 meters by the mass to its equilibrium position. Find
the value of k. You may use g = 0 as the gravitational constant.
(b) Suppose k = 20. For what value(s) of γ would this system be critically damped?
(c) Suppose γ = 9 and k = 6. Will any nonzero solution of the equation cross the equilib-
rium position exactly once?
(d) Suppose γ = 10 and k = 10. Find the quasi frequency of the system.
(e) Suppose a force of F (t) = 100 cos(αt) is applied to the system and given k = 150. What
are the values of γ and α if the system exhibits resonance?
10. A mass-spring system is described by the equation
5u00 + γu0 + 80u = 0.
(a) When γ = 0, what is the system’s natural period?

(b) When γ = 40, find the displacement u(t) that satisfies u(0) = 6 and u0 (0) = −4.
(c) When γ = 40, is the system underdamped, overdamped, or critically damped?
(d) True or False: Regardless of the initial conditions, no solution of the system described
in (b) can cross the equilibrium position more than once.
11. A mass-spring system is described by the initial value problem
4u00 + γu0 + 20u = 0, u(0) = 0, u0 (0) = 4.
(a) Suppose γ = 8. Find the real-valued particular solution of this initial value problem.
(b) What is the quasi-period of this mass-spring system described in a)?
(c) True or false: Some, but not at all, nonzero solutions of this mass-spring system, with
γ = 8 (regardless of initial conditions), will cross the equilibrium position more than
once.
(d) Find all the value(s) of γ that would make the system to be overdamped.
12. Suppose a mass-spring system described by the equation
u00 + ku = 2 cos 4t − sin 4t
is undergoing resonance. What is the value of the spring constant k?

(a) 0
(b) 2
(c) 4
(d) 16
Answers
√
1. a) 41 u00 + 12 u0 + 2u = 0, u(0) = 12 , u0 (0) = −2, b) µ = 7, T = √
2π
7
√ −9π
2. a) k = 25, b) γ = 30, c) ω0 = 4 5, d) u = −2e 2 , e) µ = 1
3. 2
4. a) 0 < k < 12, b) critically damped, c) k = 15, d) T = 2π, v) True
5. a) k = 50, b) 2u00 + 50u = 0, u(0) = 1, u0 (0) = 4
6. b) √
7. a) 13, b) 2, c) γ ≥ 8
8. b)
√ √
9. a) 15, b) 4 30, c) No, d) 635 , e) 5
10. a) π2 , b) u(t) = 6e−4t + 20te−4t , c) critically damped, d) T
√
11. a) u(t) = 2e−t sin 2t, b) π, c) F, d) γ > 8 5
12. d)
Chapter 4
Higher Order Linear Equation
4.1 General Theory of nth Order linear Equations

The mathematical theory of nth order differential equation is completely analogous to those for
the 2nd order equation.
Theorem 7 (Existence and Uniqueness Theorem for nth order linear equation).
If the functions p1 , p2 , . . . , pn , and g are continuous on the open interval I, then there exists
exactly one solution y = φ(t) of the differential equation
y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pn (t)y = g(t)
that also satisfies

(n−1)
y(t0 ) = y0 , y 0 (t0 ) = y00 , ... , y (n−1) (t0 ) = y0 ,
where t0 is any point in I. This solution exists throughout the interval I.
Example 67. Determine intervals in which solutions are sure to exist.
(t2 − 4)y (6) + t2 y 000 + 9y = 0.
Solution Writing the equation in standard form,
t2 9
y (6) + y 000 + y = 0.
(t − 2)(t + 2) (t − 2)(t + 2)
Then pi and g have discontinuities at t = −2, 2.

Hence the intervals of validity are (−∞, −2), (−2, 2) and (2, ∞).
Also we have the generalization of the principle of superposition.
Theorem 8. If y1 (t), . . . , yn (t) are solutions to
y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pn (t)y = 0
on an interval I, then so is their linear combination y = c1 y1 + . . . + cn yn .
95
CHAPTER 4. HIGHER ORDER LINEAR EQUATION 96
It is then natural to ask whether every possible solution of
y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pn (t)y = 0
can be expressed as a linear combination of y1 , . . . , yn ,
c1 y1 + c2 y2 + . . . + cn yn .
Just as for second order linear equation, we have the following theorem involving the Wronskian.
Theorem 9. If the functions p1 , p2 , . . . , pn , and g are continuous on the open interval I, if

the functions y1 , . . . , yn are solutions of
y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pn (t)y = 0,
and if W (y1 , y2 , . . . , yn )(t) 6= 0 for at least one point in I, then every solution of the above
homogeneous equation can be expressed as a linear combination of the solutions y1 , y2 , . . . , yn .
In this theorem, the Wronskian of y1 , y2 , . . . , yn is defined by

y
1 y2 ··· yn

y0 y20 ··· 0
yn
1
W (y1 , y2 , . . . , yn )(t) :=
... ..
.
..
.

y (n−1) y (n−1) · · · (n−1)
yn
1 2
and a set of solutions y1 , y2 , . . . , yn of
y (n) + p1 (t)y (n−1) + · · · + pn−1 (t)y 0 + pn (t)y = 0,
whose Wronskian is not a zero function is referred to as a fundamental set of solutions.
Linearly Independent vs Linearly Dependent:

• The functions f1 , f2 , . . . , fn are said to be linearly dependent on an interval I if there
exists a set of constants k1 , k2 , . . . , kn , not all zero, such that
k1 f1 (t) + k2 f2 (t) + · · · + kn fn (t) = 0
for all t in I.
• The functions f1 , f2 , . . . , fn are said to be linearly independent on I if they are not linearly
dependent there.
• In special case n = 2,
f and g are linearly dependent on I if we can find a constant c such that f = cg on I.

f and g are linearly independent on I if we assume c1 f + c2 g = 0 then it implies that
c1 = c2 = 0.
Wronskian vs Linear (In)dependence:

There is a close relationship between the Wronkian and linear (in)dependence we’ve just talked
about.
Theorem 10. If f and g are differentiable functions on an open interval I such that
W (f, g)(t0 ) 6= 0 for some point t0 in I, then f and g are linearly independent on I. Equiva-
lently, if f and g are linearly dependent on I, then W (f, g)(t) = 0 for all t in I.
It is worth noting that

• The converse of theorem above does not hold in general.
• That is to say, it is not true in general that two differentiable functions f and g with a
vanishing Wronskian on an open interval must be linearly dependent on the same interval.
• But the converse may hold if additional conditions are imposed on the pair of functions
involved. The next Theorem is one of them.
Theorem 11. If f and g are differentiable functions on an open interval I such that f (t) 6= 0
and W (f, g)(t) = 0 for all t in I, then f and g are linearly dependent which means f (t) = cg(t)
for a constant c.
In a nutshell, we have
To determine whether f, g are linearly independent or linearly dependent

1. Compute their Wronskian.
2. If their Wronskian is not a zero function then they are linearly independent.
3. Otherwise, find a constant c such that f = cg, then they are linearly dependent. (since
a vanishing Wronskian does not necessarily imply linear dependence)
Example 68. Determine whether each pair of functions is linearly independent or linearly
dependent.
a) e2t − 1, 2 − 2e2t
Solution They are linearly dependent since we can write
2 − 2e2t = (−2)(e2t − 1).
b) 2 cos 4t, 3 sin 4t

Solution Consider their Wronskian,

2 cos 4t 3 sin 4t
= 24 cos2 4t + 24 sin2 4t = 24 6= 0.

W (2 cos 4t, 3 sin 4t) =
−8 sin 4t 12 cos 4t

Therefore they are linearly independent.

c) 2e5t , e5t − 5

2e5t e5t − 5
W (2e5t , e5t − 5) = = 10e10t − (10e10t − 50e5t ) = 50e5t 6= 0.

10e5t 5t
5e

So they are linearly independent.

d) 1, e−2t

1 −2t
e
W (1, e−2t ) = = −2e−2t 6= 0.

0
−2e−2t
So linear independence follows.

e) e2t , e2t+3
Solution They are linearly dependent since
e2t+3 = (e3 )(e2t ).
f ) sin 3t, 0
0 = (0)(sin 3t).
g) e−t , e4t

e−t e4t
−t 4t = 4e3t + e4t = 5e3t 6= 0.

W (e , e ) =
−e−t 4e4t

Hence they are linearly independent.

h) sin 2t, cos t sin t
sin 2t = (2)(cos t sin t),
as the trig identity.

i) sin( π2 − t), 2 cos t
π π π 1
sin( − t) = sin cos t − cos sin t = cos t = ( )(2 cos t).
2 2 2 2
Exercises 4.1
1. Which pair of functions below is linearly independent?
(a) et , −et
(b) sin 3t, 0
2t 2t+1
(c) e , e
−t
(d) e , 2 + e−t
2. Indicate whether the two functions are linearly independent on the interval (−∞, ∞):
(a) e2t , e2t + 2
(b) sin(2t), sin(2t + 2π)
(c) t + 1, 2t + 2
3. All of the following pairs of functions are linearly independent on (−∞, ∞) EXCEPT
a) e−t , e4t
b) et , et − 5
c) e3t , e3t+3
d) cos(2t), sin(2t + 4π)
4. Which of the following pairs of the functions is not linearly independent on (−∞, ∞)?
1 1
(a) e 3 t , e− 3 t
(b) e−t , 2e−(t−4)
(c) 3 cos(πt), 2 sin(πt)
5t
(d) 5, e
5. Let y1 (t) and y2 (t) be two solutions of a second order homogeneous, linear differential equa-
tion. Suppose the Wronskian W (y1 , y2 ) = e−t . Which of the following is FALSE?
(a) y1 (t) and y2 (t) are linearly independent functions.
(b) 2y1 (t) − 3y2 (t) is also a solution of the differential equation.
(c) y1 (t) and y2 (t) do not constitute a fundamental set of solutions.
(d) All solutions of the differential equation can be expressed as c1 y1 + c2 y2 , where c1 and
c2 are constants.
(e) W (2y1 (t), 3y2 (t)) = 6e−t .
6. Which of the following pairs of functions is linearly dependent?

(a) sin 5t, cos 5t
2t
(b) e sin t, e2t cos t
(c) e2t , e4t
(d) et , et+4
1
7. Are the functions y1 (t) = t3 and y2 (t) = t linear independent? Justify your answer.
8. Show y1 (t) = t2 and y2 (t) = t3 , t > 0 are linear independent by calculating the Wronskian.
Answers
1. d)
2. a) yes, b) no, c) no
3. c)
4. b)
5. c)
6. d)
7. yes
8. -
4.2 Homogeneous Equations with Constant Coefficients

Consider the nth order linear homogeneous equation with constant coefficients,
a0 y (n) + a1 y (n−1) + . . . + an y = 0
where a0 6= 0, a1 , . . . , an are constant.

a0 rn + a1 rn−1 + . . . + an = 0.
It is an nth degree polynomial which gives us n roots r1 , r2 , . . . , rn .
To get the general solution, apply the followings:
• If r is a distinct real root, then ert is a solution.
• If r = λ ± µ are complex conjugate roots, then
eλ cos µt, eλ sin µt are solutions
• If r is a real root appearing k times then
ert , tert , ... , tk−1 ert are solutions
• If r = λ ± µ are complex conjugate roots appearing k times then,
eλ cos µt, eλ sin µt

teλ cos µt, teλ sin µt
.. ..
. .
tk−1 eλ cos µt, tk−1 eλ sin µt
are solutions.
• The general solution is the linear combination of all above solutions.


What is the general solution of
y (4) + 16y 00 = 0?
Solution The characteristic equation of y (4) + 16y 00 = 0 is
r4 + 16r2 = 0 ⇒ r2 (r2 + 16) = 0 ⇒ r = 0, 0, ±2i.
y(t) = c1 e0t + c2 te0t + c3 cos 2t + c4 sin 2t = c1 + c2 t + c3 cos 2t + c4 sin 2t.
Example 70. Find the general solution of y (4) − y = 0.

Solution Its characteristic equation is
r4 − 1 = 0 ⇒ (r2 − 1)(r2 + 1) = 0 ⇒ (r − 1)(r + 1)(r2 + 1) = 0.
So r = 1, −1, ±i are roots. Thus the general solution is
y = c1 et + c2 e−t + c3 cos t + c4 sin t.

Find the general solution of
y (5) − 2y (4) + y 000 = 0.
Solution The corresponding characteristic equation is
r5 − 2r4 + r3 = 0 ⇒ r3 (r2 − 2r + 1) = 0 ⇒ r3 (r − 1)2 = 0.
So r = 0, 0, 0, 1, 1 are all roots. Hence the general solution is
y = c1 e0t + c2 te0t + c3 t2 e0t + c4 et + c5 tet = c1 + c2 t + c3 t2 + c4 et + c5 tet .
Example 72. Let (r2 + 6r + 13)3 = 0 be the characteristic equation of certain differential
equation. What is the general solution to that equation?
Solution From
√
2 −6 ± 36 − 52
r + 6r + 13 = 0 ⇒ r = = −3 ± 2i.
2
So for (r2 + 6r + 13)3 = 0, we have r = −3 ± 2i, −3 ± 2i, −3 ± 2i as all roots. Therefore the
general solution is
y = c1 e−3t cos 2t+c2 e−3t sin 2t+t(c3 e−3t cos 2t+c4 e−3t sin 2t)+t2 (c5 e−3t cos 2t+c6 e−3t sin 2t).
Exercises 4.2
1. Find the general solution of the linear equation
y (5) + 16y 000 = 0.
2. Find the general solution of the equation below by examining its characteristic equation.
y 000 + 4y 0 = 0
3. Consider the fourth order linear equation
y (4) − 2y (3) + y 00 = 0.
What is its general solution?

(a) y(t) = c1 + c2 t + c3 et + c4 tet
(b) y(t) = c1 et + c2 e−t + c3 cos t + c4 sin t
(c) y(t) = c1 + c2 t + c3 cos t + c4 sin t
(d) y(t) = c1 + c2 t + c3 e−t + c4 te−t
2y (4) + 50y 00 = 0.

(a) y(t) = c1 et + c2 e−t + c3 cos 5t + c4 sin 5t
(b) y(t) = c1 + c2 t + c3 cos 5t + c4 sin 5t
√ √ √ √
(c) y(t) = c1 cos 5t + c2 sin 5t + c3 t cos 5t + c4 t sin 5t
(d) y(t) = c1 + c2 et + c3 cos 5t + c4 sin 5t
5. Find the general solution of the linear equation
y (5) + 2y (4) + 5y 000 = 0.
y (4) − 8y 00 + 16y = 0.

(a) y(t) = c1 cos 2t + c2 sin 2t + c3 t cos 2t + c4 t sin 2t
(b) y(t) = c1 + c2 t + c3 cos 4t + c4 sin 4t
(c) y(t) = c1 e2t + c2 e−2t + c3 te2t + c4 te−2t
(d) y(t) = c1 + c2 t + c3 e4t + c4 te4t
7. Consider the third order linear equation
y 000 + 4y 00 = 0.
(a) y(t) = c1 + c2 cos 2t + c3 sin 2t

(b) y(t) = c1 + c2 t + c3 e−4t
(c) y(t) = c1 + c2 e2t + c3 e−2t
(d) y(t) = c1 + c2 t + c3 t4
8. Which of the following functions is the general solution of
y 000 + y 00 + y 0 + y = 0.
(a) y(t) = c1 e−t + c2 te−t + c3 t2 e−t

(b) y(t) = c1 et + c2 tet + c3 t2 et
(c) y(t) = c1 cos t + c2 sin t + c3 e−t
(d) y(t) = c1 et + c2 e−t + c3 te−t
y 0000 + 6y 00 + 9y = 0.

(a) y(t) = c1 + c2 t + c3 e3t + c4 te3t
√ √ √ √
(b) y(t) = c1 e 3t
+ c2 e− 3t
+ c3 te 3t
+ c4 te− 3t
(c) y(t) = c1 + c2 t + c3 cos 3t + c4 sin 3t

√ √ √ √
(d) y(t) = c1 cos 3t + c2 sin 3t + c3 t cos 3t + c4 t sin 3t
Answers
1. y = c1 + c2 t + c3 t2 + c4 cos 4t + c5 sin 4t
2. y = c1 + c2 cos 2t + c3 sin 2t
3. a)
4. b)
5. y = c1 + c2 t + c3 t2 + c4 e−t cos 2t + c5 e−t sin 2t
6. c)
7. b)
8. c)
9. d)
Chapter 5
Laplace Transform
Laplace transform can be used to solve IVPs that we can’t use any previous methods we’ve learned
so far.
5.1 Definition of Laplace Transform

The Laplace transform of f , which we denote it by L{f (t)} or F (s), is defined by the equation
Z ∞
L{f (t)} = e−st f (t) dt,
0
whenever this proper integral converges.
Example 73. Compute L{1}.

Solution By the definition of Laplace transform,
∞ A
e−st t=A e−sA 1
Z Z
1
L{1} = e−st 1 dt = lim e−st dt = lim − t=0
= lim − + = , if s > 0.
0 A→∞ 0 A→∞ s A→∞ s s s
1
So L{1} = , if s > 0.
s
Example 74. Compute L{eat }.

∞ A
e(a−s)t t=A
Z Z
L{eat } = e−st eat dt = lim e(a−s)t dt = lim
0 A→∞ 0 A→∞ a − s t=0
(a−s)A
e 1 1
= lim − = , if s > a.
A→∞ a−s a−s s−a
1
So L{eat } = , if s > a.
s−a
104
CHAPTER 5. LAPLACE TRANSFORM 105
Example 75. Compute L{sinat}.

Solution First, let’s find e−st sin at dt by integration-by-parts. We have
R
e−st sin at a
Z Z
e−st sin at dt = − + e−st cos at dt
s s
e−st cos at a
Z Z
e−st cos at dt = − − e−st sin at dt
s s
Substitute the second equation into the first one,
e−st sin at a e−st cos at a

Z Z
−st
e sin at dt = − + (− − e−st sin at dt)
s s s s
e−st sin at a2
Z
a
=− − 2 e−st cos at − 2 e−st sin at dt
s s s
a2 e−st sin at
Z
a
(1 + 2 ) e−st sin at dt = − − 2 e−st cos at
s s s
s2 + a2 e−st sin at
Z
a
( 2
) e−st sin at dt = − − 2 e−st cos at
s s s
Z 2 −st
s e sin at a
e−st sin at dt = 2 2
(− − 2 e−st cos at) + c
s +a s s
Hence
Z ∞ Z A
L{sin at} = e−st sin at dt = lim e−st sin at dt
0 A→∞ 0
s2 e−st sin at a t=A
= lim 2 2
(− − 2 e−st cos at)t=0
A→∞ s + a s s
s2 s2 a
= 2 (0 + 0) − 2 (− )
s + a2 s + a2 s2
a
= 2 , if s > 0.
s + a2
a
Therefore L{sin at} = if s > 0.
s2 + a2
• Notice that we had to put a restriction on s in order to actually compute the Laplace
transform. All Laplace transform will have restrictions on s. From now on we omit this but
we really shouldn’t forget that it’s there for our convenience.
• Properties of Laplace transform

1. L{cf (t)} = cL{f (t)}.
2. L{f (t) ± g(t)} = L{f (t)} ± L{g(t)}.
3. But L{f (t)g(t)} is not necessary the same as L{f (t)}L{g(t)}.
4. L{0} = 0.
• As above examples, computing Laplace transform of even simple functions is often messy. So
usually we just use a table of Laplace transform when actually computing it without using
the definition directly.
f (t) = L−1 {F (s)} F (s) = L{f (t)}

1
1. 1 , s>0
s
1
2. eat , s>a
s−a
n!
3. tn ; n = positive integer , s>0
sn+1
Γ(p + 1)
4. tp , p > −1 , s>0
sp+1
a
5. sin at , s>0
s2 + a2
s
6. cos at , s>0
s2 + a2
eat − e−at a
7. sinh at := , s > |a|
2 s2 − a2
eat + e−at s
8. cosh at := , s > |a|
2 s2 − a2
b
9. eat sin bt , s>a
(s − a)2 + b2
s−a
10. eat cos bt , s>a
(s − a)2 + b2
n!
11. tn eat , n = positive integer , s>a
(s − a)n+1
e−cs
12. uc (t) , s>0
s
13. uc (t)f (t − c) e−cs F (s)
14. ect f (t) F (s − c)
1 s
15. f (ct) c F ( c ), c>0
Rt
16. 0
f (t − τ )g(τ ) dt F (s)G(s)
17. δ(t − c) e−cs
18. f (n) (t) sn F (s) − sn−1 f (0) − · · · − f (n−1) (0)
n = 1, f 0 (t) sF (s) − f (0)
n = 2, f 00 (t) s2 F (s) − sf (0) − f 0 (0)
19. (−t)n f (t) F (n) (s)
n = 1, tf (t) −F 0 (s)
n = 2, t2 f (t) F 00 (s)
Table 5.1: Elementary LaplaceTransforms

Let’s see how we apply properties of Laplace transform and formulas in Laplace transform table.
Example 76 (Fall 2009 Midterm Exam II).

Evaluate Z ∞
e−st t dt.
0
Z ∞
1
e−st t dt = L{t} = 2 .
0 s

Evaluate the integral Z ∞
e−(s+1)t sin 2t dt.
0
Solution Using the definition of Laplace transform,
Z ∞ Z ∞
2 2
e−(s+1)t sin 2t dt = e−st e−t sin 2t dt = L{e−t sin 2t} = = 2 .
0 0 (s − (−1))2 + 22 s + 2s + 5
Example 78 (Spring 2009, 2011 Midterm Exam II).

True or false:
• L{f (t) − 6g(t)} = L{f (t)} − 6{g(t)}
• L{4f (t)g(t)} = 4L{f (t)}{g(t)}
• L{(t + 5)2 } = L{t + 5}L{t + 5}
Solution True for the first one by properties of Laplace transform and false for the last two
since the Laplace transform of the product does not necessarily equal the product of Laplace
transforms.
Example 79. Compute
1. L{5 + 3e−2t + 4 sin 6t}
Solution
L{5 + 3e−2t + 4 sin 6t} = L{5} + L{3e−2t } + L{4 sin 6t}

= 5L{1} + 3L{e−2t } + 4L{sin 6t}
1 1 6
= 5( ) + 3 + 4( 2 )
s s − (−2) s + 62
5 3 24
= + + 2 .
s s + 2 s + 36
2. L{6e−5t + e3t + 5t3 − 9}
Solution
L{6e−5t + e3t + 5t3 − 9} = 6L{e−5t } + L{e3t } + 5L{t3 } − 9L{1}

1 1 3! 1
= 6( )+ + 5 3+1 − 9( )
s − (−5) s−3 s s
6 1 30 9
= + + 4 − .
s+5 s−3 s s
3. L{−9 sin 4t + 2 cos 10t − e3t cos 6t}
Solution
L{−9 sin 4t + 2 cos 10t − e3t cos 6t} = −9L{sin 4t} + 2L{cos 10t} − L{e3t cos 6t}
4 s s−3
= −9( 2 )+2 2 −
s + 42 s + 102 (s − 3)2 + 62
36 2s s−3
=− 2 + 2 − 2 .
s + 16 s + 100 s − 6s + 45
4. L{5 sinh 2t − 2 sin 2t}
Solution
L{5 sinh 2t − 2 sin 2t} = 5L{sinh 2t} − 2L{sin 2t}

2 2
= 5( 2 )−2 2
s − 22 s + 22
10 4
= 2 − .
s − 4 s2 + 4
5. L{t cosh 3t}
Solution
L{t cosh 3t} = −(L{cosh 3t})0

s s
= −( 2 )0 = −( 2 )0
s − 32 s −9
(s2 − 9)(1) − (s)(2s)
=−
(s2 − 9)2
2
−s − 9 s2 + 9
=− 2 = .
(s − 9)2 (s2 − 9)2
6. L{t2 sin 2t}
Solution
2
L{t2 sin 2t} = (L{sin 2t})00 = ( )00
s2 + 4
−4s 0 12s2 − 16
=( 2 ) = .
(s + 4)2 (s2 + 4)3
Exercises 5.1
1. Which of the following is the Laplace transform of t sin(2t)?
−2
(a)
(s2
+ 4)2
4s
(b) 2
(s + 4)2
s
(c) 2
(s + 4)
3
(d) 2 2
s (s + 4)2
2. For each part below, determine whether the statement is true or false. You must justify your
answers.
(a) L{f (t)g(t)} = L{f (t)}L{g(t)}
(b) L{cos(2t) sin(2t)} =
6 L{cos(2t)}L{sin(2t)}
s2 2t (s − 2)2
(c) Suppose L{f (t)} = , then L{e f (t)} = .
s4 + 1 (s − 2)4 + 1
3. Evaluate the following definite integral
Z ∞
e−st t2 dt
2
2
a)
s3
2e−2s
b)
s3
2 4 4
c) 3 + 2 +
s s s
−2s 2 4 4
d) e ( 3 + 2 + )
s s s
Z ∞
e(1−s)t cos 3t dt.
0
s−1
a)
s2 − 2s + 10
s−1
b) e−s 2
s − 2s + 10
s+1
c) e−s 2
s + 2s + 10
s+1
d) 2
s + 2s + 10
6s
5. Given that L{f (t)} = . Use properties of the Laplace transform to determine L{tf (t)}.
s3 + 64
6. Suppose
s2
L{f (t)} = .
s3 + 27
What is L{7e5t f (t)}?
7(s − 5)2
(a)
(s − 5)3 + 27
7s2
(b)
(s − 5)(s3 + 27)
7s2
(c)
s(s − 5)(s3 + 27)
7(s + 5)2
(d)
(s + 5)3 + 27
7. (a) What is the definition of the Laplace transform L{e3t }?
(b) Use the answer to Part (a) to calculate L{e3t }. (Be sure to explain why this exists only
when s > 3).
8. Find the Laplace transform of the following functions.

(a) e−3t cos 5t
(b) te5t
9. Given that L{f (t)} = F (s), what is L{e5t tf (t)}?

1
a) F (s)
s2 (s − 5)
−1 0
b) F (s)
s−5
1
c) 2 F (s + 5)
s
d) −F 0 (s − 5)
R∞
10. Which of the following functions has 0
e−st t2 dt as its Laplace transform?
a) t2
b) e−t t2
c) t
d) δ(t − 2)t2
R∞
11. 0
e−(s−2)t dt is the Laplace transform of which of the following functions?
a) t2
b) e2t
c) u(t − 2)
d) δ(t − 2)
R∞
12. (a) Evaluate the following definite integral 0
e(4−s)t cos(3t) dt.
2s2
(b) Suppose L{f (t)} = . Use properties of the Laplace transform to determine
s4 + 100
L{e−πt f (t)}.

Z ∞
e−st (t3 + sin 2t) dt
0
3 2
a) +
s4 s2 + 4
6 2
b) 4 + 2
s s +4
6 1
c) 4 + 2
s s + 22
3 1
d) 2 + 2
s s + 22
14. For each part below, determine whether the statement is true or false. State a brief reason
that justifies each answer.
(a) L{e−2t+3 (sin(t) + cos(t))} = e3 L{e−2t sin(t)} + e3 L{e−2t cos(t)}.
(b) Suppose L{f (t)} = e−s s then L{et f (t)} = e−s+1 s − e−s+1 .
1 2s
(c) Suppose L{f (t)} = 2
, then L{tf (t)} = .
1+s (1 + s2 )2
Answers
1. b)
2. a) F, b) T, c) T
3. d)
4. a)
12s3 − 384
5.
(s3 + 64)2
6. a)
R∞ 1
7. a) 0 e−st e3t dt, b)
s−3
s+3 1
8. a) 2 , b)
s + 6s + 34 (s − 5)2
9. d)
10. a)
11. b)
s−4 2(s + π)2
12. a) 2 , b)
s − 8s + 25 (s + π)4 + 100
13. b)
14. a) T, b) T, c) T
5.2 Inverse Laplace Transform

We learn from the first section that Laplace transform sends a function of t to the function of s.
The inverse Laplace transform is simply the inverse function of Laplace transform which does
the opposite. It changes a function of s to the function of s.
L{f (t)} = F (s) ⇔ L−1 {F (s)} = f (t).

Similar to the Laplace transform, the inverse one has the following properties:
1. L−1 {cF (s)} = cL−1 {F (s)}.
2. L−1 {F (s) ± G(s)} = L−1 {F (s)} ± L−1 {G(s)}.
3. L−1 {0} = 0.
The key to compute the inverse Laplace transform is to
• first look at the denominator and try to identify what you’ve got based on that.
• Correct the numerator (if necessary) to get it in the correct form and then use the table to
find the inverse Laplace transform.
Example 80. Find the inverse Laplace transform of F (s).

2 3 4
1. F (s) = − +
s s−8 s+1
Solution
2 3 4 1 1 1
L−1 { − + } = 2L−1 { } − 3L−1 { } + 4L−1 { }
s s−8 s+1 s s−8 s − (−1)
= 2(1) − 3e8t + 4e−t
= 2 − 3e8t + 4e−t
2 11
2. F (s) = − 5
3s − 5 s
Solution
2 11 1 1
L−1 { − 5 } = 2L−1 { 5 } + 11L−1 { 4+1 }
3s − 5 s 3(s − 3 ) s
2 −1 1 11 −1 4!
= L { } + L { 4+1 }
3 s − 53 4! s
2 5 t 11 4
= e3 + t .
3 24
6s 4
3. F (s) = −
s2 + 9 s2 + 25
Solution
6s 4 s 1
L−1 { − } = 6L−1 { 2 } − 4L−1 { 2 }
s2 + 9 s2 + 25 s + 32 s + 52
s 4 5
= 6L−1 { 2 2
} − L−1 { 2 }
s +3 5 s + 52
4
= 6 cos 3t − sin 5t.
5
−4 3
4. F (s) = +
3s2 + 12 s2 − 49
Solution
−4 3 1 1
L−1 { + } = −4L−1 { 2 } + 3L−1 { 2 }
3s2 + 12 s2 − 49 3(s + 22 ) s − 72
−4 −1 2 3 7
= L { 2 2
} + L−1 { 2 }
3(2) s +2 7 s − 72
2 3
= − sin 2t + sinh 7t.
3 7
2s − 5
5. F (s) =
s2 + 7
Solution
2s − 5 2s 5
L−1 { } = L−1 { 2 − }
s2 + 7 s + 7 s2 + 7
s 1
= 2L−1 { √ 2 } − 5L {
−1
√ 2}
s2 + 7 s2 + 7
√
−1 s 5 −1 7
= 2L { √ 2} − √ L { √ 2}
2
s + 7 7 2
s + 7
√ 5 √
= 2 cos 7t − √ sin 7t.
7

2s − 1
1. F (s) =
s2 + 4s + 5
Solution Note that s2 + 4s + 5 = (s2 + 4s + 22 ) − 22 + 5 = (s + 2)2 + 12 .
2s − 1 2(s + 2 − 2) − 1 2(s + 2) − 5
L−1 { } = L−1 { } = L−1 { }
s2 + 4s + 5 2
(s + 2) + 1 2 (s + 2)2 + 12
s+2 1
= 2L−1 { } − 5L−1 { }
(s + 2)2 + 12 (s + 2)2 + 12
= 2e−2t cos t − 5e−2t sin t.
1 − 3s
2. F (s) =
s2 + 8s + 21
√ 2
Solution We have s2 + 8s + 21 = (s2 + 8s + 42 ) − 42 + 21 = (s + 4)2 + 5 .
1 − 3s 1 − 3(s + 4 − 4) −1 1 − 3(s + 4) + 12
L−1 { } = L−1 { √ 2 }=L { √ 2 }
s2 + 8s + 21 (s + 4)2 + 5 (s + 4)2 + 5
13 3(s + 4)
= L−1 { √ 2 − √ 2}
(s + 4)2 + 5 (s + 4)2 + 5
√
13 5 s+4
= √ L−1 { −1
√ 2 } − 3L { √ 2}
5 (s + 4)2 + 5 (s + 4)2 + 5
13 √ √
= √ e−4t sin 5t − 3e−4t cos 5t.
5
2s − 3
3. F (s) =
s2 + 6s + 13
Solution We have s2 + 6s + 13 = (s2 + 6s + 32 ) − 32 + 13 = (s + 3)2 + 22 .
2s − 3 2(s + 3 − 3) − 3 2(s + 3) − 9
L−1 { } = L−1 { } = L−1 { }
s2 + 6s + 13 (s + 3)2 + 22 (s + 3)2 + 22
s+3 9 2
= 2L−1 { } − L−1 { }
(s + 3)2 + 22 2 (s + 3)2 + 22
9
= 2e−3t cos 2t − e−3t sin 2t.
2
Now it’s time to remind ourselves how to use the partial fraction decomposition.
Factor in denominator Term in fraction decomposition

A
ax + b
ax + b
A1 A2 Ak
(ax + b)k + + ··· +
ax + b (ax + b)2 (ax + b)k
Ax + B
ax2 + bx + c
ax2 + bx + c
A1 x + B1 A2 x + B2 Ak x + B k
(ax2 + bx + c)k + + ··· +
ax2 + bx + c (ax2 + bx + c)2 (ax2 + bx + c)k
2s2 − 1
1. F (s) =
s3 − 3s2 − 10s
Solution First we would like to use partial fraction decomposition to rewrite F as
follows
2s2 − 1 A B C
3 2
= + + .
s − 3s − 10s s s−5 s+2
Then we have,
2s2 − 1 A B C
= + +
s3− 3s2 − 10s s s−5 s+2
A(s − 5)(s + 2) Bs(s + 2) Cs(s − 5)
= + +
s(s − 5)(s + 2) s(s − 5)(s + 2) s(s − 5)(s + 2)
2s2 − 1 A(s − 5)(s + 2) + Bs(s + 2) + Cs(s − 5)
∴ 3 = .
s − 3s2 − 10s s(s − 5)(s + 2)
And so
2s2 − 1 = A(s − 5)(s + 2) + Bs(s + 2) + Cs(s − 5) (5.1)
In order to find the coefficients A, B and C, we can do in different ways.

Method 1: Impose sample points into the Equation (5.1):

1
s=0 ⇒ 2(0)2 − 1 = A(0 − 5)(0 + 2) ⇒ A=
10
7
s=5 ⇒ 2(5)2 − 1 = B(5)(5 + 2) ⇒ B=
5
1
s = −2 ⇒ 2(−2)2 − 1 = C(−2)(−2 − 5) ⇒ C=
2
Method 2: Comparison of the coefficients, by (5.1), we have
2s2 − 1 = A(s − 5)(s + 2) + Bs(s + 2) + Cs(s − 5)

= A(s2 − 3s − 10) + B(s2 + 2s) + C(s2 − 5s)
= (As2 − 3As − 10A) + (Bs2 + 2Bs) + (Cs2 − 5Cs)
2s2 + 0s − 1 = (A + B + C)s2 + (−3A + 2B − 5C)s + (−10A)
Hence
A + B + C = 2, −3A + 2B − 5C = 0, −10A = −1
1 7 1
which also give us A = , B = and C = . So
10 5 2
L−1 {F (s)} = L−1 {2s2 − 1}
1 7 1
= L−1 { 10 + 5 + 2 }
s s−5 s+2
1 7 1
= L−1 { 10 } + L−1 { 5 } + L−1 { 2 }
s s−5 s+2
1 7 5t 1 −2t 1 7 5t 1 −2t
= (1) + e + e = + e + e .
10 4 2 10 4 2
s2 + 1
2. F (s) =
s(s + 1)2
Solution Get a new form for F by using partial fraction decomposition,
s2 + 1 A B C
= + +
s(s + 1)2 s s + 1 (s + 1)2
A(s + 1)2 Bs(s + 1) Cs
= + +
s(s + 1)2 s(s + 1)2 s(s + 1)2
A(s + 1)2 + Bs(s + 1) + Cs
=
s(s + 1)2
And so
s2 + 1 = A(s + 1)2 + Bs(s + 1) + Cs (5.2)
Method 1: Impose sample points into the Equation (5.1):
s=0 ⇒ (0)2 + 1 = A(0 + 1)2 ⇒ A=1

2
s = −1 ⇒ (−1) + 1 = C(−1) ⇒ C = −2
2 2
s=1 ⇒ 1 + 1 = A(1 + 1) + B(1)(1 + 1) + C(1) ⇒ B = 0.
Method 2: Now let’s find the coefficients A, B and C, by comparison of the coefficients
of (5.2),
s2 + 1 = A(s + 1)2 + Bs(s + 1) + Cs

= A(s2 + 2s + 1) + B(s2 + s) + Cs
s2 + 0s + 1 = (A + B)s2 + (2A + B + C)s + A
Hence
A + B = 1, 2A + B + C = 0, A=1
which give us A = 1, B = 0 and C = −2.
Therefore, we have
s2 + 1
L−1 {F (s)} = L−1 { }
s(s + 1)2
1 0 −2
= L−1 { + + }
s s + 1 (s + 1)2
1 1
= L−1 { } − 2L−1 { }
s (s + 1)2
= 1 − 2te−t .
5s2 − 20s + 34
3. F (s) =
(s − 6)(s2 + 11)
Solution Using partial fraction decomposition to rewrite F as follows
5s2 − 20s + 34 A Bs + C
= + 2
(s − 6)(s2 + 11) s−6 s + 11
2
A(s + 11) + (Bs + C)(s − 6)
=
(s − 6)(s2 + 11)
(As + 11A) + (Bs2 − 6Bs + Cs − 6C)
2
=
(s − 6)(s2 + 11)
2
(A + B)s + (−6B + C)s + (11A − 6C)
= .
(s − 6)(s2 + 11)
Then we have
5s2 − 20s + 34 = (A + B)s2 + (−6B + C)s + (11A − 6C). (5.3)
By the comparison of the coefficients of (5.3),
A + B = 5, −6B + C = −20, 11A − 6C = 34
which gives us A = 2, B = 3 and C = −2.

Therefore, we have
5s2 − 20s + 34
L−1 {F (s)} = L−1 { }
(s − 6)(s2 + 11)
2 3s − 2
= L−1 { + }
s − 6 s2 + 11
2 3s 2
= L−1 { + − }
s − 6 s2 + 11 s2 + 11
√
1 s 2 −1 11
= 2L−1 { } + 3L−1 { √ 2 } − √ L { √ 2}
s−6 2
s + 11 11 2
s + 11
√ 2 √
= 2e6t + 3 cos 11t − √ sin 11t.
11
4s4 + s3 − 4s2 − 14s − 5

4. F (s) =
s3 (s2 + 4s + 5)
Solution Rewrite F by using partial fraction decomposition,
4s4 + s3 − 4s2 − 14s − 5 A B C Ds + E

= + 2+ 3+ 2
s3 (s2 + 4s + 5) s s s s + 4s + 5
As2 (s2 + 4s + 5) + Bs(s2 + 4s + 5) + C(s2 + 4s + 5) + (Ds + E)s3
=
s3 (s2 + 4s + 5)
(A + D)s + (4A + B + E)s3 + (5A + 4B + C)s2 + (5B + 4C)s + 5C
4
= .
s3 (s2 + 4s + 5)
Then we have
4s4 +s3 −4s2 −14s−5 = (A+D)s4 +(4A+B +E)s3 +(5A+4B +C)s2 +(5B +4C)s+5C.
(5.4)
Therefore by the comparison of the coefficients of (5.4),
A + D = 4, 4A + B + E = 1, 5A + 4B + C = −4, 5B + 4C = −14, 5C = −5.
which give us A = 1, B = −2, C = −1, D = 3 and E = −1.

Therefore, we have
4s4 + s3 − 4s2 − 14s − 5

L−1 {F (s)} = L−1 { }
s3 (s2 + 4s + 5)
1 2 1 3s − 1
= L−1 { − 2 − 3 + 2 }
s s s s + 4s + 5
1 1 1 3s − 1
= L−1 { } − 2L−1 { 2 } − L−1 { 3 } + L−1 { }
s s s (s + 2)2 + 12
1 2! 3(s + 2 − 2) − 1
= 1 − 2t − L−1 { 3 } + L−1 { }
2 s (s + 2)2 + 12
1 3(s + 2) − 7
= 1 − 2t − t2 + L−1 { }
2 (s + 2)2 + 12
t2 s+2 1
= 1 − 2t − + 3L−1 { } − 7L−1 { }
2 (s + 2)2 + 12 (s + 2)2 + 12
t2
= 1 − 2t − + 3e−2t cos t − 7e−2t sin t.
2
Exercises 5.2
1. Find the inverse Laplace transform of:
−2s + 1
s2 − 2s + 5
(a) −2et cos(2t)
(b) −2et cos(2t) + 3et sin(2t)
(c) −2et cos(2t) + et sin(2t)
(d) −2et cos(2t) − 21 et sin(2t)
2. What is the inverse Laplace transform of

2s + 7
s2 + 8s + 65
(a) 2e−8t cos t − 11e−8t sin t
(b) 2e−4t cos 7t − 71 e−4t sin 7t
(c) 2e−7t cos 4t − 41 e−7t sin 4t
(d) 3 cos 7t − 4 sin 7t
8
3. The inverse Laplace transform of is
s3 − 4s
(a) −2 + e2t + e−2t
(b) −2t + e2t + e−2t
(c) 2t + cos 2t
1
(d) −2 + 2 sin 2t
4. Find the inverse Laplace transform of each function:
s+4
(a)
s2 − 4s + 13
2s2 − 10s − 18
(b)
s3 − s2 − 6s
2s + 2
(c)
(s − 3)2
5. True or false. L−1 {F (s) + 7} = L−1 {F (s)} + 7.
6. Find the inverse Laplace transform of the function given below
s2 − 6s + 10
F (s) = .
(s − 1)(s − 2)(s − 3)
7. Find the function y(t) whose Laplace transform is the expression
7s2 − 8s + 12
F (s) = .
(s − 2)(s2 + 4)
8. Which of the following is the Laplace transform of t sin(2t)?

−2
(a)
(s2 + 4)2
4s
(b) 2
(s + 4)2
s
(c) 2
s +4
3
(d) 2 2
s (s + 4)
9. Find the inverse Laplace transform of
3s + 4
(a)
s2 + 4s + 20
s2 − 2s − 1
(b)
s(s + 1)2
3s2 − 2s + 8
(c)
s3 + 4s
√ 4s + 10
(d) 7 + 2
s − 4s + 20
2s2 − 1
(e) 3
s − 3s2 − 10s
s2 + 11s − 31
(f)
(s + 5)(s2 + 36)
Answers
1. d)
2. b)
3. a)
4. a) e2t cos 3t + 2e2t sin 3t, b) 3 + e−2t − 2e3t , c) 2e3t + 8te3t
5. F
5 t 1
6. e − 2e2t + e3t
2 2
7. 3e2t + 4 cos 2t
8. b)
9. a) 3e−2t cos 4t − 21 e−2t sin 4t, b) −1 + 2e−t − 2te−t , c) 2 + cos 2t − sin 2t,
d) δ(t) + 4e2t cos 4t + 29 e2t sin 4t, e) 10
1
+ 75 e5t + 12 e−2t , f) −e−5t + 2 cos 6t + 1
6 sin 6t
5.3 Solving Initial Value Problems with Laplace Transform

We can derive the Laplace transform formula for nth derivative of y as follows:
L{y (n) (t)} = sn L{y(t)} − sn−1 y(0) − . . . − y (n−1) (0).
Most of the time we use this formula in special cases(when n = 1 and n = 2)
L{y 0 (t)} = sL{y(t)} − y(0),

L{y 00 (t)} = s2 L{y(t)} − sy(0) − y 0 (0).
Notice two functions evaluations, y(0) and y 0 (0), that appear in these formula are often what we’ve
been using for initial conditions in our initial value problem. So this mean if we want to apply
these formulas, we need initial conditions at t = 0 only.
• When we solve the initial value problem in the past, we need to first find a general solution,
differentiate it, plug in the initial conditions and then solve for the constants to get the
particular solution.
• But by using Laplace transform, the initial conditions are applied during the first step and
at the end we get the particular solution instead of a general solution.
The process of solving initial value problem with Laplace transform,
1. Take Laplace transform to both sides of the equation.

2. Substitute two initial conditions, y(0) and y 0 (0).
3. Find L{y(t)} = Y (s).
4. Take inverse Laplace transform both sides of the equation to get y(t) = L−1 {Y (s)}.
Example 83. Use the Laplace transform to solve the IVP
y 00 + y 0 − 6y = 1, y(0) = 0, y 0 (0) = 2.
Solution First, we take the Laplace transform to both sides of the equation,
L{y 00 + y 0 − 6y} = L{1}

L{y } + L{y 0 } − 6L{y} = L{1}
00
1
s2 L{y} − sy(0) − y 0 (0) + (sL{y} − y(0)) − 6L{y} =

s
2
1
s L{y} − s(0) − 2 + (sL{y} − 0) − 6L{y} =
s
2 1
s L{y} + sL{y} − 6L{y} − 2 =
s
2 1
(s + s − 6)L{y} = + 2
s
2s + 1
(s + 3)(s − 2)L{y} =
s
2s + 1
L{y} =
s(s + 3)(s − 2)
Finally, we take an inverse Laplace transform to get the particular solution,

2s + 1
y = L−1 { }
s(s + 3)(s − 2)
−1 − 31 1
= L−1 { 6 + + 2 }
s s+3 s−2
1 1 −3t 1 2t
∴y=− − e + e .
6 3 2
Example 84 (Spring 2004 Midterm Exam II).

Solve the IVP
5
y 00 + 2y 0 − 3y = 3et , y 0 (0) = − .
y(0) = 0,
4
Solution Taking the Laplace transform to both sides of the equation,
L{y 00 + 2y 0 − 3y} = L{3et }

L{y 00 } + 2L{y 0 } − 3L{y} = 3L{et }
3
s2 L{y} − sy(0) − y 0 (0) + 2 (sL{y} − y(0)) − 3L{y} =

s−1

2 5 3
s L{y} − s(0) − (− ) + 2 (sL{y} − 0) − 3L{y} =
4 s−1
5 3
s2 L{y} + 2sL{y} − 3L{y} + =
4 s−1
2 3 5
(s + 2s − 3)L{y} = −
s−1 4
5
3 (s − 1)
(s + 3)(s − 1)L{y} = − 4
s−1 s−1
− 45 s + 174
L{y} =
(s − 1)2 (s + 3)
By partial fraction decomposition,
− 54 s + 17
4 A B C
2
= + 2
+
(s − 1) (s + 3) s − 1 (s − 1) s+3
A(s − 1)(s + 3) B(s + 3) C(s − 1)2
= + +
(s − 1)2 (s + 3) (s − 1)2 (s + 3) (s − 1)2 (s + 3)
A(s + 2s − 3) + B(s + 3) + C(s2 − 2s + 1)
2
=
(s − 1)2 (s + 3)
(A + C)s2 + (2A + B − 2C)s + (−3A + 3B + C)
=
(s − 1)2 (s + 3)
By the comparison of the coefficients of the numerator, we have

5 17
A + C = 0, 2A + B − 2C = − , −3A + 3B + C = .
4 4
Then we can find that A = − 12 , B = 3
4 and C = 12 .
Hence
− 12 3
4
1
2
L{y} = + + .
s − 1 (s − 1)2 s+3
Taking an inverse Laplace transform to get the particular solution,
− 12 3 1
y = L−1 { } + L−1 { 4
} + L −1
{ 2
}
s−1 (s − 1)2 s+3
1 1 3 1 1 1
= − L−1 { } + L−1 { } + L−1 { }
2 s−1 4 (s − 1)2 2 s+3
1 3 1
∴ y = − et + tet + e−3t .
2 4 2
Exercises 5.3
1. Find L{(eπt sin 3t)000 }. Hint:
(eπt sin 3t)0 = πeπt sin 3t + 3eπt cos 3t
(eπt sin 3t)00 = (π 2 + 9)eπt sin 3t + 6πeπt cos 3t
2. Suppose that the Laplace transform of y is Y . If y(0) = 2 and y 0 (0) = 3, then find the
Laplace transform of y 00 .
3. Suppose y(t) is the solution of the second order linear initial value problem
y 00 + 4y = t, y(0) = 1, y 0 (0) = 0.
What is the Laplace transform of y(t)?
1
(a) Y (s) =
s2 (s2
+ 4)
1 − s2
(b) Y (s) = 2 2
s (s + 4)
s2 + 1
(c) Y (s) = 2 2
s (s + 4)
s3 + 1
(d) Y (s) = 2 2
s (s + 4)
4. Suppose y(t) is the solution of the first order linear initial value problem
y 0 + 2y = t3 e−4t , y(0) = −3.
What is the Laplace transform of y(t)?
6 3
(a) Y (s) = −
(s + 2)(s + 4)4 s+4
6 3
(b) Y (s) = +
(s + 2)(s + 4)4 s+4
3 3
(c) Y (s) = +
(s + 2)(s − 4)4 s+2
3 3
(d) Y (s) = −
(s + 2)(s − 4)4 s+2
Answers
3s3
1. − 3s − 6π
(s − π)2 + 9
2. s2 Y − 2s + 3
3. d)
4. a)
5.4 Step Functions

The unit step function or Heaviside function, denoted by uc (t) or u(t − c), is denoted by
(
0 if t < c
uc (t) = u(t − c) = , where c ≥ 0.
1 if t ≥ c.
Figure 5.1: Graph of y = uc (t).
Example 85. By the definition,

(
0 if t < π
uπ (t) = .
1 if t ≥ π.
And so uπ (1) = 0, uπ (3.14) = 0 but uπ (π) = 1, uπ (2) = 1 and uπ (100) = 1.

Suppose f (t) = 2 + u3 (t)(t − 1) + u5 (t)t2 . What is f (4)?
Solution
f (4) = 2 + u3 (4) · (4 − 1) + u5 (4) · 42 = 2 + 1 · 3 + 0 · 16 = 5.

If f (t) = u(t − 2) − 2u(t − 3) + 3u(t − 4) − 4u(t − 5). What is f (π)?
Solution
f (π) = u2 (π) − 2u3 (π) + 3u4 (π) − 4u5 (π) = 1 − 2 · 1 + 3 · 0 − 4 · 0 = −1.

Example 88. Write the following function in terms of unit step function.

3,
 t<2
f (t) = −1, 2≤t<7.

2, t ≥ 7.

Solution
Then f (t) = 3 + (−1 − 3)u2 (t) + (2 − (−1))u7 (t) = 3 − 4u2 (t) + 3u7 (t).
Example 89. Rewrite the following piecewise continuous function f (t) in terms of unit step
functions.

2
t ,
 t<1
f (t) = 1 − t, 1≤t<3.

 2t
2e , t ≥ 3.
Solution Then f (t) = t2 + (1 − t − t2 )u1 (t) + (2e2t − (1 − t))u3 (t).
Example 90. Sketch the graph of f (t) = u1 (t) + 2u3 (t) − 6u4 (t)
Solution It is easier to graph if we can rewrite the function as a piecewise continuous one.
In order to do that, we consider its value on each subintervals.
On t<1 ⇒ Pick t = 0 ⇒ f (0) = u1 (0) + 2u3 (0) − 6u4 (0) = 0

On 1 ≤ t < 3 ⇒ Pick t = 2 ⇒ f (2) = u1 (2) + 2u3 (2) − 6u4 (2) = 1
On 3 ≤ t < 4 ⇒ Pick t = 3 ⇒ f (3) = u1 (3) + 2u3 (3) − 6u4 (2) = 1 + 2 = 3
On t≥4 ⇒ Pick t = 5 ⇒ f (5) = u1 (5) + 2u3 (5) − 6u4 (5) = 1 + 2 − 6 = −3
And so we have 
0,


t<1
1, 1≤t<3
f (t) = .
3,

 3≤t<4
−3, t ≥ 4.

Now let’s consider the Laplace transform formulas involving unit step functions. By the definition,
Z ∞ Z c Z ∞
L{uc (t)} = e−st uc (t) dt = e−st (0) dt + e−st (1) dt
0 0 c
Z ∞ Z A
−st −st
= e dt = lim e dt
c A→∞ c
e−st t=A
−sA
e−sc

e
= lim − = lim − − (− )
A→∞ s t=c A→∞ s s
e−cs
= , if s > 0.
s
And also by changing the variable technique, let ξ = t − c(or t = ξ + c), we have
Z ∞ Z t=∞
L{uc (t)f (t − c)} = e−st uc (t)f (t − c) dt = e−st f (t − c) dt
0 t=c
Z ξ=∞ Z ∞
= e−s(ξ+c) f (ξ) dξ = e−cs e−sξ f (ξ) dξ
ξ=0 0
= e−cs L{f (t)} = e−cs F (s).
So
L−1 {e−cs F (s)} = uc (t)f (t − c) = uc (t)f (t)t7→t−c = uc (t)L−1 {F (s)}t7→t−c .
Laplace transform formulas involving unit step functions:

e−cs
−cs
−1 e
• L {uc (t)} = ⇔ L = uc (t)
s s
• L {uc (t)f (t − c)} = e−cs L{f (t)} ⇔ L−1 {e−cs F (s)} = uc (t)L−1 {F (s)}t7→t−c .

Find the Laplace transform of u2 (t)(t2 + 1).
Solution Here c = 2, so we let f (t − 2) = t2 + 1. Substitute t by t + 2, we have
f (t) = f ((t + 2) − 2) = (t + 2)2 + 1 = (t2 + 4t + 4) + 1 = t2 + 4t + 5.
Therefore, by using L {u2 (t)f (t − 2)} = e−2s L{f (t)},

2 4 5
L{u2 (t)(t2 + 1)} = e−2s L{t2 + 4s + 5} = e−2s ( + 2 + ).
s3 s s

Find the Laplace transform of u4 (t)(t − 2)2 .
Solution Now c = 4, let f (t − 4) = (t − 2)2 . Then we substitute t by t + 4,
f (t) = f ((t + 4) − 4) = (t + 4 − 2)2 = (t + 2)2 = t2 + 4t + 4.
So, by using L {u4 (t)f (t − 4)} = e−4s L{f (t)},

2 4 4
L{u4 (t)(t − 2)2 } = e−4s L{t2 + 4t + 4} = e−2s ( + 2 + ).
s3 s s
Example 93 (Summer 2011 Midterm Exam II).

Rewrite the following piecewise continuous function f (t) in terms of the unit step function.
Then find its Laplace transform.
(
2t2 + t, 0 ≤ t < 3
f (t) =
e4t , 3≤t
Solution We can rewrite f as 2t2 + t + (e4t − 2t2 − t)u3 (t). And so
L{f (t)} = L 2t2 + t + (e4t − 2t2 − t)u3 (t)

= 2L{t2 } + L{t} + L (e4t − 2t2 − t)u3 (t)

2! 1 n o
= 2 · 3 + + e−3s L (e4(t+3) − 2(t + 3)2 − (t + 3)
s s
4 1
= 3 + + e−3s L e4t · e12 − 2(t2 + 6t + 9) − t − 3

s s
4 1
= 3 + + e−3s L e4t · e12 − 2t2 − 12t − 18 − t − 3

s s
4 1
= 3 + + e−3s L e4t · e12 − 2t2 − 13t − 21

s s
4 1
= 3 + + e−3s e12 L{e4t } − 2L{t2 } − 13L{t} − 21L{1}

s s
4 1 −3s 12 1 2 1 1
= 3 + +e e · − 2 · 3 − 13 · 2 − 21 ·
s s s−4 s s s
12
4 1 e 4 13 21
∴ L{f (t)} = 3 + + e−3s − − 2 − .
s s s − 4 s3 s s

e−3s
Find the inverse Laplace transform of .
s3 + s2
Solution Using partial fraction decomposition,
1 A B C
= + 2+
s2 (s + 1) s s s+1
As(s + 1) + B(s + 1) + Cs2
=
s2 (s + 1)
2 2
0s + 0s + 1 (A + C)s + (A + B)s + B
2
=
s (s + 1) s2 (s + 1)
Hence
A + C = 0, A + B = 0, B = 1,
which give us A = −1, B = 1 and C = 1.
Therefore
1 −1 1 1
3 2
= + 2+ .
s +s s s s+1
Using L−1 e−3s F (s) = u3 (t)L−1 {F (s)}t7→t−3 ,

e−3s

1
L−1 = u3 (t)L−1
s + s2
3 s3 + s2 t7→t−3

−1 −1 1 1
= u3 (t)L + 2+
s s s + 1 t7→t−3
−t

= u3 (t) −1 + t + e t7→t−3

= u3 (t) −1 + (t − 3) + e−(t−3)
e−3s
∴ L−1 { 3 } = u3 (t) t − 4 + e −t+3)
.
s + s2
Example 95.
2s − 8
Find the inverse Laplace transform of e−4s 2 .
s + 2s + 26
Solution We apply L−1 e−4s F (s) = u4 (t)L−1 {F (s)}t7→t−4 to get

−1 −4s 2s − 8 −1 2s − 8
L e = u4 (t)L
s2 + 2s + 26 s2 + 2s + 26 t7→t−4

−1 2(s + 1 − 1) − 8
= u4 (t)L
(s + 1)2 + 52 t7→t−4

2(s + 1) − 10
= u4 (t)L−1
(s + 1)2 + 52 t7→t−4

−1 s+1 10 −1 5
= u4 (t) 2L − L
(s + 1)2 + 52 5 (s + 1)2 + 52 t7→t−4
= u4 (t) 2e−t cos 5t − 2e−t sin 5t t7→t−4

2s − 8
∴ L−1 {e−4s } = u4 (t) 2e−(t−4) cos 5(t − 4) − 2e−(t−4) sin 5(t − 4) .
s2 + 2s + 26

−s + 3
Find the inverse Laplace transform of e−9s 2 .
s + 6s + 25
Solution By L−1 e−9s F (s) = u9 (t)L−1 {F (s)}t7→t−9 ,

−s + 3 −s + 3
L−1 e−9s = u9 (t)L−1
s2 + 6s + 25 s2 + 6s + 25 t7→t−9

−1 −(s + 3 − 3) + 3
= u9 (t)L
(s + 3)2 + 42 t7→t−9

−(s + 3) + 6
= u9 (t)L−1
(s + 3)2 + 42 t7→t−9

s+3 6 −1 4
= u9 (t) −L−1 + L
(s + 3)2 + 42 4 (s + 3)2 + 42 t7→t−9

3
= u9 (t) −e−3t cos 4t + e−3t sin 4t
2 t7→t−9

−1 −4s −s + 3 −3(t−9) 3 −3(t−9)
∴ L {e } = u9 (t) −e cos 4(t − 9) + e sin 4(t − 9)
s2 + 6s + 25 2
Next we’ll use Laplace transform to solve initial value problems containing unit step functions.
This is the benefit of Laplace transform since we can not solve these kind of initial value problems
with the method of undetermined coefficients.

Solve the following IVP
y 00 − 3y 0 − 4y = u2 (t) − u6 (t), y(0) = 0, y 0 (0) = 0.
Solution We use the same strategy as before by firstly taking Laplace transform to both sides
of the equation.
L{y 00 − 3y 0 − 4y} = L{u2 (t) − u6 (t)}

L{y 00 } − 3L{y 0 } − 4L{y} = L{u2 (t)} − L{u6 (t)}
e−2s e−6s
s2 L{y} − sy(0) − y 0 (0) − 3 (sL{y} − y(0)) − 4L{y} =

−
s s
−2s
e e−6s
s2 L{y} − s(0) − 0 − 3 (sL{y} − 0) − 4L{y} =

−
s s
2 e−2s e−6s
s L{y} − 3sL{y} − 4L{y} = −
s s
−2s
e e−6s
(s2 − 3s − 4)L{y} = −
s s
e−2s e−6s
(s − 4)(s + 1)L{y} = −
s s
e−2s e−6s
L{y} = −
s(s − 4)(s + 1) s(s − 4)(s + 1)
By taking an inverse Laplace transform, we have
e−2s e−6s

−1 −1
y=L −L
s(s − 4)(s + 1) s(s − 4)(s + 1)

−1 1 −1 1
= u2 (t)L − u6 (t)L
s(s − 4)(s + 1) t7→t−2 s(s − 4)(s + 1) t7→t−6
1 1 1 1 1 1
−4 −4
= u2 (t)L−1 + 20 + 5 − u6 (t)L−1 + 20 + 5
s s − 4 s + 1 t7→t−2 s s − 4 s + 1 t7→t−6

1 1 4t 1 −t 1 1 4t 1 −t
= u2 (t) − + e + e − u6 (t) − + e + e
4 20 5 t7→t−2 4 20 5 t7→t−6

1 1 4(t−2) 1 −(t−2) 1 1 4(t−6) 1 −(t−6)
∴ y = u2 (t) − + e + e − u6 (t) − + e + e .
4 20 5 4 20 5
Example 98.
Solve the IVP
(
00 0 0 1, 0 < t < 10
y + 3y + 2y = f (t), y(0) = 0, y (0) = 0, f (t) =
0, t ≥ 10.
Solution We can rewrite f as f (t) = 1 − u10 (t). Then take the Laplace transform to both
sides of the equation,
L{y 00 + 3y 0 + 2y} = L{1 − u10 (t)}

L{y } + 3L{y 0 } + 2L{y} = L{1} − L{u10 (t)}
00
1 e−10s
s2 L{y} − sy(0) − y 0 (0) + 3 (sL{y} − y(0)) + 2L{y} = −

s s
2
1 e−10s
s L{y} − s(0) − 0 + 3 (sL{y} − 0) + 2L{y} = −
s s
1 e−10s
(s2 + 3s + 2)L{y} = −
s s
1 e−10s
(s + 2)(s + 1)L{y} = −
s s
1 e−10s
L{y} = − .
s(s + 1)(s + 2) s(s + 1)(s + 2)
Taking an inverse Laplace transform to get the particular solution
e−10s

1
y = L−1 − L−1
s(s + 1)(s + 2) s(s + 1)(s + 2)

1 1
= L−1 − u10 (t)L−1
s(s + 1)(s + 2) s(s + 1)(s + 2) t7→t−10
1 1 1 1
−1 2 −1 2 −1 2 −1 2
=L + + − u10 (t)L + +
s s+1 s+2 s s + 1 s + 2 t7→t−10

1 1 1 1
= − e−t + e−2t − u10 (t) − e−t + e−2t
2 2 2 2 t7→t−10

1 −t 1 −2t 1 −(t−10) 1 −2(t−10)
= −e + e − u10 (t) −e + e .
2 2 2 2
Exercises 5.4
1. Rewrite the following function f (t) in terms of step functions, and find its Laplace transform.
(
3t2 , 0 ≤ t < 2π
f (t) = 2
3t + sin t, t ≥ 2π.
2. (a) Sketch the function

f (t) = t − u1 (t)(t − 1) + u3 (t)(2 − t)
(b) Write the given function in terms of unit step functions and find its Laplace transform.

2
t ,
 0≤t<2
f (t) = 6 − t, 2≤t<6

0, 6≤t

3. Find the solution to the following initial value problem
y 00 + 2y 0 − 8y = e3t u3 (t), y(0) = 0, y 0 (0) = 1.
4. Find the Laplace transform of u π2 (t) cos(2t). The following identity may be useful:
cos(α ± β) = cos α cos β ∓ sin α sin β

s sπ
(a) − e− 2
s2 + 4
2 sπ
(b) − 2 e− 2
s +4
1 sπ
(c) 2 e− 2
s +4
s sπ
(d) − 2 e− 2
s + 4s + 4
5. Suppose f (t) = 2 − u2 (t) + t(u3 (t) − u6 (t)), what is f (5)?
(a) 2
(b) 4
(c) 6
(d) 7
6. Rewrite the following function in terms of unit step functions and find it’s Laplace transform.
(
t+1 0≤t<3
f (t) = 2 t
t +e , 3≤t
7. Let g(t) = tu1 (t) − t2 u3 (t) + tu5 (t). What is g(4)?

(a) 4
(b) −12
(c) −16
e−s 2e−3s e−5s
(d) 2
− 3
+ 2
s s s
8. Which of the following functions corresponds to this graph?
(a) tu1 (t) − 1

(b) t + u1 (t)
(c) (1 − t)u2 (t) + t
(d) t + (1 − t)u1 (t)
9. True or false:
1
(a) L{uπ (t) sin t} = e−πs
s2 + 1
(b) If f (t) = 2 − u2 (t) + 3u4 (t), then f (3) = 1.
10. Find the inverse Laplace transform of each function given below:
e−5s
(a)
s2 (s
− 2)
1 + 4s
(b) e−3s 2
s + 10s + 41
5s + 4
(c) F (s) = e−101s 2
s − 2s + 10
11. Find the Laplace transform of f (t) = u3 (t)t2 e−t .
2
(a) e−3s
(s + 1)3
2
(b) e−3s
s(s + 1)3
2 3
(c) e−3s+3 ( + )
(s + 1)3 s+1
2 6 9
(d) e−3s−3 ( + + )
(s + 1)3 (s + 1)2 s+1
12. Rewrite the following piecewise continuous function f (t) in terms of the unit step function.
Then find its Laplace transform.
(
t2 + 3t, 0 ≤ t < 7
f (t) = 2
t + 3t + te−t + 5, t ≥ 7

y 00 + 2y 0 + 5y = 6uπ (t), y(0) = 2, y 0 (0) = 0.
Find its Laplace transform Y (s) = L{y(t)}.
6e−πs 2s + 2
(a) Y (s) = 2
− 2
s(s + 2s + 5) s + 2s + 5
6e−πs + 2s
(b) Y (s) =
s(s2 + 2s + 5)
6e−πs
(c) Y (s) =
s(s2 + 2s + 5)
6e−πs 2s + 4
(d) Y (s) = +
s(s2 + 2s + 5) s2 + 2s + 5
14. Find L{uπ (t) sin( 2t )}
Hint: sin(α ± β) = sin(α) cos(β) ± cos(α) sin(β).
15. Find the Laplace transform L{u1 (t)tet−1 }.
s
(a) e−s
(s − 1)2
1
(b) e−s
s(s − 1)2
1
(c) e−s−1
s(s − 1)2
s
(d) e−s−1
(s − 1)2
Answers
6 1
1. f (t) = 3t2 + u2π (t) sin t, F (s) = + e−2πs 2
s3 s +1
2 2 5 1
2. a) -, b) f (t) = t2 + (6 − t − t2 )u2 (t) + (t − 6)u6 (t), 3 + e−2s (− 3 − 2 ) + e−6s ( 2 )
s s s s
1 1 1 1 1
3. − e−4t + e2t + ( e3(t−3) − e2(t−3) + e−4(t−3) )e9 u3 (t)
6 6 7 6 42
4. a)
5. c)
1 1 2 5 5 e3
6. f (t) = (t + 1) + u3 (t)(t2 − t − 1 + et ), F (s) = 2 + + e−3s ( 3 + 2 + + )
s s s s s s−1
7. b)
8. d)
9. a) F, b) T
1 1 1 19
10. a) u5 (t)(− − (t − 5) + e2(t−5) ),b) u3 (t)(4e−5(t−3) cos 4(t − 3) − e−5(t−3) sin 4(t − 3)), c)
4 2 4 4
u101 (t)(5et−101 cos(3t − 303) + 3et−101 sin(3t − 303))
11. d)
e−7 7e−7

2 3 5
12. 3 + 2 + e−7s + +
s s (s + 1)2 s+1 s
12. d)
4s
13. e−πs 2
4s + 1
14. a)
5.5 Impulse Functions

The idealized impulse function is called Dirac Delta function at t = c if it has two following
properties:
1. δ(t − c) = 0, if t 6= c,
R∞
2. −∞ δ(t − c) dt = 1.
Note here that
• In elementary calculus, there is no such function to satisfy above two properties.
• The function δ is an example of what are known as generalized function or distribution.
So if this is an ideal function, how do mathematicians work with it? The answer is that mathe-
maticians use Dirac Delta function in terms of the limit of tangible function dτ ,
δ(t − c) = lim dτ (t − c). (5.5)

τ →0+
Here is how dτ is defined,

(
1
2τ if − τ < t − c < τ,
dτ (t − c) =
0 if t − c ≤ −τ or t − c ≥ τ,
where τ is a positive constant. For simplicity, we consider its graph when c = 0.
Figure 5.2: Graph of y = dτ (t).
Figure 5.3: Graphs of y = dτ (t) as τ → 0+ .

By using Equation (5.5), we have the following results,
Laplace transform formulas involving Dirac Delta functions:

• L {δ(t)} = 1 ⇔ L−1 {1} = δ(t).
• L {δ(t − c)} = e−cs ⇔ L−1 {e−cs } = δ(t − c).
• L {δ(t − c)f (t)} = f (c)e−cs ⇔ L−1 {f (c)e−cs } = δ(t − c)f (t).

Solve the IVP
y 0 + 3y = tu(t − 2) + δ(t − 3), y(0) = 1.
Solution Again firstly taking Laplace transform to both sides of the equation.
L{y 0 + 3y} = L{tu(t − 2) + δ(t − 3)}

L{y 0 } + 3L{y} = L{tu2 (t)} + L{δ(t − 3)}
(sL{y} − y(0)) + 3L{y} = e−2s L {t + 2} + e−3s

−2s 1 2
(sL{y} − 1) + 3L{y} = e + + e−3s
s2 s
2s + 1
(s + 3)L{y} − 1 = e−2s · + e−3s
s2
2s + 1 e−3s 1
L{y} = e−2s 2 + + .
s (s + 3) s + 3 s + 3
By taking an inverse Laplace transform, we have

−1 −2s 2s + 1 −1 −3s 1 −1 1
y=L e +L e +L
s2 (s + 3) s+3 s+3

−1 2s + 1 −1 1 −1 1
= u2 (t)L + u3 (t)L +L
s2 (s + 3) t7→t−2 s + 3 t7→t−3 s+3
5 1 5
−9
= u2 (t)L−1 9 + 32 + + u3 (t) e−3t t7→t−3 + e−3t

s s s + 3 t7→t−2

5 1 5
= u2 (t) + t − e−3t + u3 (t)e−3(t−3) + e−3t
9 3 9 t7→t−2

5 t − 2 5 −3(t−2)
∴ y = u2 (t) + − e + u3 (t)e−3(t−3) + e−3t .
9 3 9
Example 100.
Solve the IVP
y 00 + y = δ(t − 2π) cos t, y(0) = 0, y 0 (0) = 1..
Solution Apply the Laplace transform to both sides of the equation.
L{y 00 + y} = L{δ(t − 2π) cos t}

L{y 00 } + L{y} = e−2πs cos 2π
s2 L{y} − sy(0) − y 0 (0) + L{y} = e−2πs

s2 L{y} − s(0) − 1 + L{y} = e−2πs

(s2 + 1)L{y} − 1 = e−2πs

e−2πs 1
L{y} = +
s2 + 1 s2 + 1
Hence, by taking the inverse Laplace transform to both sides of the equation,

−1 −2πs 1 −1 1
y=L e · 2 +L
s +1 s2 + 1

−1 1 −1 1
= u2π (t)L +L
s2 + 12 t7→t−2π s2 + 12
= u2π (t) (sin t)t7→t−2π + sin t
= u2π (t) sin(t − 2π) + sin t
∴ y = u2π (t) sin(t) + sin t
Exercises 5.5
1. Solve the initial value problem,
y 00 − 4y = δ(t − 2π), y(0) = 0, y 0 (0) = 1.
y 00 + 5y 0 + 4y = 2δ(t − 3), y(0) = 1, y 0 (0) = 0.
3. Solve the initial value problem:
y 00 − 4y 0 + 4y = 2δ(t − 4), y(0) = 1, y 0 (0) = 0
4. Use the Laplace transform to solve the initial value problem
y 00 − 2y 0 − 15y = e2t + 2δ(t − 1), y(0) = 0, y 0 (0) = 0.
5. Evaluate L{δ(t − 1)t10 tan2 (tπ/4)}

6. (a) Use the Laplace transform to solve the following initial value problem
y 00 + 9y = t − u2 (t) + δ(t − 1), y(0) = 0, y 0 (0) = 1.
(b) Evaluate y(π/3).

7. True or false. Justify your answers.
L{δ(t − 2π)et cos(t)} = e−2π(s+1)

y 00 + 3y 0 + 2y = u6 (t) − 2δ(t − 1), y(0) = 0, y 0 (0) = 2.
(b) Evaluate y(ln 2).

t
9. Find L{δ(t − π) sin( )(3t2 − 1)}.
2
10. Solve the following using Laplace transforms.
y 00 + 2y 0 + 2y = δ(t − 3), y(0) = 0, y 0 (0) = 0.
11. Solve the initial value problem:
y 00 + 2y 0 + 5y = 3δ(t − 2π), y(0) = 1, y 0 (0) = 0.
12. Solve the following IVP:
y 0 + 3y = δ(t − 1) + u(t − 2), y(0) = −4.
13. Use the Laplace transform to solve the following
y 0 + 3y = tu(t − 2) + δ(t − 2), y(0) = 1.
y 00 + 4y 0 + 8y = δ(t − π), y(0) = 0, y 0 (0) = −1.
y 000 − 2y 00 + y 0 + 4y = e−2t + δ(t − π), y(0) = 0, y 0 (0) = 1, y 00 (0) = 0.
Find L{f (t)} = Y (s), the Laplace transform of its solution. You do not need to simplify
your answer. Do not solve for its inverse transform, y(t)!
y 00 + 6y 0 + 9y = δ(t) − 2u5 (t), y(0) = 0, y 0 (0) = −1.

Z ∞
π
e−st δ(t − ) cos t dt.
0 2
(a) 0
π 1
(b) e− 2 s
s2
+1
−π s s
(c) e 2 2
s +1
−π s π
(d) e 2
2
18. Use the Laplace transform to solve the following initial value problem.
y 00 + 4y 0 + 3y = δ(t) − u6 (t), y(0) = 2, y 0 (0) = 0.


Z ∞
π
e−st δ(t − ) sin 3t dt.
0 6
π
(a) e− 6 s
π 3
(b) e− 6 s
s2 +9
π s
(c) e 6 s
s2 + 9
π
(d) −e 6 s
20. True or false.
−s
(a) L{δ(t − 1)t10 tan2 ( tπ
4 )} = e .
(b) Suppose f (2) = 0, then L{δ(t − 2)ef (t) } = e−2s
y 00 + 9y = t − u2 (t) + δ(t − 1), y(0) = 0, y 0 (0) = 1.
(b) Evaluate y( π3 )
Answers
1 2t 1 −2t 1 1
1. y(t) = e − e + u2π (t)[ e2(t−2π) − e−2(t−2π) ]
4 4 4 4
1 −4t 4 −t 2 −(t−3) 2 −4(t−3)
2. y(t) = − e + e + u3 (t)( e − e )
3 3 3 3
2t 2t 2(t−4)
3. y(t) = e − 2te + u4 (t)2(t − 4)e
−1 2t 1 1 1 1
4. y(t) = e + e−3t + e5t + u1 (t)( e5(t−1) − e−3(t−1) )
15 40 24 4 4
5. e−s
t sin 3t 1 cos 3(t − 2) sin 3(t − 1) sin 3t π sin 3
6. a) − − u2 (t)( − ) + u1 (t) + , b) +
9 27 9 9 3 3 27 3
7. F
1 1 1
8. a) y(t) = 2e−t − 2e−2t + 2u1 (t)(e−2(t−1) − e−(t−1) ) + u6 (t)( − e−(t−6) + e−2(t−6) ), b)
2 2 2
9. (3π 2 − 1)e−πs
10. y(t) = u3 (t)e−(t−3) sin(t − 3)
1 3
11. y(t) = e−t cos 2t + e−t sin 2t + u2π (t)e−t+2π sin 2t
2 2
1
12. y(t) = −4e−3t + u(t − 1)e−3(t−1) + 1 − e−3(t−2) u(t − 2)
3
2 −3(t−2) 7 1
13. y(t) = u(t − 2) e + − (t − 2)
9 9 3
1 −2t 1
14. y(t) = − e sin 2t + uπ (t)e−2(t−π) sin(2t − 2π)
2 2
1 e−πs s−2
15. Y (s) = + 3 +
(s +2)(s − 2s + s + 4) s − 2s2 + s+ 4 s3 − 2s2 + s + 4
3 2
2 2 2
16. y(t) = u5 (t) − + e−3t+15 + (t − 5)e−3t+15
9 9 3
17. a)
7 −t 3 −3t 1 −t+6 1 1 −3t+18
18. y(t) = e − e + u6 (t) e − − e
2 2 2 3 6
19. a)
20. a) T, b) T
t sin 3t 1 cos 3(t − 2) sin 3(t − 1) sin 3t π sin(π − 3)
21. a) y(t) = − − u2 (t) − + u1 (t) + , b) +
9 27 9 9 3 3 27 3
Chapter 6
Systems of First Order Linear

Equations
6.1 Introduction
A system of n-linear first order differential equations in n unknowns has the following general
form:
x01 = a11 x1 + a12 x2 + · · · + a1n xn + g1

x02 = a21 x1 + a22 x2 + · · · + a2n xn + g2
..
.
x0n = an1 x1 + an2 x2 + · · · + ann xn + gn
where aij ’s and gi ’s are arbitrary functions of t.

• If every gi is a zero function, then the system is called to be homogeneous.
• Otherwise, it is a nonhomogeneous system if one of the gi ’s is a nonzero function.
Actually a system of n-linear first order differential equations is equivalent to, mathematically the
same thing, an nth order differential equation.The followings are how we can transform them back
and forth.
We start with the method to transform an nth order differential equation to a system of n-first
order linear equations. Given
an y (n) (t) + an−1 y (n−1) (t) + . . . + a1 y 0 (t) + a0 y(t) = g(t)
be an nth order linear differential equation.
138
CHAPTER 6. SYSTEMS OF FIRST ORDER LINEAR EQUATIONS 139
How to transform nth order equation to a system of n first order linear equations:
1. Introduce n new variables x1 , x2 , . . . , xn defined by
x1 = y, x2 = y 0 , ..., xn = y (n−1) .
2. Then it follows that
x01 = x2 , x02 = x3 , ..., x0n−1 = xn
are n − 1 first order linear equations.

3. Rewrite the given equation in the following form,
an−1 (n−1) a1 0 a0 g(t)

y (n) = − y − ... − y − y+ .
an an an an
Then change the equations to new variables introduced in step 1 to receive
an−1 a1 a0 g(t)
x0n = − xn − . . . − x2 − x1 + .
an an an an
4. Equations from step 2 and 3 form a system of n first order linear equations.

Transform the equation below into a system of 1st order equations.
2y (4) + t2 y 000 − 2y 0 + 6ty = 3 sin t.
Solution Introduce 4 new variables since the given equation is the 4th order one,
x1 = y, x2 = y 0 , x3 = y 00 , x4 = y 000 .
Then it follows that

x01 = x2 , x02 = x3 , x03 = x4 .
Now rewrite the given equation and change its variables to new ones introduced earlier
t2 000 3 t2 3
y (4) = − y + y 0 − 3ty + sin t ⇒ x04 = − x4 + x2 − 3tx1 + sin t.
2 2 2 2
Then  0

x1 = x2
x0

= x3
2
0

x 3 = x4
t2

 0 3
x4 = −3tx1 + x2 − 2 x4 + 2 sin t
form the system of 1st order linear equations.

Find the system of 1st order linear equations which is equivalent to the 3rd order linear
equation
y 000 + y 00 − 2y 0 + 3y = 0.
Solution The equation above is the third order one so we need to introduce 3 new variables,
x1 = y, x2 = y 0 , x3 = y 00 .
Then it follows that

x01 = x2 , x02 = x3 .
Rewrite the given equation and change its variables to new ones introduced earlier
y 000 = −y 00 + 2y 0 − 3y ⇒ x03 = −x3 + 2x2 − 3x1 .
Hence the system of 1st order equations is


0
x 1 = x 2

x02 = x3
 0

x3 = −3x1 + 2x2 − x3 .
The reverse is also true. Given a system of n 1st order linear equations, it can be rewritten into
a single nth order linear equation. We take a look here for the special case when n = 2.
Let
x01 = ax1 + bx2

x02 = cx1 + dx2 .
be a given system of 1st order linear equations.
How to transform a system of two 1st order linear equations to a 2nd order
equation:
1. From the first equation, solve for x2 :
x01 ax1
x2 = − .
b b
2. Then substitute x2 we get from step 1 to the second equation.

Find 2nd order equation which is equivalent to the given linear system
x01 = x2
x02 = 3x1 − 2x2 .
Solution From the first equation, we have x2 = x01 . Substitute it back into the second one,
x02 = 3x1 − 2x2 ⇒ (x01 )0 = 3x1 − 2(x1 )0 ⇒ x001 = 3x1 − 2x01 .
Hence the equivalent second order equation is
x001 + 2x01 − 3x1 = 0.

Example 104. Transform the given system into a single equation of second order
5 3
x01 = x1 + x2 , x1 (0) = −2
4 4
3 5
x02 = x1 + x2 , x2 (0) = 1.
4 4
Solution From the first equation, we have

5 3 4 5 4 0 5
x01 = x1 + x2 ⇒ x2 = 0
x1 − x1 ⇒ x2 = x − x1 .
4 4 3 4 3 1 3
Substitute it back into the second equation,

3 5 4 5 3 5 4 5
x02 = x1 + x2 ⇒ ( x01 − x1 )0 = x1 + ( x01 − x1 )
4 4 3 3 4 4 3 3
4 00 5 0 3 5 0 25
⇒ x − x = x1 + x1 − x1
3 1 3 1 4 3 12
4 00 10 0 4
⇒ x − x1 + x1 = 0
3 1 3 3
⇒ 4x001 − 10x01 + 4x1 = 0
⇒ 2x001 − 5x01 + 2x1 = 0
Therefore we have
4x001 − 10x01 + 4x1 = 0
as an equivalent single equation of second order. Now you can see that the equation we just
derived is in terms of x1 . So an initial condition x1 (0) = −2 works but we need to change the
second one x2 (0) = 1 to variable x1 . To do this, we use the first equation from the problem,
5 3 5 3 5 3 7
x01 = x1 + x2 ⇒ x01 (0) = x1 (0) + x2 (0) ⇒ x01 (0) = (−2) + (1) = − .
4 4 4 4 4 4 4
In the end, we have the following IVP
7
2x001 − 5x01 + 2x1 = 0, x1 (0) = 2, x01 (0) = − .
4
6.2 Review of Matrices

An m × n (called the size or dimension of the matrix) matrix is a rectangular array of numbers
with m rows and n columns and the entry in the ith row and jth column is denoted by aij .
 
 a11 a12 ··· a1n 
 
 
a
 21 a22 ··· a2n 

A=
 .
 = (aij )m×n
 .. .. .. .. 
 . . . 

 
 
am1 am2 ··· amn
m×n
Example 105.  
1
 
1 −1 4  
 



 , 2 0 3 4 5 ,
2

1×5 
2 3 5  

2×3
−3
3×1
Le’t consider special-name matrices here:
1. Square matrix is any matrix which has the same number of rows as columns.
 
 a11 a12 ··· a1n 
 
 
a
 21 a22 ··· a2n 

A=
 .

 .. .. .. .. 
 . . . 

 
 
am1 am2 ··· ann
n×n
The diagonal that start in the upper left and end in the lower right is called the main
diagonal. The sum of the entires in the main diagonal is called trace and write it as tr. So
here
Xn
tr A = aii = a11 + a12 + . . . + ann .
i=1
2. Zero matrix, denoted 0m×n or 0, is a matrix which all entries are zeros. It satisfies two
properties:
A+0=A=0+A
A0 = 0 = 0A.
3. Identity matrix is a square matrix, denoted by In or I, whose main diagonals are all 1’s
and other entries are zero.  

1 0 ··· 0
 
 
0 1
 ··· 0 
 
. . .. .. 
 .. .. . .
 
 
 
0 0 ··· 1
It satisfies AI = A = IA.
4. Column(row) matrix or Column(row) vector is a matrix consisting of a single column
(row).
 
 x1 
 
 
x 
 2
x=
 . 
 is a column vector and y = y
1 y2 ··· yn is a row vector.
 .. 
 
 
 
xn
Next let’s consider the arithmetic of matrices.

1. Addition, subtraction and scalar multiplication:
     
a b  e f  a ± e b ± f 
 ± = 
     
c d g h c±g d±h
   
a b  αa αb 
α

=
 


c d αc αd
2. Matrix multiplication of A and B: It is defined only when A has the same number of
columns as B has rows. Otherwise, we can’t compute AB.
Am×p Bp×n = ABm×n
where its ijth entry (AB)ij can be found by the dot product of row vector i of matrix A
and the transpose of column vector j of matrix B. Here are some examples:
     
a b e f ae + bg af + bh







 = 



c d g h ce + dg cf + dh
2×2 2×2 2×2
     
a b e ae + bf 




 
  = 



c d f ce + df
2×2 2×1 2×1
 
a b
But 


 e f is not defined.
1×2
c d
2×2
Unlike multiplication of real numbers , matrix multiplication does not commute. In other
words, it is NOT true that AB equals to BA.
3. Determinant, denoted by det, is a function that takes a square matrix into a number in a
certain way. In particular for 2 × 2 matrix, we define determinant as follows
 

a b  a b

det  = := ad − bc
 
c d c d

• If the determinant of a matrix is zero, we call that matrix singular.

• We call the matrix nonsingular if its determinant is nonzero.
4. Inverse matrix of a square matrix: We call B the inverse of A, denoted by A−1 , if
AB = BA = In .
If A−1 exists, then we say A is invertible.

 
a b
• For 2 × 2 matrix if A = 

 then

c d
   
1  d −b  d −b
A−1 =  = 1  .
ad − bc   det A  
−c a −c a
• It is the fact that a square matrix is invertible if and only if its determinant is nonzero.
   
2 −1 2 −1
Example 106. Let A = 

 and B = 
 
 . Then

3 1 −6 3
det A = 2(1) − 3(−1) = 5 6= 0, det B = 2(3) − (−6)(−1) = 0.
So B is not invertible while A is invertible with

   
1 1
1 1 1  5 5
A−1 =  = .
5   
−3 2 − 35 2
5
5. Eigenvalues and eigenvectors: Given a square matrix A, suppose there are constant r
and a nonzero column vector x such that
Ax = rx,
then r is called eigenvalue of A and x is called eigenvector of A corresponding to r. How

do we find them? First, we rewrite the equation,
Ax = rx ⇒ Ax − rx = 0 ⇒ (A − rI)x = 0.
Then apply the following theorem,

Theorem 12. Let A be n × n matrix then the system Ax = 0 has

• a unique solution x = 0 if and only if A is invertible (det A 6= 0).
• infinitely many solutions if and only if A is not invertible (det A = 0).
Since we’re looking for a nontrivial solution x, we receive

det(A − rI) = 0.
Now take a look in the particular case of 2 × 2 matrix A,
     
a b 1 0 a − r 0 
A − rI = 

−r
 
=
 
.

c d 0 1 0 d−r
And so det(A − rI) = 0 means

(a − r)(d − r) − 0(0) = 0 ⇒ r2 − (a + d)r + (ad − bc) = 0
⇒ r2 − tr(A)r + det A = 0.
We call r2 − tr(A)r + det A the characteristic polynomial corresponding to A and the
equation
r2 − tr(A)r + det A = 0
the characteristic equation of matrix A.
 
a b
In conclusion, for A = 

, we have

c d
How to find eigenvalues and eigenvectors for A:
1. Write the characteristic equation
r2 − tr(A)r + det A = 0.
Its roots give eigenvalues of A.

2. Plug in each eigenvalue back to (A − rI)x = 0 and solve for x to find the corresponding
eigenvectors.
Example 107. Find the eigenvalues and eigenvectors of

 
3 −1
A=

.

4 −2
Solution First, find the roots of its characteristic equation:
r2 − (3 − 2)r + 3(−2) − 4(−1) = r2 − r − 2 = 0 ⇒ (r − 2)(r + 1) = 0 ⇒ r = 2, −1.
Hence the eigenvalues are 2 and −1.

For r = 2: Consider (A − 2I)x = 0,

            
3 −1 1 0 x1  0 3 − 2 −1  x1  0
  − 2    =   ⇒    =  
            
4 −2 0 1 x2 0 4 −2 − 2 x2 0
    
1 −1 x1  0
⇒ 

  =  
   
4 −4 x2 0
(
x1 − x2 = 0,
⇒
4x1 − 4x2 = 0.
⇒ x1 = x2 .
     
x1  x2  1
So the eigenvectors are 
  =   = x2   . For simplicity, we choose x2 = 1. Then
    
x2 x2 1
 
1
eigenvector corresponding to r = 2 is 
 .

1
For r = −1: Consider (A − (−1)I)x = 0,

            
3 −1 1 0 x1  0 3 + 1 −1  x1  0
  − (−1)     =   ⇒    =  
            
4 −2 0 1 x2 0 4 −2 + 1 x2 0
    
4 −1 x1  0
⇒ 

  =  
   
4 −1 x2 0
(
4x1 − x2 = 0,
⇒
4x1 − x2 = 0.
⇒ x2 = 4x1 .
     
x1   x1  1
So the eigenvectors are 
 =
   = x1   . For simplicity, we choose x1 = 1. Then
  
x2 4x1 4
 
1
eigenvector corresponding to r = −1 is 
 .

4
Note that for each value of an eigenvalue, its corresponding eigenvector is not unique since nonzero
scalar multiplication of an eigenvector is also an eigenvector.
Exercises 6.1 − 6.2

1. Which system of first order linear equations below is equivalent to the third order linear
equation
y 000 + y 00 − 3y 0 − y = 0?

0
x 1 = x 2

(a) x02 = x3
 0

x3 = x1 − 3x2 − x3

0
x 1 = x 2

(b) x02 = x3
 0

x3 = −x1 + 3x2 + x3

0
x 1 = x 2

(c) x02 = x3
 0

x3 = −x1 − 3x2 + x3

0
x 1 = x 2

(d) x02 = x3
 0

x3 = x1 + 3x2 − x3
2. Rewrite the following third order linear equation into an equivalent system of first order
linear equations.
y 000 + 3y 00 − 2y 0 + 4y = sin 2t
3. Determine whether the statement is true or false. The system of first order linear equations

0
x 1 = x 2

x02 = x3
 0 3
x3 = 38 x1 − t3 x2 + 13 sin(8t)

is equivalent to the third order linear equation 3y 000 + t3 y 0 − 8y = sin(8t).
4. Which system of first order linear equations below is equivalent to the second order linear
equation
y 00 − 4y 0 + 7y = 10t3 ?
(
x01 = x2
(a)
x02 = −7x1 + 4x2 + 10t3
(
x01 = x2
(b)
x02 = 7x1 − 4x2 + 10t3
(
x01 = x2
(c)
x02 = −4x1 + 7x2 + 10t3
(
x01 = x2
(d)
x02 = 4x1 − 7x2 + 10t3
5. Which second order differential equation below is equivalent to the system

   
0 1   0 
x0 = 

x + 
 
?

1 −2 sin t
(a) y 00 + 2y 0 − y = sin t
(b) y 00 + sin ty 0 + y = 2
(c) y 00 − y 0 + 2y = sin t
(d) y 00 + y 0 − 2y = sin t
6. Which system of first order linear equations below is equivalent to the second order linear
equation
y 00 − 5y 0 + 6y = t2 − t?
(
x01 = x2
(a)
x02 = −5x1 + 6x2 + t2 − t
(
x01 = x2
(b)
x02 = 5x1 − 6x2 − t2 + t
(
x01 = x2
(c)
x02 = −6x1 + 5x2 + t2 − t
(
x01 = x2
(d)
x02 = 6x1 − 5x2 − t2 + t
7. Rewrite the following equation into an equivalent system of first order linear equations.
2y 00 + 6y 0 − 4y = e−t cos 9t
8. What second order differential equation is equivalent to the system

   
 0 1  0
x0 = 

 x +  ?
  
−5t sin t t2
(a) y 00 + 5ty 0 − sin ty = t2

(b) y 00 − 5ty 0 + sin ty = t2
(c) y 00 − t2 y 0 + 5ty = sin t
(d) y 00 − sin ty 0 + 5ty = t2
Answers
1. d)

0
x 1
 = x2
2. x02 = x3
 0

x3 = −4x1 + 2x2 − 3x3 + sin 2t
3. True
4. a)
5. a)
6. c)
(
x01 = x2
7.
x02 = 2x1 − 3x2 + 12 e−t cos 9t
8. d)
6.3 Solutions to Systems of First Order Linear Equations

Recall a system of n-linear first order differential equations is in the form
x01 = a11 x1 + a12 x2 + · · · + a1n xn + g1
x02 = a21 x1 + a22 x2 + · · · + a2n xn + g2
..
.
x0n = an1 x1 + an2 x2 + · · · + ann xn + gn
which can be rewritten into a matrix-vector equation,
      
 x01  a11 a12 ··· a1n   x1   g1 
      
      
 x0   a a22 ··· a2n 
  x2   g2 
   
 2   21
 =   +  ,
 .   . .. .. ..   .  .
 ..   .. . . .   .  .
    .   . 
      
      
x0n am1 am2 ··· ann xn gn
or just simply
x0 = Ax + g.
In this class, we consider only for homogeneous system
x0 = Ax.
• Equilibrium solutions or critical points are solutions x for which x0 = 0 (or Ax = 0).
Since we’re going to assume that det A 6= 0, by the theorem above, we will have only one
equilibrium solution, namely x = 0. So in this chapter, the only critical point is the origin.
• A sketch of a particular solution in the phase plane x1 x2 plane is called the trajectory of
the solution.
• A plot that shows a representative sample of trajectories for a given system is called a phase
portrait.
• There are three types of stability of the system based on the behavior of trajectories:
1. It is asymptotically stable if all trajectories converge to zero.
2. It is unstable if all or all but a few trajectories move away from origin.
3. It is (neutrally) stable if all trajectories stay in a fixed orbit around the origin.
• Solutions to the homogeneous system x0 = Ax: For n = 1 and write A = [r], we know that
the solution to x0 = rx is x = kert . So let’s use this as a guideline for general n. If x = kert
is a solution of a system, then x0 = rkert and so
x0 = Ax ⇒ rkert = Akert ⇒ Ak = rk,
where r is an eigenvalue of A and k is an eigenvector of A corresponding to r.
• For this class, we consider when A is 2 × 2 matrix. Just like the solution of a second order
homogeneous linear equation, there are three possibilities of solutions depending on a type
of eigenvalues of A.
1. Two distinct real eigenvalues.
2. Complex conjugate eigenvalues.
3. Repeated eigenvalues.
6.3.1 Real and Distinct Eigenvalues

If the coefficient matrix A has two real distinct eigenvalues r1 and r2 with corresponding eigen-
vectors k1 , k2 respectively. Then the system x0 = Ax has a general solution
x = c 1 k 1 e r1 t + c 2 k 2 e r2 t .
Case 1.1: r1 , r2 have opposite signs.
Example 108. a) Solve the following IVP

   
1 2 0
x0 = 

 x,
  .
x(0) =  
3 2 −4
 
1 2
Solution Firstly, we find the eigenvalues of the coefficient matrix A = 

 . Its

3 2
characteristic equation is
r2 − (1 + 2)r + 1(2) − 3(2) = r2 − 3r − 4 = 0 ⇒ (r + 1)(r − 4) = 0 ⇒ r = −1, 4.
Hence the eigenvalues are −1 and 4.
For r1 = −1: Consider (A − (−1)I)k1 = 0,

         
1 − (−1) 2  m1  0 2 2 m1  0


  =  
    ⇒ 

  =  
   
3 2 − (−1) m2 0 3 3 m2 0
(
2m1 + 2m2 = 0,
⇒
3m1 + 3m2 = 0.
⇒ m1 = −m2 .
So the eigenvector corresponding to r1 = −1 is

     
m1  −m2  Choose m =1 −1
 =
  

 =======2=⇒ k1 = 
 .

m2 m2 1
For r2 = 4: Consider (A − 4I)k2 = 0,

         
1 − 4 2  n1  0 −3 2  n1  0


  =  
    ⇒ 

  =  
   
3 2−4 n2 0 3 −2 n2 0
(
−3n1 + 2n2 = 0,
⇒
3n1 − 2n2 = 0.
2
⇒ n1 = n2 .
3
So the eigenvector corresponding to r2 = 4 is

     
2
 n1   3 n2  Choose n =3 2
 =
  

 =======2=⇒  .
k2 =  
n2 n2 3
The general solution is then

   
−1 2
x(t) = c1 e−t  4t  
  + c2 e   .

1 3
Apply the given initial condition to determine the values c1 and c2 ,

     
( (
0 −1 2 −c1 + 2c2 = 0, c1 = − 58 ,
  = c1   + c2   ⇒ ⇒ .
      c1 + 3c2 = −4 c2 = − 45
−4 1 3

     
8 −t 8 4t
8 −1 4 4t 2  − 5e 5e
x(t) = − e−t 

− e  = .
5   5    
1 3 − 85 e−t − 12 4t
5 e
 
1 2
b) Sketch the phase portrait of x0 = 

 x. Then find its type and stability.

3 2
Type: The critical point (0, 0) is called a saddle point.

Stability: Since some solutions move away from the origin (0, 0) as t increases, it is
unstable.
Case 1.2: r1 , r2 have same signs. (they are either both positive or both negative)

   
−5 1 1
x0 = 

 x,
 x(0) = 
 .

4 −2 2
 
−5 1
Solution Firstly, we find the eigenvalues of the coefficient matrix A = 

.

4 −2
r2 −(−5−2)r+(−5)(−2)−4(1) = r2 +7r+6 = 0 ⇒ (r+1)(r+6) = 0 ⇒ r = −1, −6.
So the eigenvalues are −1 and −6.

         
−5 − (−1) 1  m1  0 −4 1  m1  0


  =  
    ⇒ 

  =  
   
4 −2 − (−1) m2 0 4 −1 m2 0
(
−4m1 + m2 = 0,
⇒
4m1 − m2 = 0.
⇒ m2 = 4m1 .

     
m1   m1  Choose m =1 1
 =
  

 =======1=⇒ k1 = 
 .

m2 4m1 4

         
−5 − (−6) 1  n1  0 1 1 n1  0


  =  
    ⇒ 

  =  
   
4 −2 − (−6) n2 0 4 4 n2 0
(
n1 + n2 = 0,
⇒
4n1 + 4n2 = 0.
⇒ n1 = −n2 .

     
n1  −n2  Choose n =1 −1
 =
  

 =======2=⇒  .
k2 =  
n2 n2 1
Then the general solution is

   
1 −1
x(t) = c1 e−t 
  + c2 e
 −6t 
 .

4 1
Apply the given initial condition to get the values c1 and c2 ,

     
( (
1 1 −1 c1 − c2 = 1, c1 = 53 ,
  = c1   + c2   ⇒ ⇒ .
      4c1 + c2 = 2 c2 = − 25 .
2 4 1

     
3 −t 2 −6t
3 −t 1 2 −1  + 5e 5e
 − e−6t   = 

x(t) = e  .
5   5   
12 −t 2 −6t

4 1 5 e − 5 e
 
−5 1


4 −2
Type: The critical point (0, 0) is called a node.

Stability: Since all trajectories move towards (0, 0) as t increases, it is asymptotically

stable since both r1 , r2 are negative. (But this node can be unstable if both r1 , r2
are positive numbers.)
6.3.2 Complex Conjugate Eigenvalues

If the coefficient matrix A has complex conjugate eigenvalues r1,2 = λ ± µi, then similar to the
first case, the general solution is
x = c1 k1 e(λ+µi)t + c2 k2 e(λ−µi)t .
But what we want is to rewrite it in the form of real valued solution. In order to do this, we again
use the help from Euler’s formula. Let k1 = a + ib be the eigenvector corresponding to r1 = λ + µi.
Then it follows that k2 = a − ib is the eigenvector corresponding to r2 = λ − µi. Then we can
derive new form of general solution as follows:
x = c1 k1 e(λ+µi)t + c2 k2 e(λ−µi)t
= c1 (a + ib)eλt eµit + c2 (a − ib)eλt e−µit
= c1 (a + ib)eλt (cos µt + i sin µt) + c2 (a − ib)eλt (cos µt − i sin µt)
= c1 eλt (a cos µt − b sin µt + i(a sin µt + b cos µt))
+ c2 eλ (a cos µt − b sin µt − i(a sin µt + b cos µt))
= (c1 + c2 )eλt (a cos µt − b sin µt) + (c1 − c2 )ieλ (a sin µt + b cos µt)
After renaming the constant, the new form of the general solution is
x(t) = c1 eλt (a cos µt − b sin µt) + c2 eλ (a sin µt + b cos µt) ,
where the eigenvalues of A is λ ± µi and the eigenvector corresponding to λ + µi is a + ib.

Case 2.1: Real part is zero. (r1,2 = ±µi)

   
3 −9 2
x0 = 

 x,
 x(0) = 
 .

4 −3 −4
Solution The characteristic equation is

√ √
r2 − (3 − 3)r + (3)(−3) − 4(−9) = r2 + 27 = 0 ⇒ r = ± 27i = ±3 3i.
√ √
So the eigenvalues are r1,2 = 0 ± 3 3i. Note that λ = 0 and µ = 3 3.
√ √
For r1 = 3 3i: Consider (A − (3 3i)I)k1 = 0,
    
√ ( √
3 − 3 3i −9  m1  0 (3 − 3 3i)m1 − 9m2 = 0,
   =   ⇒ √
 √     4m1 + (−3 − 3 3i)m2 = 0.
4 −3 − 3 3i m2 0
1 √
⇒ m2 = (1 − 3i)m1 .
3
√
So the eigenvector corresponding to r1 = 3 3i is
     
m1   m1  Choose m =3  3 
 =  =======1=⇒ k1 =  .
  
1
√   √ 
m2 3 (1 − 3i)m1 1 − 3i
Write k1 in terms of a + ib,

       
 3   3 + 0i  3  0 
k1 =  =  =   + i
 √  = a + ib.

 √   √   
1 − 3i 1 − 3i 1 − 3

         
3 √  0  √  3 √  0  √ 
x(t) = c1 e0t    0t  
  cos 3 3t −  √  sin 3 3t + c2 e   sin 3 3t +  √  cos 3 3t
     
1 − 3 1 − 3
   
√ √
 3 cos 3 3t   3 sin 3 3t 
= c1   + c2  .
 √ √ √   √ √ √ 
cos 3 3t + 3 sin 3 3t sin 3 3t − 3 cos 3 3t.
Apply initial condition,

     
( (
2 3  0  3c1 = 2, c1 = 32 ,
  = c1   + c2   ⇒ √ ⇒
     √  c1 − 3c2 = −4. c2 = 314
√ .
3
−4 1 − 3

   
√ √
2 3 cos 3 3t 3 sin 3 3t
 + 14
  
x(t) =  √  
3 √ √ √  3 3 √ √ √ 
cos 3 3t + 3 sin 3 3t sin 3 3t − 3 cos 3 3t
   
√ 14
√
 2 cos 3 3t   √ sin 3 3t
3 
= + 

2
√ √
2 3
√   14 √ 14
√ 
√ sin 3 3t −
3 cos 3 3t + 3 sin 3 3t 3 3 3 cos 3 3t
 
√ 14
√
 2 cos 3 3t + √ 3
sin 3 3t 
= 

2 14
√ √
2 3 14
√ 
( 3 − 3 ) cos 3 3t + ( 3 + 3√3 ) sin 3 3t
 
√ 14
√
 2 cos 3 3t + √3 sin 3 3t 
∴ x(t) =  .
 √ 20
√ 
−4 cos 3 3t + 3√3 sin 3 3t
 
−5 1


4 −2
Type: The critical point (0, 0) is called a center.

Stability: It is (neutrally) stable since the trajectories of the solutions are circles or
ellipses centered at the origin. To check the orientation, pick a sample point and
see what we get.
      
1 3 −9 1 3
x= 
  ⇒ x0 = 

  =  .
   
0 4 −3 0 4
At the point (0, 1), the trajectory points in upward direction. So the trajectories
travel in counter clockwise direction.
Case 2.2: Real part is nonzero. (r1,2 = λ ± µi, where λ 6= 0)

   
3 −13  3 
x0 = 

 x,
 x(0) = 

.

5 1 −10
Solution Consider the characteristic equation,
r2 − (3 + 1)r + (3)(1) − 5(−13) = r2 − 4r + 68 = 0 ⇒ r = 2 ± 8i.
So the eigenvalues are r1,2 = 2 ± 8i. Here λ = 2 and µ = 8.

For r1 = 2 + 8i: Consider (A − (2 + 8i)I)k1 = 0,

    
(
3 − (2 + 8i) −13  m1  0 (1 − 8i)m1 − 13m2 = 0,
   =   ⇒
     5m1 + (−1 − 8i)m2 = 0.
5 1 − (2 + 8i) m2 0
1 + 8i
⇒ m1 = m2 .
5
So the eigenvector corresponding to r1 = 2 + 8i is
     
1+8i
m1   5 m2  Choose m =5 1 + 8i
 =
  

 =======2=⇒ k1 = 

.

m2 m2 5
Write k1 in terms of a + ib,

       
1 + 8i 1 + 8i 1 8
k1 = 

=
 
 =   + i   = a + ib.
    
5 5 + 0i 5 0

         
1 8 1 8
x(t) = c1 e2t  2t  
 
  cos 8t −   sin 8t + c2 e   sin 8t +   cos 8t
       
5 0 5 0
   
cos 8t − 8 sin 8t sin 8t + 8 cos 8t
= c1 e2t 

 + c2 e2t 
 
.

5 cos 8t 5 sin 8t
Apply the given initial condition,

     
( (
 3  1 8 c1 + 8c2 = 3 c1 = −2,
  = c1   + c2   ⇒ ⇒ .
      5c1 = −10 c2 = 85
−10 5 0
Hence the particular solution is

   
cos 8t − 8 sin 8t 5 2t sin 8t + 8 cos 8t
x(t) = −2e2t  + e  
  8  
5 cos 8t 5 sin 8t
 
133
 3e2t cos 8t + 8 e2t sin 8t 
=

.

2t 25 2t
−10e cos 8t + 8 e sin 8t
 
3 −13


5 1
Type: The critical point (0, 0) is called a spiral.

Stability: It is unstable since real part λ = 2 of the eigenvalues is positive one. (But this
spiral point can be asymptotically stable if real part of the eigenvalues is negative
one). The trajectories spiral into (if λ < 0) or out of the origin (if λ > 0) which
in the case in this example. To check the orientation, pick a sample point and see
what we get.
      
1 3 −13 1 3
x=
 
 ⇒ x0 = 

  =  .
   
0 5 1 0 5
At the point (0, 1), the trajectory points in upward direction. So the trajectories
travel in counter clockwise direction.
6.3.3 Repeated Real Eigenvalues

Suppose the coefficient matrix A has repeated real eigenvalues r1 = r2 = r.
 
r 0
Case 3.1: A is in the form of   where r 6= 0. If this happens, we can choose arbitrary two
 
0 r
linearly independent eigenvectors k1 and k2 (Choose in such a way that k1 6= ck2 ). So the general
solution to x0 = Ax is
x(t) = c1 k1 ert + c2 k2 ert .
 
3 0
Example 112. a) Solve the following system x0 = 

 x.

0 3
Solution Consider the characteristic equation,
r2 − (3 + 3)r + 3(3) − 0(0) = r2 − 6r + 9 = (r − 3)2 = 0 ⇒ r = 3, 3.
So the eigenvalues are r1 = r2 = 3.

For r1 = r2 = 3: Consider (A − 3I)k1 = 0,
    
3 − 3 0  m1  0


  =  
    ⇒ 0 = 0,
0 3−3 m2 0
which gives us nothing. So we can choose arbitrary two linearly independent eigenvec-
tors, for example,    
1 0
k1 = 
 ,
 k2 = 
 .

0 1

     
3t
1 3t 0 3 c1 e 
  e + c2   e t = 
x(t) = c1      .

3t
0 1 c2 e
 
3 0


0 3
The phase portrait has a star-burst shape.

Type: The critical point (0, 0) is called a proper node or star point.
Stability: It is unstable since repeated eigenvalues, r = 3, are positive. (But this proper
node can be asymptotically stable if repeated eigenvalues are negative ones). The
trajectories move towards the origin (if r < 0) or move away from the origin (if
r > 0) which is the case in this example.
 
r 0
Case 3.2: Matrix A gives repeated eigenvalues but is not in the form of   where r 6= 0.
 
0 r
In this case, we will find only one eigenvector k corresponding to repeated eigenvalues r. So for
sure, one solution is kert . We need to find the second solution. As we studied before, it is natural
to guess that the second one is ktert . To check whether it work or not, plug it back into x0 = Ax,
x0 = Ax ⇒ (ktert )0 = Aktert ⇒ rktert + kert = Aktert .

(
rk = Ak ⇒ This is the same as saying k is the eigenvector corresponding to r.
rt
ke ⇒ k=0 ⇒ This is a contradiction. Since an eigenvector k can’t be zero.
So our first guess is incorrect. Let’s choose something new but not far from the first guess. How’s
about ktert + ηert . Checking whether it work or not, substitute it back into x0 = Ax,
x0 = Ax ⇒ (ktert + ηert )0 = A(ktert + ηert )

⇒ rktert + kert + rηert = Aktert + Aηert
⇒ rktert + (k + rη)ert = Aktert + Aηert

(
rk = Ak ⇒ This is again the same as saying k is the eigenvector corresponding to r.
k + rη = Aη ⇒ (A − rI)η = k ⇒ This is new and it gives us how to find the vector η.

x(t) = c1 kert + c2 ktert + ηert ,

where η can be found by (A − rI)η = k.

   
7 1 2
x0 = 

 x,
 x(0) = 
 

−4 3 −5
Solution Its characteristic equation is,
r2 − (7 + 3)r + 7(3) − (−4)(1) = r2 − 10r + 25 = (r − 5)2 = 0 ⇒ r = 5, 5.
So the eigenvalues are r1 = r2 = 5.

For r1 = r2 = 5: Consider (A − 5I)k = 0,

    
(
7 − 5 1  m1  0 2m1 + m2 = 0
   =   ⇒ ⇒ m2 = −2m1 .
     −4m1 − 2m2 = 0
−4 3 − 5 m2 0
Hence the eigenvector corresponding to r = 5 is

     
m1   m1  Choose m =1 1
 =
  

 =======1=⇒  .
k= 
m2 −2m1 −2
Solve for η: Consider (A − 5I)η = k,

    
(
7 − 5 1  n1   1  2n1 + n2 = 1
   =   ⇒ ⇒ n2 = 1 − 2n1 .
     −4n1 − 2n2 = −2
−4 3 − 5 n2 −2
Hence η is      
n1   n1  Choose n =0 0
 =
  

 =======1=⇒ η=
 .

n2 1 − 2n1 1

      
 1  5t  1  0 
x(t) = c1   e + c2   te5t +   e5t  .
      
−2 −2 1
Apply the given initial condition,

     
( (
2 1 0 c1 = 2, c1 = 2,
  = c1   + c2   ⇒ ⇒ .
      −2c1 + c2 = −5 c2 = −1.
−5 −2 1
So the particular solution is

      
 1  5t  1  5t 0 5t 
  e −   te +   e 
x(t) = 2       
−2 −2 1
     
5t 5t 5t 5t
 2e   te   2e − te 
= − = .
     
−4e5t −2te5t + e5t −5e5t + 2te5t
 
7 1


−4 3
Type: The critical point (0, 0) is called a improper node.

Stability: It is unstable since repeated eigenvalues, r = 5, are positive. (But this
improper node can be asymptotically stable if repeated eigenvalues are negative
ones). The trajectories move towards the origin (if r < 0) or move away from
the origin (if r > 0) which is the case in this example. They keep paralleling to
eigenvector. To check the orientation, pick a sample point and see what we get.
      
1 7 1 1  7 
x=
 
 ⇒ x0 = 

  =  .
   
0 −4 3 0 −4
(a) Case 1.1: Saddle Point (b) Case 1.2: Node
(c) Case 2.1: Center (d) Case 2.2: Spiral Point
(e) Case 3.1: Proper Node (f) Case 3.2: Improper Node
Figure 6.1: Phase Portraits

6.3.4 Summary Sheets
Eigenvalues Type Stability General Solutions
1.1) r2 < 0 < r1 Saddle Point Unstable x(t) = c1 k1 er1 t + c2 k2 er2 t ,
1.2) r1 > r2 > 0 Node Unstable where ki is the eigenvector corresponding to ri
r1 < r2 < 0 Node Asymptotically Stable
2.1) r1,2 = µi Center Stable x(t) = c1 (a cos µt − b sin µt) + c2 (a sin µt + b cos µt)
2.2) r1,2 = λ ± µi Spiral Point x(t) = c1 eλt (a cos µt − b sin µt) + c2 eλt (a sin µt + b cos µt),
if λ > 0, Unstable where a + ib is the eigenvector to λ + iµ
if λ < 0, Asymptotically Stable
3.1) Proper node 3.1) x(t) = c1 k1 ert + c2 k2 ert

3) r1 = r2 = r > 0 Unstable
3.2) Improper node where k1 and k2 are linearly independent vectors.
3.1) Proper node 3.2) x(t) = c1 k1 ert + c2 (k2 tert + ηert )

r1 = r2 = r < 0 Asymptotically Stable
CHAPTER 6. SYSTEMS OF FIRST ORDER LINEAR EQUATIONS
3.2) Improper node where η can be found from (A − rI)η = k.
Table 6.1: Systems of First Order Linear Equations

164
Exercises 6.3
1. Match the sketches of phase portraits for 2 × 2 homogeneous linear systems x0 = Ax with
the names of their critical points at the origin
(a) saddle
(b) node
(c) proper node
(d) center
(e) spiral
2. Match the following formulas for general solutions of 2 × 2 homogeneous linear systems
x0 = Ax with the sketches of the phase portraits given in Problem 1:
   
1 1
(a) c1 e2t   + c2 e−t  
   
−1 1
   
 cos t   sin t 
(b) c1 et 

 + c2 et 
 


2 sin t 2 cos t
   
1 0
(c) c1 e−t  −t  
  + c2 e  

0 1
   
1 0
−t  
(d) c1 et 
  + c2 e  

0 1
   
(e) c1 

 + c2 
 


2 sin t 2 cos t
 
0
  of a 2 × 2 homogeneous linear systems
3. Match the three adjectives for the critical point  
0
x0 = Ax with the five general solutions given in the table below by placing one of the letters
A, U, or S in each of the five blanks. Use A for asymptotically stable, U for unstable and S
for stable.
   
1 1
−t  
(a) c1 e2t 
  + c2 e  

−1 1
   
(b) c1 et 

 + c2 et 
 


2 sin t 2 cos t
   
1 0
(c) c1 e−t  −t  
  + c2 e  

0 1
   
1 0
(d) c1 et   + c2 e−t  
   
0 1
   
(e) c1 

 + c2 
 


2 sin t 2 cos t
4. Suppose the phase portrait of the system

 
2 1
x0 = 

x

−1 d
has an improper node at (0, 0). Find the value (or range of values) of d.
5. Find the general solution of the system
 
5 0
x0 = 

 x.

0 5
6. Consider the system of linear equations

   
3 −2 1
x0 = 

 x, x(0) =   .
  
4 −1 −4
Solve the initial value problem and classify the type and stability of the critical point at
(0, 0).
7. Consider the 2 × 2 linear homogeneous system with constant coefficients x0 = Ax.
(a) If the eigenvalues of A are ±i, classify the type and stability of the critical point (0, 0).
 
2
(b) If in addition an eigenvector corresponding to the eigenvalue i is 
 , then write down

i
the real-valued general solution x(t).
8. Consider the system  
3 −4
x0 = 

 x.

1 −2
(a) Find the general solution x(t)

 
6
(b) If x(0) = 
  and t→∞
 lim x(t) = 0, what is β?
β
9. (a) Solve the initial value problem

   
−2 1 4
x0 = 

 x, x(0) =  
  
2 −3 1
(b) What is lim |x(t)|?

t→∞
10. Consider a certain linear system x0 = Ax, where A is a matrix of real numbers with complex
conjugate eigenvalues. Suppose some of its solutions do not reach a limit either as t → +∞,
or as t → −∞. Then the critical point (0, 0) must be a(n)
(a) unstable spiral point.
(b) unstable saddle point.
(c) asymptotically stable improper node.
(d) (neutrally) stable center.
 
2 − α 2 0 
11. Suppose the linear system x0 = 

 x has an unstable proper node at (0, 0).

0 −2α − 1
Determine all possible vlaue(s) of α.
12. For each part below, consider a certain system of two first order linear linear differential
equations in two unknowns, x0 = Ax, where A is a 2 × 2 matrix of real numbers. Based
solely on the information given in each part, determine the type and stability of the system’s
critical point at (0, 0).
(a) Eigenvalues of A are −1 and −6.

(b) Eigenvalues of A are 3 + 7i and 3 − 7i.
(c) Eigenvalues of A are 9i and −9i.
√
(d) Eigenvalues of A are 11 and π 2 .
   
1 1
(e) The general solution is x(t) = c1 e−5t 
 
 + c2 e−5t  
 
−1 0
   
−1  −t 
(f) The general solution is x(t) = c1 et 
 
 + c2 e t 
 √


1 − 3+t
13. Consider a certain linear system x0 = Ax, where A is a matrix of real numbers with distinct
nonzero real eigenvalues. Suppose all of its solutions have a finite limit as t → +∞. Then
the critical point (0, 0) must be a(n)
(a) (neutrally) stable center.

(b) asymptotically stable spiral point.
(c) unstable saddle point.
(d) asymptotically stable node.
   
2 1 −2
x0 = 

 x, x(0) =  
  
−1 0 6
(a) Solve the initial value problem.

(b) Classify the type and stability of the critical point at (0, 0).
15. For each part below, consider a certain system of two first order linear linear differential
equations in two unknowns, x0 = Ax, where A is a 2 × 2 matrix of real numbers. Based
solely on the information given in each part, determine the type and stability of the system’s
critical point at (0, 0).
 
5 −2
(a) The coefficient matrix is 

 x.

0 5
(b) Eigenvalues of A are −e and −e2 .
(c) One of the eigenvalues of A is r = 1 − 49i.
   
 2 cos t   2 sin t 
(d) The general solution is x(t) = c1 

 + c2 
 
.

cos t − 2 sin t sin t + 2 cos t
   
√ −1 √ 2
(e) The general solution is x(t) = c1 e− 7t   + c2 e− 7t  .
   
2 1
16. Consider a certain system of two first order linear differential equations in two unknowns,
x0 = Ax, where A is a matrix of real numbers. Suppose one of the eigenvalues  of the
 4 
coefficient matrix A is r = −5 + 7i, which has a corresponding eigenvector 

. Write

5 + 6i
down the system’s real valued general solution.
17. (a) Find the general solution of the system of linear equations
 
2 12 
x0 = 

 x.

−3 −11
 
5
(b) Find the solution satisfying x(0) = 
 .

0
18. For each part below, consider a certain system of two first order linear differential equations
in two unknowns, x0 = Ax, where A is a 2 × 2 matrix of real numbers. Based solely on
the information given in each part, determine the type and stability of the system’s critical
point at (0, 0).
 
3 2
(a) The coefficient matrix is 

.

−2 −1
(b) One of the eigenvalues of A is r = −2 + 11i.
(c) The characteristic equation of A can be rewritten as r2 + 4 = 0.
   
−4 2
(d) The general solution is x(t) = c1 e2t 
 
 + c2 e2t  .
 
1 −3
19. For what value(s) of β will the linear system below have a (neutrally) stable center at (0, 0)?
 
 2β 5
x0 = 

x

−5 β2
(a) 0
(b) 2
(c) −2, 2
(d) 0, −2
20. Consider a certain system of two first order linear differential equations in two unknowns,
x0 = Ax, where A is a matrix of real numbers. Suppose one of the eigenvalues  of the

−2 − 5i
coefficient matrix A is r = 3 + i, which has a corresponding eigenvector x0 = 

.

4
What is the system’s real-valued general solution?
   
−2 cos t − 5 sin t 5 cos t − 2 sin t
(a) x(t) = c1 e3t 

 + c2 e3t 
 


4 cos t 4 sin t
   
−2 cos t + 5 sin t −5 cos t − 2 sin t
(b) x(t) = c1 e3t 

 + c2 e3t 
 


4 cos t 4 sin t
   
−2 cos t + 5 sin t −5 cos t − 2 sin t
(c) x(t) = c1 e3t 

 + c2 e3t 
 


4 sin t −4 cos t
   
−2 cos t − 5 sin t 5 cos t − 2 sin t
(d) x(t) = c1 e3t 

 + c2 e3t 
 


4 sin t −4 cos t
 
4 −1
21. True or false. Consider the linear system x0 = 

 x. The critical point (0, 0) is an

−5 0
asymptotically stable node.
 
0 4
22. True or false. Consider the linear system x0 = 

 x. The critical point (0, 0) is an

−4 0
unstable saddle point.
23. For what range of values of β will the linear system below have a saddle point at (0, 0)?
 
−β 1
x0 = 

x

−1 β
(a) (−∞, −1)

(b) (−∞, −1) ∪ (1, ∞)
(c) (−1, 1)
(d) (1, ∞)
24. (a) Find the general solution of the system of linear equations
 
0 −1
x0 = 

 x.

1 2
 
−1
 .
(b) Find the solution satisfying x(0) =  
2
(c) Classify the type and stability of the critical point at (0, 0).
25. For each part below, consider a certain system of two first order linear differential equations
in two unknowns, x0 = Ax, where A is a 2 × 2 matrix of real numbers. Based solely on the
information given in each part, answer each question.
(a) Suppose one of the eigenvalues
 of the
 coefficient matrix A is r = 2 + i, which has
 3 
a corresponding eigenvector 

. Write down the system’s real-valued general

1 − 6i
solution.
(b) Classify the type and stability of the critical point at (0, 0) for the system described in
(a).
(c) Suppose one of the eigenvalues of the coefficient matrix A is r = −9i, classify the type
and stability of the critical point at (0, 0) for the system.
(d) Suppose the coefficient matrix A  only
 has one distinct
 eigenvalue, r = −8, which has
0 −2
corresponding eigenvectors both   and   . Write down the system’s general
   
4 5
solution.
(e) Classify the type and stability of the critical point (0, 0) for the system described in
(d).
Answers
1. a) 5, b) 1, c) 4, d) 2, e) 3
2. a) none, b) 3, c) 4, d) 5, e) 2
3. a) U, b) U, c) A, d) U, e) S
4. 0, 4   
1 5t 0
5. c1   e + c2   e5t
   
0 1
   
 cos 2t   sin 2t 
6. et 

 + 5et 
 
, sprial point, unstable

cos 2t + sin 2t sin 2t − cos 2t
   
 2 cos t  2 sin t
7. a) center, (neutrally) stable, b) x(t) = c1 

 + c2 
 


− sin t cos t
   
1 −t 4
8. a) x(t) = c1   e + c2   e2t , b) β = 6
   
1 1
   
3 −t  1  −4t
9. a) x(t) = 
  e +   e , b) 0
  
3 −2
10. d)
11. α = −1 only
12. a) node, asymptotically stable, b) spiral point, unstable, c) center, (neutrally) stable, d) node,
unstable, e) proper node, asymptotically stable, f) improper node, unstable
13. d)  
 4t − 2  t
14. a) x(t) = 

 e , b) improper node, unstable

−4t + 6
15. a) improper node, unstable, b) node, asymptotically stable, c) spiral point, unstable, d) center,
stable, e) proper node,
 asymptotically
 stable  
 4 cos 7t   4 sin 7t 
16. x(t) = c1 e−5t 

 + c2 e−5t 
 


5 cos 7t − 6 sin 7t 6 cos 7t + 5 sin 7t
     
−2t −7t
−3 −2t −4 −7t  9e − 4e 
17. a) x(t) = c1 
 e
 + c2 
  e , b) x(t) = 
  

1 3 −3e−2t + 3e−7t
18. a) improper node, unstable, b) spiral point, asymptotically stable, c) center, (neutrally) stable,
d) proper node, unstable
19. d)
20. b)
21. False
22. False
23. b)      
−1 t −t + 1 t −tet − et 
24. a) x(t) = c1 
 
 e + c2 

 e , b) x(t) = 
 
, c) improper node, unstable

1 t tet + 2et
   
 3 cos t   3 sin t 
25. a) x(t) = c1 e2t 

 + c2 e2t 
 
, b) spiral, unstable, c) center,

cos t + 6 sin t −6 cos t + sin t
   
0 −8t −2 −8t
(neutrally) stable, d) x(t) = c1 
 e
 + c2 
  e , e) proper node, asymptotically stable

4 5
Chapter 7
Nonlinear Differential Equations

and Stability
7.1 Autonomous Systems

Consider a nonlinear system of two simultaneous differential equations of the form
dx dy
= F (x, y), = G(x, y).
dt dt
Observe that the functions F and G do not depend on the independent variable t but only on the
dependent variables x and y.
• A system in this form is said to be autonomous.
• The critical points are points (x, y) such that x0 = y 0 = 0. (or F (x, y) = G(x, y) = 0)
• Unlike a linear system which has only one critical point, namely (0, 0), a nonlinear system
could have none, one, two, three or any number of critical points.

Find all critical points of the nonlinear system:
x0 = (2x + y)(2x − y)
y 0 = (x − 2)(y + 2)
Solution Critical points can be found by setting x0 = 0 and y 0 = 0. So
x0 = 0 ⇒ (2x + y)(2x − y) = 0 ⇒ 2x + y = 0 or 2x − y = 0.
0
y =0 ⇒ (x − 2)(y + 2) = 0 ⇒ x − 2 = 0 or y + 2 = 0.
Hence we have four pairs of equations to solve:

(
2x + y = 0
⇒ x = 2, y = −2x = −2(2) = −4 ⇒ (2, −4) is a critical point.
x−2=0
(
2x + y = 0 y −2
⇒ y = −2, x = − = =1 ⇒ (1, −2) is a critical point.
y+2=0 2 2
173
CHAPTER 7. NONLINEAR DIFFERENTIAL EQUATIONS AND STABILITY 174
(
2x − y = 0
⇒ x = 2, y = 2x = 2(2) = 4 ⇒ (2, 4) is a critical point.
x−2=0
(
2x − y = 0 y −2
⇒ y = −2, x= = = −1 ⇒ (−1, −2) is a critical point.
y+2=0 2 2
From all 4 cases, (2, −4), (1, −2), (2, 4) and (−1, −2) are all critical points of the given non-
linear system.

Find all critical points of the nonlinear system:
x0 = x2 − xy
y 0 = xy + 2y 2 − 6y.
Solution Note that the nonlinear system can be written as
x0 = x(x − y)
y 0 = y(x + 2y − 6).
To find all critical points, we set x0 = 0 and y 0 = 0. Then
x0 = 0 ⇒ x(x − y) = 0 ⇒ x = 0 or x − y = 0.
0
y =0 ⇒ y(x + 2y − 6) = 0 ⇒ y = 0 or x + 2y − 6 = 0.
Hence we have four pairs of equations to solve:

(
x=0
⇒ (0, 0) is a critical point.
y=0
(
x=0 −x + 6 −0 + 6
⇒ x = 0, y = = =3 ⇒ (0, 3) is a critical point.
x + 2y − 6 = 0 2 2
(
x−y =0
⇒ y = 0, x = y = 0 ⇒ (0, 0) is a critical point.
y=0
(
x−y =0
⇒ x = 2, y = 2 ⇒ (2, 2) is a critical point.
x + 2y − 6 = 0
From all 4 cases, (0, 0), (0, 3) and (2, 2) are all critical points of the above system.
Exercises 7.1
1. Which of the points below is not a critical point of the nonlinear system of equation
x0 = x2 − y 2
y 0 = xy + x + y + 1?
a) (1, 1)
b) (1, −1)
c) (−1, 1)
d) (−1, −1)
2. True or false. The point (0, 1) is a critical point of the nonlinear system
x0 = −x + 2xy
y 0 = y − x2 − y 2 .
3. Which point below is NOT a critical point of the system
x0 = y − y 2
y 0 = x − x3 ?
a) (1, 0)
b) (−1, 1)
c) (0, 1)
d) (1, −1)
4. Find the critical points of the following nonlinear system.
x0 = (x + 1)(y − 2)
y 0 = y(x + y).
5. Find all critical points of the following nonlinear system.

dx
= (1 − y)(2x − y)
dt
dy
= (2 + x)(x − 2y).
dt
6. Find all critical points of the following autonomous nonlinear system.
x0 = x2 − xy
y 0 = 4xy + y 2 − 2y.
Answers
1. a)
2. True
3. d)
4. (−1, 1), (−1, 0) and (−2, 2)
5. (−2, 1), (2, 1), (−2, −4) and (0, 0)
2 2
6. (0, 2), (0, 0) and ( , )
5 5
7.2 Locally Linear Systems

Consider an autonomous nonlinear system,
x0 = F (x, y), y 0 = G(x, y).
Since there might be multiple critical points presenting on the phase portrait, the type and stability
of each critical point need to be determined locally (in a small neighborhood around each point).
We’ll approximate the behavior of the trajectories using the linearizations of F and G about each
critical point. This will convert the globally nonlinear system into a locally linear system around
each critical point. First, using Taylor series about the critical point (x, y) = (α, β),
x0 = F (x, y) ≈ F (α, β) + Fx (α, β)(x − α) + Fy (α, β)(y − β),
y 0 = G(x, y) ≈ G(α, β) + Gx (α, β)(x − α) + Gy (α, β)(y − β).
Since (α, β) is a critical point, F (α, β) = G(α, β) = 0. Then
x0 = F (x, y) ≈ Fx (α, β)(x − α) + Fy (α, β)(y − β)
y 0 = G(x, y) ≈ Gx (α, β)(x − α) + Gy (α, β)(y − β).
Using translation by substitution x − α by x and y − β by y, we have
x0 = F (x, y) ≈ Fx (α, β)x + Fy (α, β)y
y 0 = G(x, y) ≈ Gx (α, β)x + Gy (α, β)y.
Write it in matrix form,      
x0   Fx Fy  x
 =   .
     
y0 Gx Gy y
(α,β)
 
 Fx Fy 
The coefficient matrix 

 is called the Jacobian matrix and denoted by J.

Gx Gy
To consider the type and stability of each critical point of nonlinear system, we first compute
Jacobean matrix, evaluate it at each critical point and then use its eigenvalues to determine their
types and stabilities.
Example 116 (Summer 2002 Final Exam).

Consider the system:
x0 = x + y
y 0 = x2 + y 2 − 8.
a) Find all critical points of the system.

Solution Set x0 = 0 and y 0 = 0. Then
x0 = 0 ⇒ x+y =0
0
y =0 ⇒ x2 + y 2 − 8 = 0.
From the first equation, we have x = −y. Apply this condition into the second equation,
x2 + y 2 − 8 = 0 ⇒ (−y)2 + y 2 − 8 = 0 ⇒ 2y 2 − 8 = 0
⇒ (y − 2)(y + 2) = 0 ⇒ y = 2, −2.
Use x = −y again to find the corresponding values x,
y=2 ⇒ x = −y = −2 ⇒ (−2, 2) is a critical point.

y = −2 ⇒ x = −y = 2 ⇒ (2, −2) is a critical point.
In conclusion, (−2, 2) and (2, −2) are all critical points of the above system.
b) Find the linearized matrix of the system.

Solution We have F (x, y) = x + y, G(x, y) = x2 + y 2 − 8. Then the Jacobean matrix is
   
 Fx Fy   1 1
J =

=
 
.

Gx Gy 2x 2y
At the critical point (−2, 2), the linearized matrix of the system is
   
 1 1  1 1
A1 = J|(−2,2) = 

=
 
.

2(−2) 2(2) −4 4
At the critical point (−2, 2), the linearized matrix of the system is
   
 1 1  1 1
A2 = J|(2,−2) = 

=
 
.

2(2) 2(−2) 4 −4
c) Classify the type and stability of each critical point.

Solution
At (−2, 2), the characteristic equation of A1 is

√
2 2 5 ± 7i
0 = r − (1 + 4)r + 1(4) − (−4)1 = r − 5r + 8 ⇒ r= .
2
Since we get complex eigenvalues with positive real part (case 2.2), (−2, 2) is an unstable
spiral point. While at (2, −2), the characteristic equation of A2 is
√
2 2 −3 ± 41
0 = r − (1 − 4)r + 1(−4) − 4(1) = r + 3r − 8 ⇒ r = .
2
Since we get real distinct eigenvalues with opposite signs (case 1.1), (2, −2) is an un-
stable saddle point.

Consider the nonlinear system:
x0 = x(1 − x + y)
y 0 = y(x + y).
a) Verify that (1, 0) is critical point.

Solution Substitute x = 1 and y = 0,
x0 = x(1 − x + y) = 1(1 − 1 + 0) = 0
y 0 = y(x + y) = 0(1 + 0) = 0.
Therefore (1, 0) is a critical point.

b) Linearize the system about the point (1, 0). Clearly identify the matrix for the linearized
system.
Solution Note F (x, y) = x(1 − x + y) = x − x2 + xy and G(x, y) = y(x + y) = yx + y 2 .
Then we compute Jacobean matrix,
   
 Fx Fy  1 − 2x + y x 
J =

=
 
.

Gx Gy y x + 2y
Now we evaluate the Jacobean matrix at the critical point (1, 0) to get the matrix for
the linearized system,
   
1 − 2(1) + 0 1  −1 1
A = J|(1,0) = 

=
 
.

0 1 + 2(0) 0 1
(This means that near the point (1, 0), the nonlinear system acts like the linear system
 
−1 1
x0 = 

 x.)

0 1
c) Classify the type and stability of the critical point (1, 0) by examining the linearized
system obtained in part b)
Solution Find the eigenvalues from the characteristic equation of A,
0 = r2 − (−1 + 1)r + (−1)1 − 0(0) = r2 − 1 = (r − 1)(r + 1) ⇒ r = 1, −1.
Since we get two distinct real eigenvalues with opposite signs (case 1.1), (1, 0) is an
unstable saddle point.
In this example, we can check further that there are two more critical points (three in total):
• (1, 0) is an unstable saddle point.
• (0, 0). At this critical point, the coefficient matrix has zero determinant.
• ( 12 , − 12 ) is an asymptotically stable spiral point. (case 2.2, we have complex eigenvalues
with negative real part)

Linearize the system
x0 = 1 − y
y 0 = x2 − y 2
around its critical point (1, 1) and classify its type and stability.
Solution Here F (x, y) = 1 − y and G(x, y) = x2 − y 2 . So the Jacobean matrix is,
   
 Fx Fy   0 −1 
J =

=
 
.

Gx Gy 2x −2y
Then the linearized system at (1, 1) has coefficient matrix,

   
 0 −1  0 −1
A = J|(1,1) = 

=
 
.

2(1) −2(1) 2 −2
Now let’s find the eigenvalues from the characteristic equation of A,

√
2 2 −2 ± 4−8
0 = r − (0 − 2)r + 0(−2) − 2(−1) = r + 2r + 2 ⇒ r= = −1 ± i.
2
Since we get complex eigenvalues with negative real part (case 2.2), (1, 1) is an asymptotically
stable spiral point.
(You can check further that another critical point is (−1, 1). And it is an unstable saddle
point since its linearized matrix gives distinct real eigenvalues with opposite signs. (case 1.1))

Linearize the following nonlinear system about its critical point (4, 2) and classify its type and
stability.
x0 = x2 − 4y 2
y 0 = xy − 2x.
Solution We have F (x, y) = x2 − 4y 2 and G(x, y) = xy − 2x. Then the Jacobean matrix is,
   
 Fx Fy   2x −8y 
J =

=
 
.

Gx Gy y−2 x
Then the linearized system at (4, 2) has a coefficient matrix,

   
 2(4) −8(2) 8 −16
A = J|(4,2) = 

=
 
.

2−2 4 0 4
So the characteristic equation of A is
0 = r2 − (8 + 4)r + 8(4) − 0(−16) = r2 − 12r + 32 = (r − 8)(r − 4) ⇒ r = 4, 8.
Since we have a pair of positive real distinct eigenvalues (case 1.2), (4, 2) is an unstable node.
In this example, we can check further that there are two more critical points (three in total):
• (4, 2) is an unstable node.
• (0, 0). At this critical point, the coefficient matrix has vanishing determinant.
• (−4, 2) is an asymptotically stable node. (case 1.2: negative real distinct eigenvalues)
Exercises 7.2
1. (a) Find the critical points of the following nonlinear system
(
x0 = x(x + y)
y 0 = y(2 − x + y)
(b) Linearize the following nonlinear system about its critical point (0, 2) and classify its
type and stability. (
x0 = xy − 6x
?
y 0 = xy − 2x + y − 2
2. Given that the point (0, 1) is a critical point of the nonlinear system of equations
x0 = x2 y − xy 2
y 0 = xy − x − 3y + 3
The critical point (0, 1) is a(n)

(a) unstable saddle point.
(b) unstable spiral point.
(c) asymptotically stable node.
(d) asymptotically stable proper node (star point).
3. Consider the nonlinear system:
x0 = x − y
y 0 = 4y − x2 y.
(a) One of the critical points of the system is (2, 2). Verify that (2, 2) is indeed a critical
point. That is, show that (2, 2) satisfies the condition(s) of being a critical point.
(b) Besides (2, 2), there are 2 other critical points. Find those other 2 critical points of the
system.
(c) Linearize the system about the point (2, 2). Classify the type and stability of the critical
point at (2, 2) by examining the linearized system. Be sure to clearly state the linearized
system’s matrix and its eigenvalues.
x0 = x2 − xy
y 0 = xy − 3x + 2.
(a) One of the critical points of the system is (1, 1). There is another critical point. Find
it.
(b) Linearize the system about the point (1, 1). Classify the type and stability of the critical
point at (1, 1) by examining the linearized system. Be sure to clearly state the linearized
x0 =(x + 1)(y − 2) = xy − 2x + y − 2
y 0 =(x + y)(2x − y) = 2x2 + xy − y 2
(a) The system has 4 critical points. One of the critical points of the system is (−1, 1).
Find the other 3 critical points of the system.
(b) Linearize the system about the critical point (−1, 1). Identify the coefficient matrix of
the linearized system.
(c) What are eigenvalues of the coefficient matrix? Classify the type and stability of the
critical point at (−1, 1) by examining the linearized system found in (b).
6. Consider the autonomous nonlinear system:
x0 =4(x − 1) − y
y 0 =y(2x2 − 4 − y)
(a) The system has 3 critical points. One of the critical points is (1, 0). Verify that (1, 0)
is indeed a critical point. Then find the other 2 critical points.
(b) Linearize the system about the point (1, 0). Classify the type and stability of this
critical point by examining the linearized system. Be sure to clearly state the linearized
7. Given that (−2, 4) is a critical point of the nonlinear system
x0 = xy + 2y
y 0 = xy − 4x.
Linearize this system about (−2, 4) to determine the critical point as a(n)

(c) unstable node.
(d) asymptotically stable proper node.
8. Given that (3, −2) is a critical point of the nonlinear system
x0 = 3y − xy
y 0 = xy + 2x.
Linearize this system about (3, −2) to determine the critical point as a(n)
(c) unstable node.
(d) asymptotically stable improper node.
9. Given that (1, 1) is a critical point of the system
x0 = x2 − y 2
y 0 = 2 − 2xy
Which of the following statement is TRUE regarding (1, 1)?

(a) It is an asymptotically stable spiral point.
(b) It is an unstable node.
(c) It is an asymptotically stable proper node.
(d) It is an unstable saddle point.
Answers
1. a) (0, 0), (0, −2), (1, −1), b) saddle point, unstable
2. c)
3. a) Substitute (2, 2) into the equations directly to see that x0 = y 0 = 0, b) (0, 0) and (−2, −2),
c) saddle point, unstable
4. a) (2, 2), b) saddle point, unstable 
−1 0
5. a) (−1, −2), (−2, 2) and (1, 2), b) 

, c) −1, −3, asymptotically stable node

−3 −3
6. a) (2, 4), (0, −4), b) saddle point, unstable
7. a)
8. c)
9. d)
Chapter 8
Partial Differential Equation
Recall that a partial differential equation is any differential equation that contains two or more
independent variables. Therefore the derivatives appearing in the equation are partial derivative.
In this class, we’ll examine the second order partial differential equation with two independent
variables such as
α2 Uxx =Ut (Heat equation)

2
a Uxx =Utt (Wave equation)
The goal of this chapter is to solve the above two partial differential equations by using the method
of separation of variables.
8.1 Boundary Value Problem

8.1.1 BVP vs IVP
Recall that the initial value problem for second order differential equation is in the form of
y 00 + p(t)y 0 + q(t)y = g(t), y(t0 ) = y0 , y 0 (t0 ) = y00 .
Notice that initial conditions of IVP evaluate at the same point. (in this case at t0 )
If the conditions are specified at two different points, such conditions are called boundary con-
ditions. A differential equation with boundary conditions form a (two-point) boundary value
problem, or simply, BVP. A typical example is the differential equation
y 00 + p(t)y 0 + q(t)y = g(t) (8.1)
with boundary conditions
y(α) = y0 , y(β) = y1 (8.2)

( or y(α) = y0 , y 0 (β) = y1 (8.3)
0 0
or y (α) = y0 , y (β) = y1 ) (8.4)
where α 6= β. If the function g in Equation (8.1) is zero and if the boundary values y0 and y1 in
Equation (8.2)-(8.4) are also zero, then BVP is called homogeneous. Otherwise, the problem is
nonhomogeneous.
184
CHAPTER 8. PARTIAL DIFFERENTIAL EQUATION 185
Example 120. The followings are all BVPs since the boundary conditions evaluate at dif-
ferent points. (at 0 and π)
y 00 + 2y = 0, y(0) = 1, y(π) = 0
00
y + y = 2, y(0) = 0, y 0 (π) = 0
y 00 − 2y = 0, y 0 (0) = 0, y 0 (π) = 0
Although IVP and BVP may superficially appear to be quite similar, their situations differ in
some very important way. Under mild conditions on the coefficients of the differential equation,
IVP are certain to have a unique solution by the Existence and Uniqueness Theorem below. (for
second order ODEs)
Theorem 13 (Existence and Uniqueness Theorem). Consider the initial value problem
y 00 + p(t)y 0 + q(t)y = g(t), y(t0 ) = y0 , y 0 (t0 ) = y00 ,
where p, q and g are continuos on an open interval I that contains the point t0 . Then there is
exactly one solution y = φ(t) of this problem, and the solution exists throughout the interval
I.
On the other hand, BVP under similar conditions may have a unique solution, no solution or even
infinitely many solutions. Let’s take a look in the next example.
Example 121. Consider the solution to y 00 + y = 0 y(0) = 0 y(π) = 0.

Solution The characteristic equation is 0 = r2 + 1 and so r = ±i. So the general solution is
y = c1 cos t + c2 sin t.
Invoke condition y(0) = 0, we have c1 = 0, and y(π) = 0 also gives us c1 = 0. Hence solution
is y(t) = c2 sin t where c2 remains arbitrary. So this BVP has infinitely many solutions.
Example 122 (Fall 2013 Final Exam). Which initial or boundary value problem below is
guaranteed to have a unique solution according to the Existence and Uniqueness Theorems?
√
a) t2 y 00 + ty 0 + e2t y = 2, y(2) = 3, y 0 (2) = 1e .
b) y 00 + sin(t)y 0 + 10ty = ln t, y(−5) = 1, y 0 (−5) = 4.
c) y 0 + tan(t)y = sec t, y( 3π
2 ) = 9.
d) y 00 − t3 y 0 + e−3t y = 0, y 0 (1) = −1, y 0 (π) = 1.

Solution Notice that differential equations in multiple choices a) to c) are IVP while the one
in d) is BVP. So d) is not a correct answer since there is no theorem to guarantee about
existence and uniqueness of solution for BVP.
b) also is not the correct answer since t0 = −5 is not defined for the function ln while in c),
t0 = 3π 1
2 is a point of discontinuity of sec t = cos t .
So a) is the correct answer by the Existence and Uniqueness Theorem since t0 = 2 is not the
discontinuous point. (The only discontinuity for a) is 0.)
8.1.2 Eigenvalues Problems

Next let’s consider a certain BVP
y 00 + λy = 0
together with two boundary conditions, where λ is a constant real number.
• We call it the eigenvalue problem.
• The values of λ for which nontrivial solutions y occur are called eigenvalues.
• The nontrivial solutions (y 6= 0) themselves are called eigenfunctions.
For eigenvalue problems, the following trig identities are helpful:
sin(x) = 0 ⇒ x = nπ for all integer n

(2n − 1)π
cos(x) = 0 ⇒ x= for all integer n
2
Example 123. Find the eigenvalues and eigenfunctions of the given BVP
y 00 + λy = 0, y(0) = 0, y(L) = 0
where L is a positive real number. (L > 0)

Solution We will consider into three cases: λ > 0, λ = 0 and λ < 0.
Case 1: λ > 0. To avoid the radical signs, it’s convenient to let λ = µ2 , where µ > 0. So the
equation becomes
y 00 + µ2 y = 0.
So its characteristic equation is 0 = r2 + µ2 and so r = ±µi. So the general solution is
y(x) = c1 cos µx + c2 sin µx. (8.5)
Invoking the boundary conditions,
y(0) = 0 ⇒ 0 = c1 cos(0) + c2 sin(0) = c1 ⇒ c1 = 0. (8.6)

y(L) = 0 ⇒ 0 = c1 cos(µL) + c2 sin(µL) = c2 sin(µL) ⇒ c2 = 0 or sin(µL) = 0.
(8.7)
In Equation (8.6), if c2 = 0 together with c1 = 0 from Equation (8.7), then the general solution
is a trivial solution y = 0 by Equation (8.5). Since we’re seeking nontrivial solutions (by the
definition of eigenfunctions), so we must require that c2 6= 0. So by Equation (8.7), we get
nπ
sin(µL) = 0 ⇒ µL = nπ ⇒ µn = , for n = 1, 2, . . .
L
The last step we use the fact that L > 0. So in this first case, the eigenvalues are
nπ 2
λn = (µn )2 = ( ) ,
L
and the corresponding eigenfunctions are
nπ
yn (x) = c2 sin µn x = c2 sin x,
L
where n = 1, 2, . . . by Equation (8.5).
Case 2: λ = 0. Then the equation is y 00 = 0 and so y(x) = c1 x + c2 . Now let’s use the
boundary conditions:
y(0) = 0 ⇒ 0 = c1 (0) + c2 = c2 ⇒ c2 = 0.
y(L) = 0 ⇒ 0 = c1 L + c2 = c1 L ⇒ c1 = 0 since L 6= 0.
So there is only the trivial solution y0 = 0 in this case. Hence λ0 = 0 is not an eigenvalue
and there is no eigenfunction.
Case 3: λ < 0. Similar to case 1, let λ = −µ2 , where µ > 0. So the equation becomes
y 00 − µ2 y = 0.
So its characteristic equation is 0 = r2 − µ2 = (r − µ)(r + µ) and so r = ±µ. Thus the general

solution is
y(x) = c1 eµx + c2 e−µx . (8.8)
By using the given boundary conditions, we have
y(0) = 0 ⇒ 0 = c1 + c2 ⇒ c2 = −c1 .
−Lx
y(L) = 0 ⇒ 0 = c1 e Lx
+ c2 e = c1 eLx − c1 e−Lx = c1 (eLx − e−Lx )
⇒ c1 = 0 and hence c2 = 0.
Here we use the fact that eLx − e−Lx 6= 0 for all x since if it is, we get
eLx = e−Lx ⇒ Lx = −Lx ⇒ 2Lx ⇒ x = 0,
which is a contradiction.
For all 3 cases, the eigenvalues are
nπ 2
λn = ( ) ,
L
nπ
yn (x) = sin x,
L
where n = 1, 2, . . .
Example 124 (Fall 2005, Final Exam).

Determine all positive eigenvalues λ and the corresponding eigenfunctions for the following
two point boundary value problem:
y 00 + λy = 0, y 0 (0) = 0, y(3) = 0.
Solution As usual, write λ = µ2 , where µ > 0. So the equation becomes
y 00 + µ2 y = 0.
So its characteristic equation is 0 = r2 + µ2 and so r = ±µi. So the general solution is
y(x) = c1 cos µx + c2 sin µx. (8.9)
Note here that

y 0 (x) = −c1 µ sin µx + c2 µ cos µx.
Invoking the boundary conditions,
y 0 (0) = 0 ⇒ 0 = c2 µ ⇒ c2 = 0 since µ 6= 0.
y(3) = 0 ⇒ 0 = c1 cos(3µ) ⇒ c1 = 0 or cos(3µ) = 0.
Again if both c1 and c2 are zero, then the general solution is a trivial solution y = 0. Since
we’re seeking nontrivial solutions, we need that c1 6= 0. So we get
(2n − 1)π (2n − 1)π

cos(3µ) = 0 ⇒ 3µ = ⇒ µn = , for n = 1, 2, . . . .
2 6
Therefore the eigenvalues are
(2n − 1)π 2
λn = (µn )2 = ( ) ,
6
(2n − 1)π
yn (x) = c1 cos µn x = c1 cos x,
6
where n = 1, 2, . . . by Equation (8.9).
Exercises 8.1
1. Which of the following has a unique solution on the whole interval (0, π)?
(a) y 00 + y = 0, y(0) = 0, y(π) = 0.
00 0
(b) y + 4y = 0, y (0) = 0, y 0 (π) = 0.
(c) (t + 1)y 00 + ty = 0, y(1) = 1, y 0 (1) = 0.
(d) (t − 1)y 0 + 2y = 0, y(0) = 0, y 0 (0) = 1.
2. Of the four initial or boundary value problems below, only one is guaranteed to have a unique
solution according to the Existence and Uniqueness Theorems. Which one is it?
(a) ty 00 − t2 y 0 + et y = tan(t), y(1) = 0, y 0 (1) = π.
00 0 0
(b) ty + 2y − 3y = 0, y(0) = 2, y (0) = 0.
00
(c) y + 9y = 0, y(0) = 0, y(3π) = 0.
0
(d) y + sec(t)y = sin(2t), y( π2 ) = 0.
3. Consider the two initial/ boundary value problems below. Which is certain to have a unique
solution for every value of α?
(I) y 00 + 9y = 0, y(α) = α2 , y 0 (α) = −α.

(II) y 00 + 9y = 0, y(0) = 0, y 0 (α2 ) = 0.
(a) I only.
(b) II only.
(c) Both I and II.
(d) Neither.
4. Consider the second order equation
y 00 + λy = 0,
Where λ is a real number. Which statement below about the equation is false?
(a) For any value of λ, its general solution is in the form of y = C1 y1 + C2 y2 , where y1 and
y2 are two fundamental solutions of the equation.
(b) If λ = 0, then its general solution is y(t) = C1 + C2 t.
(c) For any value of λ, there is a unique solution satisfying boundary conditions y(0) = 0
and y 0 (2π) = 0.
(d) For any value of λ, there is a unique solution satisfying initial conditions y(π) = 0 and
y 0 (π) = 32.
5. Consider the following two differential equations:
(I) y 00 + ay 0 + by = 0, y(0) = 0, y 0 (0) = 2.

(II) y 00 + ay 0 + by = 0, y(0) = 0, y(π) = 2.
where a, b are real numbers. Which of the following statements are true?
(a) Both I and II always have a unique solution on some interval.
(b) Only I always has a solution on some interval.
(c) Only II always has a solution on some interval.
(d) None of the above

6. Consider the two-point boundary value problem
X 00 + λX = 0, X(0) = 0, X 0 (3) = 0
(a) Find all positive eigenvalues and corresponding eigenfunctions of the boundary value
problem.
(b) Is λ = 0 an eigenvalue of this problem? If yes, find its corresponding eigenfunction. If
no, briefly explain why it is not an eigenvalue.
7. Find all eigenvalues and corresponding eigenfunctions of the boundary value problem
X 00 + λX = 0, X 0 (0) = 0, X 0 (π) = 0.
Make sure to consider, and show your work for, all three possibilities: λ < 0, λ = 0, and
λ > 0.
8. Consider the 2-point boundary value problem:
X 00 + λπ 2 X = 0, X(0) = 0, X 0 (1) = 0.
1
Is λ = an eigenvalue? If not, justify your answer. If so, find a corresponding eigenfunction.
4
9. (a) For what positive values of λ does the boundary value problem
X 00 + λX = 0, X 0 (0) = 0, X(π) = 0
have a nontrivial solution?

(b) What are the corresponding eigenfunctions?
10. Find the eigenvalues and eigenfunctions of the boundary value problem
π
X 00 + λX = 0, X(0) = 0, X( ) = 0.
2
Show your work in all three cases: λ < 0, λ = 0, and λ > 0.
Answers
1. c)
2. a)
3. a)
4. c)
5. b)
2 2
6. a) The positive eigenvalues are λn = (2n−1)36
π
, n = 1, 2, 3, . . . ;
The corresponding eigenfunctions are Xn (x) = sin( 2n−1 6 πx) b) No.
7. Eigenvalues are λ = 0 and λ = n2 , n = 1, 2, 3, . . .;
Eigenfunctions are X0 (x) = 1, and Xn (x) = cos(nx), n = 1, 2, 3, . . .
π
8. Yes, it is. Eigenfunction is sin x.
2
2
9. Eigenvalues are λn = (2n−1)
4 and the eigenfunctions are Xn (x) = cos( (2n−1)x
2 ), for all n ∈ N.
2
10. Eigenvalues are λn = 4n , n = 1, 2, 3, . . . and the corresponding eigenfunctions are Xn (x) =
sin(2nx), n = 1, 2, 3, . . .
8.2 Fourier Series

A function f is said to be periodic with period T > 0 if
f (x + T ) = f (x)
for every value of x.
• Notice here that a period of T is not unique, because if T is a period of f , then also 2T and
so indeed is any integral multiple of f .
• We call the smallest value of T such that f (x + T ) = f (x), the fundamental peiod of f .
• The fundamental period does not necessarily exist by if it does, it is unique.
• Moreover the sum of any finite terms or even the sum of a convergent infinite series of
functions of period T is also periodic with period T .
Example 125.
• sin x, cos x are periodic functions with fundamental period 2π.

• Constant function is a periodic function with an arbitrary period but no fundamental
period.
A function f is said to be piecewise continuous on an interval a ≤ x ≤ b if the interval can be

partitioned by a finite number of point a = x0 < x1 < . . . < xn = b so that f is continuous on
each open subinterval xi−1 < x < xi for all i.
8.2.1 The Fourier Convergence Theorem

Next let’s consider the most important theorem in this chapter.
Theorem 14 (The Fourier Convergence Theorem).

Suppose that f and f 0 are piecewise continuous on the interval −L ≤ x < L. Further suppose
that f is defined outside the interval −L ≤ x < L so that it is periodic with period 2L, i.e.,
f (x + 2L) = f (x). Then f has a Fourier series, denoted by F ,
∞
a0 X mπx mπx
F (x) = + (am cos + bm sin ),
2 m=1
L L
whose coefficients are given by Euler-Fourier Formulas,

Z L
1
a0 = f (x) dx,
L −L
Z L
1 mπx
am = f (x) cos dx, m = 1, 2, . . . ,
L −L L
Z L
1 mπx
bm = f (x) sin dx, m = 1, 2, . . . .
L −L L
f (x+ )+f (x− )

The Fourier series converges to f (x) at all points where f is continuous and to 2
at all points where f is discontinuous. In other words,

f (x0 ), if f is continuous at x0
F (x0 ) = f (x+ ) + f (x− ) .
0 0
 , if f is discontinuous at x0
2
Recall that
f (x−
0 ) = lim f (x), f (x+
0 ) = lim f (x)
x→x0 − x→x0 +
is the left-hand (right-hand) limit of f at x0 . Let’s consider the following example to recall
ourselves about it.

−x − 2
 if x < 0
Example 126. Let f (x) = −1 if x = 0

 2
x +1 if x > 0.
Then we have
f (0) = −1, f (0− ) = −2 and f (0+ ) = 1.
Example 127 (Spring 2004, Final Exam). Let


−2 − x, if − 2 < x < −1

f (x) = x, if − 1 ≤ x ≤ 1 and f (x + 4) = f (x).

2 − x, if 1 < x < 2.

Set up, but do not integrate, the integral(s) to find the Fourier coefficients of this function.
Solution: With L = 2, for m = 0, 1, 2, . . ., we have
1 2
Z
mπx
am = f (x) cos dx
2 −2 2
Z −1 Z 1 Z 2
1 mπx mπx mπx
= (−2 − x) cos dx + x cos dx + (2 − x) cos dx ,
2 −2 2 −1 2 1 2
and, for m = 1, 2, . . . ,
1 2
Z
mπx
bm = f (x) sin dx
2 −2 2
Z −1 Z 1 Z 2
1 mπx mπx mπx
= (−2 − x) sin dx + x sin dx + (2 − x) sin dx .
2 −2 2 −1 2 1 2

Consider the Fourier series of function given below:
(
x2 + 1, if − 2 < x < 21
f (x) = and f (x + 4) = f (x).
sin 2πx, if 12 ≤ x < 2
To what value does the Fourier series converges to at x = 12 .

1
Solution Notice that f is discontinuous at x = 2. By the Fourier convergence theorem,
Fourier series of f converges to
+ −
f ( 12 ) + f ( 21 ) sin 2π( 12 ) + ( 12 )2 + 1 5

sin π + 4 5
= = = ,
2 2 2 8
at x = 12 .
The following trig identities are very helpful in this topic:
sin(mπ) = 0 for all integer m

m
cos(mπ) = (−1) for all integer m
Example 129 (Spring 2005, Final Exam).

Consider the function 
0
 if − 2 ≤ x < 0
f (x) = 3 if 0 ≤ x < 1

0 if 1 ≤ x ≤ 2.

a) Sketch the graph of f for three periods.
b) Find the Fourier series of f (x) on [−2, 2].

Solution Note that L = 2. Then

x=1
1 2 1 1
Z Z
3 3
a0 = f (x) dx = 3 dx = (x) = .
2 −2 2 0 2 x=0 2
And for any positive integer m, we have
1 2 1 1 3 1
Z Z Z
mπx mπx mπx
am = f (x) cos dx = 3 cos dx = cos dx
2 −2 2 2 0 2 2 0 2
x=1 x=1
3 2 mπx 3 mπx 3 mπ
= sin = sin = sin .
2 mπ 2
x=0 mπ 2
x=0 mπ 2
Also,
2
1 1 3 1
Z Z Z
1 mπx mπx mπx
bm = f (x) sin
dx = 3 sin dx = sin dx
2 −2 2 2 0 2 2 0 2
x=1 x=1
3 2 mπx 3 mπx
= − cos =− cos
2 mπ 2
x=0 mπ 2
x=0
3 mπ 3 mπ
=− cos −1 = 1 − cos .
mπ 2 mπ 2
So the Fourier series is
∞
3 X 3 mπ mπx 3 mπ mπx
+ sin cos + 1 − cos sin .
4 m=1 mπ 2 2 mπ 2 2
c) Find value that the Fourier series in a) converge at x = 9.
Solution Since f is a periodic function of period 4, so is its Fourier series. So the the
value that the Fourier series converge at x = 9 is the same one at x = 1. Since at x = 1,
f is discontinuous. By the Fourier convergence theorem, its Fourier series converge to
f (1+ ) + f (1− ) 0+3 3

= = .
2 2 2
d) Sketch the graph of the function of which the series converges for three periods.

Let f be the periodic function with period 2π such that
(
2 if − π ≤ x < 0
f (x) =
−2 if 0 ≤ x < π.
a) Sketch the graph of the given function of three periods.
b) Find the Fourier series of the function f .
Solution We have L = π. Then

1 π
Z Z 0 Z π
1
a0 = f (x) dx = 2 dx + −2 dx
π −π π −π 0
1 0 π 1
= 2x −π + (−2x)0 = [(0 − (−2π)) + ((−2π) − 0)] = 0
π π
and for any positive inter m, we have
1 π
Z Z 0 Z π
mπx 1
am = f (x) cos dx = 2 cos mx dx + −2 cos mx dx
π −π π π −π 0

1 2 0 −2 π
= sin mx−π + ( sin mx)0 = 0,
π m m
and also,
1 π
Z Z 0 Z π
mπx 1
bm = f (x) sin dx = 2 sin mx dx + −2 sin mx dx
π −π π π −π 0

1 2 0 2 π
= − cos mx−π + cos mx0
π m m

1 2 2 2 2 1 4 4
= − − (− cos(−mπ)) + cos mπ − = − + cos(mπ)
π m m m m π m m

0 if m is even
1 4 4
= − + (−1)m = 8
π m m − if m is odd.
mπ
Therefore the Fourier series of f is

∞
X 1 4 4
f (x) = − + (−1)m sin mx
m=1
π m m
∞ ∞
X 8 X 8
= − sin mx = − sin(2m − 1)x.
m=1
mπ m=1
(2m − 1)π
m odd
c) To what values does the Fourier series in a) converges at x = π?
Solution At x = π, f is discontinuous. By the Fourier convergence theorem, its Fourier

series converge to
f (π + ) + f (π − ) 2 + (−2)
= = 0,
2 2
at x = π.
d) Sketch the graph of the function of which the series converges for three periods.
Just because a Fourier series could have infinitely many terms does not mean that it will always
have that many terms. If a periodic function f can be expressed by finitely many terms normally
found in a Fourier series, then f must be the Fourier series of itself.
Example 131.
1. The Fourier series (period 2π) of f (x) = 5 + cos 4x − sin 5x is f itself.

2. Let f be the periodic function 6 cos x sin x with period 2π. Even though the product
cos x sin x is not a term in a Fourier series, we can use the double angle formula to get
6 cos x sin x = 3 sin 2x.
So the Fourier series is f itself.
Example 132 (Fall 2005 Final Exam).

Which of the following already has the form of a Fourier series on the interval [−2, 2]?
π 1
f (x) = 4 sin x, g(x) = + 2 cos 3πx.
3 4
Solution Here L = 2. Note that the angle of trig functions in the Fourier series is
mπx mπx
= .
L 2
πx
For function f , the angle is . If it is already has the form of a Fourier series, then
3
πx mπx 2
= ⇒ m= ,
3 2 3
which is not an integer. So f is not in the form of Fourier series yet. On the other hand, for
g, we have
mπx
3πx = ⇒ m = 6.
2
So g is already in the form of Fourier series.
8.2.2 Even and Odd Functions and their Fourier Series

Even Function vs Odd Function
• An even function is any function f such that
f (x) = f (−x),
for all x in its domain. Common examples of even functions are
cos x, |x|, x2 , x4 , . . . , x−2 , x−4 , . . . (or any xm where m is even integer)
• An odd function is any function f such that

f (x) = −f (−x),
for all x in its domain. Common examples of odd functions are
sin x, x, x3 , . . . , x−1 , x−3 , . . . (or any xm where m is odd integer)
• Unlike integer which is either even or odd, most functions are neither even nor odd. For
instance,
f (x) = ex .
• Also there is only one function which is both even and odd, namely zero function. This is
the case since, if f is both even and odd, then
f (x) = f (−x) = −f (x) ⇒ 2f (x) = 0 ⇒ f (x) = 0.
The first equality holds since f is an even and the second one is true since f is an odd
function.
Properties of Even and Odd Functions
• Elementary properties of even and odd functions are displayed in the following table:
Functions +/− ×/÷
Even & Even Even Even
Odd & Odd Odd Even
Even & Odd Neither Odd

• The integral properties of even and odd functions:
– If f is an even function, then

Z L Z L
f (x) dx = 2 f (x) dx.
−L 0
– If f is an odd function, then

Z L
f (x) dx = 0.
−L
• By using this property, we are able to evaluate the integration faster, for examples,
Z 2
x3 dx = 0
−2
Z 8 Z 8
6
x dx = 2 x6 dx.
−8 0
since x3 is an odd function while x6 is an even function.

Fourier Series of Even and Odd Functions
• Now we will apply what we learn about even and odd functions to help us find Fourier series
faster. Suppose now that f is an even periodic function with period 2L. Then, by using the
above table, we have
mπx
f (x) cos is even and
L
mπx
f (x) sin is odd.
L
So
Z L Z L
1 mπx 2 mπx
am = f (x) cos = f (x) cos dx,
L −L L L 0 L
ZL
1 mπx
bm = f (x) sin = 0,
L −L L
for all m. Hence the Fourier series of f is
∞
a0 X mπx
+ am cos .
2 m=1
L
Since cosine is the only trig function that appears in the Fourier series, we call it the Fourier
cosine series.
• Now let f be an odd periodic function with period 2L. Then, by using the above table, we
have
mπx
f (x) cos is odd and
L
mπx
f (x) sin is even.
L
Hence
Z L
1 mπx
am = f (x) cos = 0,
L −L L
Z L Z L
1 mπx 2 mπx
bm = f (x) sin = f (x) sin dx,
L −L L L 0 L
for all m. Hence the Fourier series of f is

∞
X mπx
bm sin .
m=1
L
We call this Fourier series the Fourier sine series.

We summarize what we have done here:
The Fourier Cosine Series: If f is an even periodic function with period 2L, then
∞
a0 X mπx
+ am cos ,
2 m=1
L
is its Fourier cosine series where
2 L L
Z Z
2 mπx
a0 = f (x) dx, am = f (x) cos dx, m = 1, 2, . . .
L 0 L 0 L
The Fourier Sine Series: If f is an odd periodic function with period 2L, then
∞
X mπx
bm sin ,
m=1
L
is its Fourier sine series where

Z L
2 mπx
bm = f (x) sin dx, m = 1, 2, . . .
L 0 L
Example 133 (Fall 2003 Final Exam).

Find the Fourier series for the following function
f (x) = π 2 − x2 , −π ≤ x ≤ π, f (x + 2π) = f (x).
Solution Here L = π. Note here that f is an even function because
f (−x) = π 2 − (−x)2 = π 2 − x2 = f (x),
for all x. By the property of an even function, we have bm = 0 for all m. Also
x=π
2 π 2 x3 π3
Z
2 2 2 2 4
a0 = (π − x ) dx = (π x − ) = (π 3 − ) = π2 ,
π 0 π 3 x=0 π 3 3
and for m = 1, 2, . . . , by using integration-by-parts,
2 π 2
Z
am = (π − x2 ) cos mx dx
π 0
x=π
2 sin mx cos mx sin mx
= (π 2 − x2 ) − 2x + 2
π m m2 m3
x=0
2 cos mπ 4 4 4
= (−2π 2
) = − 2 cos mπ = − 2 (−1) = (−1)m+1 2 .
m
π m m m m
So the Fourier cosine series of f is

∞
2 2 X 4
π + (−1)m+1 2 cos mx.
3 m=1
m
Example 134 (Spring 2007 Final Exam). Consider the following periodic functions
a(x) = x, − 2 ≤ x < 2, a(x + 4) = a(x)

2
b(x) = x , − 1 ≤ x < 1, b(x + 2) = b(x)
c(x) = 1 + cos x, − π ≤ x < π, c(x + 2π) = c(x)
d(x) = 1 + sin x, − π ≤ x < π, d(x + 2π) = d(x)
3
e(x) = x − x , − 2 ≤ x < 2, e(x + 4) = e(x).
a) List all functions that can be expressed as a Fourier sine series.
Solution a is clearly an odd function. So is e since
e(−x) = (−x) − (−x)3 = −x − (−x3 ) = −(x − x3 ) = −e(x),
for all x. Therefore both a and e can be expressed as a Fourier sine series.
b) List all functions that can be expressed as a Fourier cosine series.
Solution b is obviously an even function. So is c since
c(−x) = 1 + cos(−x) = 1 + cos x = c(x),
for all x. Then both b and c can be expressed as Fourier cosine series.
Note here that d is neither even nor odd function, since
d(−x) = 1 + sin(−x) = 1 − sin x
is neither d(x) = 1 + sin x nor −d(x) = −1 − sin x.

8.2.3 The Cosine and Sine Series Extensions

If f and f 0 are piecewise continuous functions defined on the interval 0 ≤ t ≤ L then f can be
expressed into an even(odd) periodic function, f˜, of period 2L, such that f (x) = f˜(x) on [0, L]
and whose Fourier series is therefore a cosine(sine) Fourier series, respectively.
Even (cosine series) extension of f (x)

Let f be defined on [0, L]. Then its even extension of period 2L is
(
˜ f (x) if 0 ≤ x ≤ L
f (x) = , f˜(x + 2π) = f˜(x),
f (−x) if − L ≤ x < 0
where the Fourier cosine series is

∞
a0 X mπx
+ am cos ,
2 m=1
L
and Z L Z L
2 2 mπx
a0 = f (x) dx, am = f (x) cos dx, m = 1, 2, . . .
L 0 L 0 L
Odd (sine series) extension of f (x)

If f is defined on (0, L). Then its odd extension of period 2L is

f (x)
 if 0 < x < L
˜
f (x) = 0 if x = 0, L , f˜(x + 2π) = f˜(x),

−f (−x) if − L < x < 0

where the Fourier sine series is

∞
X mπx
bm sin ,
m=1
L
and Z L
2 mπx
bm = f (x) sin dx, m = 1, 2, . . .
L 0 L
Example 135 (Fall 2001 Final Exam). Consider the following function, define for 0 ≤ x ≤ 3:
(
−1 , 0 ≤ x ≤ 2
f (x) = .
2 ,2 < x ≤ 3
a) Graph the even extension of f (x) on the interval −3 ≤ x ≤ 3.
b) Graph the odd extension of f (x) on the interval −3 ≤ x ≤ 3.
Example 136 (Summer 2003 Final Exam). Let

(
1 − x ,0 < x < 2
f (x) = .
−1 ,2 ≤ x < 3
a) Sketch the even, period 6 extension of f (x) on the interval [−9, 9].
b) Sketch the odd, period 6 extension of f (x) on the interval [−9, 9].
c) Set up, but do not evaluate, the integral(s) that will give the Fourier sine coefficients of
the odd extension.
Solution We know that for Fourier sine series, a0 and all am are zero. For m =
1, 2, . . .,
2 3
Z
mπx
bm = f (x) sin dx
3 0 3
Z 2 Z 3
2 mπx mπx
= (1 − x) sin dx + (−1) sin dx .
3 0 3 2 3
d) To what value does the sine series above converge to at x = −3, x = 0 and x = 2?
Solution f is discontinuous at x = −3 and x = 0 and continuous at x = 2. So by
Fourier Convergence Theorem, we have
f (−3+ ) + f (−3− ) 1 + (−1)

F (−3) = = =0
2 2
+ −
f (0 ) + f (0 1 + (−1)
F (0) = = =0
2 2
F (2) = f (2) = −1.
Example 137 (Spring 2003 Final Exam). Consider the following function, defined on the
interval (0, 2). (
x ,0 < x < 1
f (x) = .
2 ,1 ≤ x < 2
a) Graph the even, period 4, extension of f (x) on [−4, 4].
b) Graph the odd, period 4, extension of f (x) on [−4, 4].
c) Which of the above two has a cosine series, and which has a sine series?
Solution Part a) has a cosine series because it is an even function while Part b) gives
a sine series since it is an odd one.
d) What does the Fourier series representing part b) (the odd extension) converge to at
x = −2, x = 12 , and x = 3?
Solution Since f is a periodic function with period 4, so is its Fourier series. So the
value that Fourier series converges at x = 3 is the same one the Fourier series converges
to at x = −1. f is discontinuous at x = −2 and x = −1 and continuous at x = 21 . So
by Fourier Convergence Theorem, we have
f (−2+ ) + f (−2− ) −2 + 2
F (−2) = = =0
2 2
1 1 1
F( ) = f( ) =
2 2 2
f (−1+ ) + f (−1− ) −1 + (−2) 3
F (3) = F (−1) = = =− .
2 2 2
Exercises 8.2
1. True or false:
(a) The constant term of the Fourier series representing f (x) = x2 , −2 < x < 2, f (x + 4) =
a0 4
f (x), is = .
2 3
(b) The Fourier series (of period 2π) representing f (x) = 3 − 7 sin2 (x) is a Fourier sine
series.
(c) The Fourier series of
f (x) = 3x2 − 4 cos 2x, −π < x < π, f (x + 2π) = f (x)
is a cosine series.
(d) Every Fourier sine series converges to 0 at x = 0.
(e) Every Fourier sine series has 0 as its constant term.
(f) The constant term of the Fourier series representing
f (x) = 2x3 , −1 < x < 1, f (x + 2) = f (x),
is zero.
2. Consider the Fourier series (of period 2π) representing
f (x) = 13 − 5 cos(4x) + 2 sin2 (6x).
Which statement below is true? (Hint: Is f (x) an even or an odd function?)

(a) The Fourier series is a cosine series.
(b) The Fourier series is a sine series.
(c) The Fourier series is neither a cosine series nor a sine series.
(d) The function does not have a Fourier series because it is not periodic.
3. Consider the Fourier series (of period 20) representing
f (x) = x5 , 10 < x < 10, f (x + 20) = f (x).
Which statement below is true?

4. Consider the Fourier series (of period 6π) representing
f (x) = 5x3 + 7, −3π < x < 3π, f (x + 6π) = f (x).

5. Find the Fourier cosine coefficient corresponding to n = 2, a2 , of the Fourier series of period
T = 2π representing the function f (x) = 6 cos 2x.
(a) a2 = 0
(b) a2 = 6
3
(c) a2 =
π
1
(d) a2 = −
π
6. Consider the 2π-periodic function
f (x) = |x|, when − π < x < π, and f (x + 2π) = f (x).
Find the fourth sine coefficient b4 of the Fourier series.

1
(a) 2.
(b) − 41 .
(c) 0.
(d) − 12 .
7. Find the Fourier sine coefficient corresponding to n = 6, b6 , of the Fourier series of period
T = 4π representing the function
(
1, −2π < x < 0
f (x) = , f (x + 4π) = f (x).
3, 0 < x < 2π
(a) b6 = 0.
4
(b) b6 = .
5π
3
(c) b6 = .
2π
3
(d) b6 = − .
4π
8. Let (
2, 0 < x < 1
f (x) =
1, 1 ≤ x ≤ 2
(a) Sketch the odd, period 4 extension of f (x) on the interval −6 ≤ x ≤ 6.

(b) Sketch the even, period 4 extension of f (x) on the interval −6 ≤ x ≤ 6.
(c) Find the Fourier series of the even extension.
(d) What does the series above converge to when x = −2, x = 0 and x = 1?
9. Let f (x) = 2 + sin x, 0 < x < π.
(a) Consider the odd periodic extension, of period T = 2π, of f (x). Sketch 3 periods, on
the interval −3π < x < 3π, of this odd periodic extension.
(b) Which of the integrals below can be used to find the Fourier sine coefficients of the odd
periodic extension in a)?
Z π
1
i. (2 + sin x) sin nx dx
2π −π
Z π
2
ii. (2 + sin x) sin nx dx
π 0
Z π
1 nx
iii. (2 + sin x) sin dx
2π 0 2
Z π
2 nx
iv. (2 + sin x) sin dx
π −π 2
(c) To what value does the Fourier series of this odd periodic extension converge at x = −π?
3π
At x = ?
2
(d) Consider the even periodic extension, of period T = 2π of f (x). Sketch 3 periods, on
the interval −3π < x < 3π, of this even periodic extension.
(e) To what value does the Fourier series of this even periodic extension converge at x = π?
10. Let f (x) = 2 − x, 0 < x < 2.
(a) Consider the odd periodic extension, of period T = 4, of f (x). Sketch 3 periods, on the
interval −6 < x < 6, of this odd periodic extension.
(b) True of False: The Fourier sine coefficients of the odd periodic extension in (a) can be
found by Z 0 Z 2
1 nπx nπx
bn = (−2 − x) sin dx + (2 − x) sin dx
2 −2 2 0 2
(c) To what value does the Fourier series of this odd periodic extension converge at x = −1?
At x = 4?
(d) Consider the even periodic extension, of period T = 4 of f (x). Sketch 3 periods, on the
interval −6 < x < 6, of this even periodic extension.
(e) Find the constant term of the Fourier series of the periodic function described in (d).
(f) To what value does the Fourier series of this even periodic extension converge at x = 4?
Answers
1. a) T, b) F, c) T, d) T, e) T, f) T
2. a)
3. b)
4. c)
5. b)
6. c)
7. a)
∞
3 X 2 mπ mπx
8. c) F (x) = + sin cos ,
2 m=1 mπ 2 2
3
d) At x = −2, it converges to 1. At x = 0, it converges to 2. At x = 1, it converges to 2
9. b) ii), c) At x = −π, it converges to 0. At x = 23 π, it converges to −3, e) 2.
10. b) T, c) −1 and 0, e) 1, f) 2
8.3 Heat Equations & Separation of Variables

Consider a straight bar of uniform cross section and homogeneous material. Let the x-axis be
chosen to lie along the axis of the bar, and let x = 0 and x = L denote the ends of the bar. We
also assume that the cross sectional dimension are so small that the temperature can be considered
constant on any given crosse section. Then the temperature distribution U is a function only of
the axial coordinate x and the time t.
Figure 8.1: A heat-conducting solid bar.
The equation describing this function U (x, t) is
α2 Uxx = Ut ,
where α2 is a constant known as the thermal diffusivity. This equation is called heat conduc-
tion equation or heat equation. The constant α2 depends only on material from which bar is
made. Typical values of α2 are given in the table below.
Material α2 (cm2 /s)
Silver 1.71
Copper 1.14
Aluminum 0.86
Cast iron 0.12
Granite 0.011
Brick 0.0038
Water 0.00144
Table 8.1: Values of the Thermal Diffusivity for Some Common Materials
8.3.1 Type I: Homogeneous Boundary Conditions & Separation of Vari-

ables
In addition, we assume that the initial temperature distribution in the bar is given
U (x, 0) = f (x), 0 < x < L.
Assume that the ends of the bar are held at zero temperature at all time, so
U (0, t) = 0, U (L, t) = 0, t > 0.
So the following is the problem we would like to solve
Heat Problem Type I: Homogeneous Boundary Conditions

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t > 0
(Initial condition) U (x, 0) = f (x), 0 < x < L,
(Boundary conditions) U (0, t) = 0, U (L, t) = 0, t > 0.
We call this type of problems which contain both initial condition(s) and boundary conditions an
initial-boundary value problem or IBVP. Alternatively, we can consider this problem in the
xt-plane as shown below. The solution u(x, t) is sought in the semi-infinite strip 0 < x < L, t > 0,
subject to the requirement that u(x, t) must assume a prescribed value at each point on the
boundary of this strip.
Figure 8.2: Boundary value problem for the heat conduction equation.
The method we’ll use to solve this heat equation is called the separation of variables. Assume
that U (x, t) can be written as a product of two functions, one depending only on x and the other
depending only on t. So we write
U (x, t) = X(x)T (t).
Example 138 (Spring 2010 Final Exam). True or false.
a) It is possible to separate the PDE,
x3 Utt − t3 Uxx = 0,
into two ODEs.
Solution Let U (x, t) = X(x)T (t). Then Utt = XT 00 and Uxx = X 00 T . Then we have
x3 Utt − t3 Uxx = 0 ⇒ x3 XT 00 − t3 X 00 T = 0
⇒ x3 XT 00 = t3 X 00 T
T 00 X 00
⇒ 3
= 3
t T x X
So the statement is true since the given PDE is separable.
b) It is possible to separate the PDE,
Uxx − 2Utx + Utt = 0,
into two ODEs.
Solution Let U (x, t) = X(x)T (t). Then Uxx = X 00 T, Utx = X 0 T 0 and Utt = XT 00 . So
we have
Uxx − 2Utx + Utt = 0 ⇒ X 00 T − 2X 0 T 0 + XT 00 = 0.
Since there is no common term in each pair of the expression, we can’t separate this
PDE. So the argument is false.
Example 139 (Spring 2009 Final Exam). True or false. It is possible to separate the PDE,
Uxx − 2Utt = 3Ut ,
into two ODEs.

Solution Let U (x, t) = X(x)T (t). Then Uxx = X 00 T , Utt = XT 00 and Ut = XT 0 . Then we
have
Uxx − 2Utt = 3Ut ⇒ X 00 T − 2XT 00 = 3XT 0

⇒ X 00 T = 3XT 0 + 2XT 00
⇒ X 00 T = (3T 0 + 2T 00 )X
X 00 3T 0 + 2T 00
⇒ = .
X T
So the statement is true since the given PDE is separable.
Example 140 (Spring 2004 Final Exam).

Separate the PDE,
Uxx − 2Utx + 5x3 Ut = 0
into a system of two equations of one independent variable each
Solution Again using the method of separation of variables by writing U (x, t) = X(x)T (t).
Then Uxx = X 00 T, Utx = X 0 T 0 and Ut = XT 0 . Then we have
Uxx − 2Utx + 5x3 Ut = 0 ⇒ X 00 T − 2X 0 T 0 + 5x3 XT 0 = 0

⇒ X 00 T = 2X 0 T 0 − 5x3 XT 0
⇒ X 00 T = (2X 0 − 5x3 X)T 0
X 00 T0
⇒ = = −λ
2X 0 − 5x3 X T
So we get
X 00
= −λ ⇒ X 00 + 2λX 0 − 5x3 λx = 0
2X 0 − 5x3 X
T0
= −λ ⇒ T 0 + λT = 0,
T
as two equations of one independent variable.
Now let’s comeback to our Heat problem. Write U (x, t) = X(x)T (t) by the separation of variables
method. From this we have Uxx = X 00 T, Ut = XT 0 . Substituting these into the heat equation,
X 00 T0
α2 Uxx = Ut ⇒ αX 00 T = XT 0 ⇒ = 2 = −λ.
X α T
We put negative sign here because it makes our life easier later. Next step is to consider them
separately,
X 00
= −λ ⇒ X 00 + λX = 0.
X
0
T
= −λ ⇒ T 0 + α2 λT = 0.
α2 T
Now let’s take a look at the boundary conditions. By using U (x, t) = X(x)T (t), we have
U (0, t) = 0 ⇒ X(0)T (t) = 0 ⇒ X(0) = 0 or T (t) = 0

U (L, t) = 0 ⇒ X(L)T (t) = 0 ⇒ X(L) = 0 or T (t) = 0
If T (t) = 0, both conditions would be satisfied but U (x, t) = X(x)T (t) = 0 would be a trivial
function which is totally uninteresting solution. Hence we have to let the new boundary conditions
to be
X(0) = 0, X(L) = 0.
For now, after using heat equation and boundary conditions, we transform the problem to
X 00 + λX = 0, X(0) = 0, X(L) = 0 (8.10)

0 2
T + α λT = 0. (8.11)
Recall that we solved BVP (8.10) before in Example 123, the eigenvalues are
nπ 2
λn = ( ) ,
L

nπx
Xn (x) = sin ,
L
where n = 1, 2, . . ..
nπ
Now substitute λn = ( )2 into Equation (8.11),
L
α 2 n2 π 2 α 2 n2 π 2 t
T 0 + α2 λT = 0 ⇒ T0 + T =0 ⇒ Tn (t) = exp(− ), n = 1, 2, . . .
L2 L2
Thus, for each n = 1, 2, . . .,
α 2 n2 π 2 t nπx
Un (x, t) = Xn (x)Tn (t) = exp(− ) sin .
L2 L
Then functions {Un } are sometimes called fundamental solutions of the heat conduction prob-
lem. Then the linear combination of Un for the general solution of the heat equation,
∞ ∞
X X α 2 n2 π 2 t nπx
U (x, t) = cn Un (x, t) = cn exp(− 2
) sin , (8.12)
n=1 n=1
L L
where cn are as yet undetermined.

To find the value of cn , apply initial condition U (x, 0) = f (x) into Equation (8.12),
∞
X nπx
f (x) = cn sin .
n=1
L
So we need to choose cn so that the series of sine functions converges to f (x). Notice that the
series on the right hand side is just the Fourier sine series for f , so its coefficients are given by
2 ∞
Z
nπx
cn = bn = f (x) sin dx.
L 0 L
In conclusion, we get the following result
Heat Problem Type I: Homogeneous Boundary Conditions

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t>0
∞
X α 2 n2 π 2 t nπx
The solution is U (x, t) = cn exp(− 2
) sin ,
n=1
L L
2 L
Z
nπx
where cn =f (x) sin dx.
L 0 L
( The steady state is lim U (x, t) = 0.)
t→∞
Example 141. Solve
8Uxx = Ut , 0 ≤ x ≤ 5, t>0
U (0, t) = 0, U (5, t) = 0,
U (x, 0) = x.
Solution We have α2 = 8, L = 5 and f (x) = x. So the general solution is

∞
X 8n2 π 2 t nπx
U (x, t) = cn exp(− ) sin
n=1
52 5
∞
X 8n2 π 2 t nπx
= cn exp(− ) sin
n=1
25 5
To find the particular solution, let’s find cn by using integration-by-parts.
2 5
Z
nπx
cn = x sin dx.
5 0 5
x=5
2 5 nπx 25 nπx
= − x cos + 2 2 sin
5 nπ 5 n π 5
x=0
2 25 10
= (− cos nπ) = − cos nπ
5 nπ nπ
10 10
= − (−1)n = (−1)n+1 .
nπ nπ
So the particular solution is
∞
X 10 8n2 π 2 t nπx
U (x, t) = (−1)n+1 exp(− ) sin .
n=1
nπ 25 5

Suppose the temperature distribution function U (x, t) of a rod that has both ends constantly
kept at 0 degree is given by the heat conduction problem
4Uxx = Ut , 0 ≤ x ≤ 6, t>0
U (0, t) = 0, U (6, t) = 0,
πx 3πx
U (x, 0) = 2 sin + 4 sin πx − 10 sin .
3 2
a) Find the particular solution of the above initial-boundary value problem.
Solution Here α2 = 4 and L = 6. So the general solution is
∞
X 4n2 π 2 t nπx
n=1
62 6
∞
X n2 π 2 t nπx
= cn exp(− ) sin
n=1
9 6
So when t = 0 we have
∞
X nπx
U (x, 0) = cn sin . (8.13)
n=1
6
On the other hand, it is given by the initial condition that

πx 3πx
U (x, 0) = 2 sin + 4 sin πx − 10 sin (8.14)
3 2
By Equations (8.13) and (8.14),
∞
X nπx πx 3πx
cn sin = 2 sin + 4 sin πx − 10 sin .
n=1
6 3 2
To find cn , comparing the angles of the expressions,

πx nπx
= ⇒ n=2 ⇒ c2 = 2
3 6
nπx
πx = ⇒ n=6 ⇒ c6 = 4
6
3πx nπx
= ⇒ n=9 ⇒ c9 = −10.
2 6
And cn = 0 where n 6= 2, 6, 9. Substitute all cn back into the general solution to get a
particular solution,
∞
X n2 π 2 t nπx
n=1
9 6
22 π 2 t 2πx 62 π 2 t 6πx 92 π 2 t 9πx
= c2 exp(− ) sin + c6 exp(− ) sin + c9 exp(− ) sin
9 6 9 6 9 6
4π 2 t πx 3πx
= 2 exp(− ) sin + 4 exp(−4π 2 t) sin πx − 10 exp(−9π 2 t) sin .
9 3 2
b) What is lim U (x, t)?

n→∞
Solution The limit is 0.
Exercises 8.3.1
1. (a) Use separation of variables to rewrite the partial differential equation below into a pair
of ordinary differential equations.
utt + 5u = 4uxx .
(b) Suppose the above partial differential equation has boundary condition ux (0, t) =
0, u(20, t) = 0. Use separations of variables to determine the corresponding bound-
ary conditions that the ordinary differential equations found in (a) must satisfy.
(c) (Yes or no) Could the partial differential equation, uxx − 2uxt = 5utt , be separated into
two ordinary differential equations?
2. Consider the following partial differential equation
uyy + 2uy = y 2 uxx
(a) Separate this equation into two ordinary differential equations.

(b) Translate the following boundary conditions on the above partial differential equation
to conditions on the ordinary differential equations found above.
u(0, y) = 0, u(L, y) = 0
3. True or false:
(a) It is possible to separate the partial differential equation, 4uxx = utt + t2 ut , into two
ordinary differential equations.
(b) Using the formula u(x, t) = X(x)T (t), the boundary conditions u(0, t) = 0 and ux (9, t) =
0 can be rewritten as X(0) = 0 and X 0 (9) = 0.
(c) Using the formula u(x, t) = X(x)T (t), the boundary conditions ux (0, t) = 0 and
ux (5, t) = 0 can be rewritten as T 0 (0) = 0 and T 0 (5) = 0.
(d) Using the formula u(x, t) = X(x)T (t), the boundary conditions ux (0, t) = 0 and
u(1, t) = 1 can be rewritten as X 0 (0) = 0 and X(1) = 1.
4. Consider the two linear partial differential equations.
(I) uxx = uxt − 2ut

(II) uxx = 2utt − x4 u
Use the substitution u(x, t) = X(x)T (t) and attempt to separate each equation into two
ordinary differential equations. Which statement below is true?
(a) Neither equation is separable.

(b) Only (I) is separable.
(c) Only (II) is separable.
(d) Both equations are separable.
5. Consider the two linear partial differential equations.
(I) uxx + 4uxt + 4utt = 0

(II) uxx − 4utx − 5u = 0
Use the substitution u(x, t) = X(x)T (t) and attempt to separate each equation into two
ordinary differential equations. Which statement below is true?
(a) Neither equation is separable.
(b) Only (I) is separable.
(c) Only (II) is separable.
(d) Both equations are separable.
6. Using the substitution u(x, t) = X(x)T (t), where u(x, t) is not the trivial solution, consider
the two statements below.
(I) The equation uxx = utt + 4u can be separated into two ordinary differential equations.
(II) The boundary conditions ux (0, t) = 0 and u(4, t) = 0 can be rewritten into X 0 (0) = 0
and X(4) = 0.
(a) Neither is true.
(b) Only (I) is true.
(c) Only (II) is true.
(d) Both are true.
7. Using the substitution u(x, t) = X(x)T (t), where u(x, t) is not the trivial solution. Consider
the two statements below.
(I) The equation uxt = utt + 4ut can be separated into two ordinary differential equations.
(II) The boundary conditions ux (0, t) = 0 and u(6, t) = 0 can be rewritten into X 0 (0) = 0
and X(6) = 0.
What can you say regarding the truthfulness of these statements?
(a) Only (I) is true.

(b) Only (II) is true.
(c) Both are true.
(d) Neither is true.
Answers
1. a) 4X 00 + λX = 0, and T 00 + (λ + 5)T = 0, b) X 0 (0) = 0, X(20) = 0, c) No
2. a) X 00 + λX = 0, and Y 00 + 2Y 0 + λy 2 Y = 0, b) X(0) = X(L) = 0
3. a) T, b) T, c) F, d) F
4) d)
5. c)
6. d)
7. c)
8.3.2 Type II: Nonhomogeneous Boundary Conditions

Suppose now that one end of the bar is held at a constant temperature T1 and the other is
maintained at a constant temperature T2 where T1 and T2 are not necessary zero. So now our
IBVP becomes
Heat Problem Type II: Nonhomogeneous Boundary Conditions

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t>0 (8.15)
(Initial condition) U (x, 0) = f (x), 0 < x < L, (8.16)
(Boundary conditions) U (0, t) = T1 , U (L, t) = T2 , t > 0. (8.17)
This problem is only slightly more difficult to solve since we can reduce it to a problem of Type I
by introducing the concept of steady-state solution, V (x).
After a long period of time (t → ∞), we anticipate that a steady temperature distribution V (x)
will be reached, which is independent of time t. Thus, by definition, we have
V (x) = lim U (x, t).

t→∞
So V must satisfy Equations (8.15) and (8.17). Note that
Vxx = V 00 , Vt = 0.
Hence
(8.15) becomes lim α2 Uxx = lim Ut ⇒ α2 Vxx = Vt ⇒ V 00 = 0,

t→∞ t→∞
(8.17) becomes lim U (0, t) = lim T1 ⇒ V (0) = T1 ,
t→∞ t→∞
lim U (L, t) = lim T2 , ⇒ V (L) = T2 .
t→∞ t→∞
Now we have IVP of homogeneous equation,
V 00 = 0, V (0) = T1 , V (L) = T2 .
Let’s solve it.
V 00 = 0 ⇒ V (x) = c1 x + c2
V (0) = T1 ⇒ c2 = T1
T2 − T1
V (L) = T2 ⇒ T2 = c1 L + T1 ⇒ c1 = .
L
Hence we finally derive the formula for steady state function V ,
T2 − T1
V (x) = x + T1 .
L
Now let’s return to the original problem, express U (x, t) as
U (x, t) = V (x) + W (x, t),
where W (x, t) is called transient part. Substitute this into Equations (8.15)-(8.17), we have
α2 Uxx = Ut ⇒ α2 (V + W )xx = (V + W )t ⇒ α2 (V 00 + Wxx ) = Vt + Wt ⇒ α2 Wxx = Wt

U (x, 0) = f (x) ⇒ V (x) + W (x, 0) = f (x) ⇒ W (x, 0) = f (x) − V (x)
U (0, t) = T1 ⇒ V (0) + W (0, t) = T1 ⇒ T1 + W (0, t) = T1 ⇒ W (0, t) = 0
U (L, t) = T2 ⇒ V (L) + W (L, t) = T2 ⇒ T1 + W (L, t) = T1 ⇒ W (L, t) = 0.
We use these facts, V 00 = 0, Vt = 0, for the first equation. Now we can see that all equations of W
is exactly the Type I heat problem. So we have the general solution for W ,
∞
W (x, t) = cn exp(− 2
) sin ,
n=1
L L
where Z L
2 nπx
cn = (f (x) − V (x)) sin dx.
L 0 L
Finally, we then get the general solution of U , as follow
U (x, t) = V (x) + W (x, t)

∞
α 2 n2 π 2 t
X
T2 − T1 nπx
= x + T1 + cn exp(− 2
) sin .
L n=1
L L
In conclusion, we then have
Heat Problem Type II: Nonhomogeneous Boundary Conditions

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t > 0
(Boundary conditions) U (0, t) = T1 , U (L, t) = T2 , t > 0.
∞
The solution is U (x, t) = V (x) + cn exp(− ) sin ,
n=1
L2 L
Z L
2 nπx
where cn = (f (x) − V (x)) sin dx,
L 0 L

T2 − T1
and V (x) = x + T1 .
L
( The steady state is lim U (x, t) = V (x).)
t→∞

Suppose the temperature distribution function U (x, t) of a rod is given by initial-boundary
value problem
3Uxx = Ut , 0 ≤ x ≤ 5, t>0
U (0, t) = 30, U (5, t) = 50,
U (x, 0) = 5x + 30.
a) Find its steady-state solution.

Solution We have T1 = 30, T2 = 50 and L = 5. Then
T2 − T1 50 − 30
V (x) = x + T1 = x + 30 = 4x + 30.
L 5
b) Find the general solution of the above initial-boudary value problem.

Solution We have α2 =3. So the general solution is

∞
U (x, t) = V (x) + cn exp(− ) sin
n=1
L2 L
∞
X 3n2 π 2 t nπx
= 4x + 30 + cn exp(− ) sin
n=1
25 5
c) Find the formula for cn .

Solution
Z L
2 nπx
cn = (f (x) − V (x)) sin dx
L 0 L
Z 5
2 nπx
= ((5x + 30) − (4x + 30)) sin dx
5 0 5
Z 5
2 nπx
= x sin dx
5 0 5
x=5
2 5 nπx 25 nπx
= − x cos + 2 2 sin
5 nπ 5 n π 5
x=0
2 25
= (− cos nπ)
5 nπ
10
= (−1)n+1 .
nπ
d) What is lim U (1, t)?

n→∞
Solution lim U (1, t) = V (1) = 4(1) + 30 = 34.
n→∞
Example 144 (Fall 2010, Final Exam). Suppose the temperature distribution function U (x, t)
of a rod that has both ends constantly kept at different temperatures is given by initial-boundary
value problem
4Uxx = Ut , 0 ≤ x ≤ 1, t>0
U (0, t) = 60, U (1, t) = 40,
U (x, 0) = 60 − 20x − 30 sin 2πx + 50 sin 7πx.
a) Find the steady-state solution of the above initial-boundary value problem.

Solution We have T1 = 60, T2 = 40 and L = 1. Then
T2 − T1 40 − 60
V (x) = x + T1 = x + 60 = −20x + 60
L 1
b) Based on the given boundary conditions, state the general form of the solution. Then
find the particular solution of the initial-boudary value problem.
Solution Here we have α2 = 4. So the general solution is

∞
U (x, t) = V (x) + cn exp(− ) sin
n=1
L2 L
∞
X
= −20x + 60 + cn exp(−4n2 π 2 t) sin nπx.
n=1
Notice that by substituting t = 0 into above equation, we get

∞
X
U (x, 0) = −20x + 60 + cn sin nπx. (8.18)
n=1
But it is given that
U (x, 0) = 60 − 20x − 30 sin 2πx + 50 sin 7πx. (8.19)
So by comparing Equations (8.18) with (8.19), we find cn as follows:
2πx = nπx ⇒ n=2 ⇒ c2 = −30,

7πx = nπx ⇒ n=7 ⇒ c7 = 50.
And cn = 0 where n 6= 2, 7. Substitute all cn back into the general solution to get a
∞
X
U (x, t) = −20x + 60 + cn exp(−4n2 π 2 t) sin nπx
n=1
= −20x + 60 + c2 exp(−4(2)2 π 2 t) sin 2πx + c7 exp(−4(7)2 π 2 t) sin 7πx
= −20x + 60 − 30 exp(−16π 2 t) sin 2πx + 50 exp(−196π 2 t) sin 7πx.
c) What is lim U (0.9, t)?

n→∞
Solution lim U (0.9, t) = V (0.9) = −20(0.9) + 60 = −18 + 60 = 42.
n→∞
8.3.3 Type III: Bar with Both Ends Insulated

Now let’s consider the situation where the two ends of the bar are also sealed with perfect insulation
so that no heat could escape to the outside environment, or vice versa. The new boundary
conditions are
Ux (0, t) = 0, Ux (L, t) = 0,
reflecting the fact that there will be no heat transferring at x = 0, x = L. Now we would like to
solve
Heat Problem Type III: Bar with Both Ends Insulated

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t > 0,
(Boundary conditions) Ux (0, t) = 0, Ux (L, t) = 0, t > 0.
By using the method of separation of variables to solve. Two ordinary differential equations we
get are the same as before which are
X 00 + λX = 0, T 0 + α2 λT = 0.
The only different conditions we have this time come from new boundary conditions,
Ux (0, t) = 0 ⇒ X 0 (0)T (t) = 0 ⇒ X 0 (0) = 0 or T (t) = 0

Ux (L, t) = 0 ⇒ X 0 (L)T (t) = 0 ⇒ X 0 (L) = 0 or T (t) = 0
Again we discard the case when T (t) = 0 since it’s gonna gives us an interesting trivial solution
U (x, t) = X(x)T (t) = 0. Hence we have to let the new boundary conditions to be
X 0 (0) = 0, X 0 (L) = 0.
So now we have transformed the problem to
X 00 + λX = 0, X 0 (0) = 0, X 0 (L) = 0, (8.20)

0 2
T + α λT = 0. (8.21)
The first equation (8.20) is the Eigenvalue problem. The result is summarized below:
Case 1: λ > 0, the eigenvalues are
nπ
λn = ( )2 ,
L
nπx
Xn (x) = cos , n = 1, 2, . . . .
L
Case 2: λ0 = 0, it is indeed an eigenvalue, with corresponding eigenfunction X0 (x) = 1.

Case 3: λ < 0, there are neither eigenvalues nor eigenfunctions in this case.
For λ0 = 0, substitute its value in Equation (8.21),
T0 = 0 ⇒ T0 = constant. (We will use 1 for simplicity.)

nπ 2
For λn = ( ) , we get
L
α 2 n2 π 2 α 2 n2 π 2 t
T 0 + α2 λT = 0 ⇒ T0 + T =0 ⇒ Tn (t) = exp(− ), n = 1, 2, . . .
L2 L2
Therefore we have the fundamental set of solutions as follows:
U0 (x, t) = X0 (x)T0 (t) = 1

α 2 n2 π 2 t nπx
Un (x, t) = Xn (x)Tn (t) = exp(− ) cos .
L2 L
Then the general solution which is their linear combination is
∞
U (x, t) = c0 + cn exp(− 2
) cos ,
n=1
L L
where cn are as yet undetermined.

In order to find the value of cn , apply initial condition U (x, 0) = f (x) into the general solution,
∞
X nπx
f (x) = c0 + cn cos ,
n=1
L
which is a Fourier cosine series. So we then have

1 2 ∞ 1 ∞
Z Z
a0
c0 = = f (x)dx = f (x) dx,
2 2 L 0 L 0
2 ∞
Z
nπx
c n = bn = f (x) cos dx, n = 1, 2, . . . .
L 0 L
In conclusion, we get the following result
Heat Problem Type III: Bar with Both Ends Insulated

(Heat equation) α2 Uxx = Ut , 0 ≤ x ≤ L, t > 0,
(Boundary conditions) Ux (0, t) = 0, Ux (L, t) = 0, t > 0.
∞
X α2 n2 π 2 t nπx
The solution is U (x, t) = c0 + cn exp(− ) cos ,
n=1
L2 L
Z L
1
where c0 = f (x) dx,
L 0
2 L
Z
nπx
and cn = f (x) cos dx, n = 1, 2, . . . .
L 0 L
( The steady state is lim U (x, t) = c0 .)
t→∞
Suppose the temperature distribution function U (x, t) of a rod that both ends perfectly insulated
is given by the initial-boundary value problem
3Uxx = Ut , 0 ≤ x ≤ 2π, t>0

Ux (0, t) = 0, Ux (2π, t) = 0,
U (x, 0) = 20 − 10 cos x + 30 cos 4x.
a) Find the particular solution of the above initial-boundary value problem.

Solution For this problem we have α2 = 3, L = 2π. So the general solution is

∞
X 3n2 π 2 t nπx
U (x, t) = c0 + cn exp(− ) cos
n=1
(2π)2 2π
∞
X 3n2 t nx
= c0 + cn exp(− ) cos
n=1
4 2
Notice that by substituting t = 0 into above equation, we get

∞
X nx
U (x, 0) = c0 + cn cos . (8.22)
n=1
2
But it is provided that
U (x, 0) = 20 − 10 cos x + 30 cos 4x. (8.23)
c0 = 20,
nx
x= ⇒ n=2 ⇒ c2 = −10,
2
nx
4x = ⇒ n=8 ⇒ c8 = 30.
2
And cn = 0 where n 6= 0, 2, 8. Substitute all cn back into the general solution to get a
3(2)2 t 2x 3(8)2 t 8x
U (x, t) = c0 + c2 exp(− ) cos + c8 exp(− ) cos
4 2 4 2
= 20 − 10 exp(−3t) cos x + 30 exp(−48t) cos 4x.
b) What is lim U (1, t)?

t→∞
Solution
lim U (x, t) = lim (20 − 10 exp(−3t) cos x + 30 exp(−48t) cos 4x) = 20,
n→∞ n→∞
or simply lim U (x, t) = c0 = 20. In particular, lim U (1, t) = 20.

t→∞ t→∞
Exercises 8.3
1. Solve the heat conduction problem described by:
ut = 9uxx , 0 < x < 5, t>0

u(0, t) = u(5, t) = 0, for t > 0
2πx 7πx
u(x, 0) = sin − 3 sin πx + 13 sin
5 5
2. Solve the nonhomogeneous heat equation
10uxx = ut , 0 < x < 6, t>0

u(0, t) = 20, u(6, t) = 80, t>0
3πx
u(x, 0) = 20 + 10x + 5 sin πx − 10 sin
2
3. Find the steady-state solution of the heat conduction problem
uxx = ut , 0 < x < 10, t>0

u(0, t) = 100, u(10, t) = 60,
u(x, 0) = f (x).
(a) v(x) = −40x + 60

(b) v(x) = −4x + 100
(c) v(x) = 4x + 60
(d) v(x) = 40x + 100
4. Suppose the temperature distribution function u(x, t) of a rod that has both ends constantly
kept at 0 degree is given by the heat conduction problem
4uxx = ut , 0 < x < 6, t>0

u(0, t) = 0, u(6, t) = 0,
πx 3πx
u(x, 0) = 2 sin + 4 sin πx − 10 sin .
3 2
(a) Find the particular solution of the above initial-boundary value problem.
(b) What is lim u(x, t)?
t→∞
5. Suppose the temperature distribution function u(x, t) of a rod that has both ends perfectly
insulated is given by the initial-boundary value problem
9uxx = ut , 0 < x < 4, t>0

ux (0, t) = 0, ux (4, t) = 0,
u(x, 0) = 2 − cos πx − 7 cos 5πx.
(a) Find the particular solution of the above initial-boundary value problem.
(b) What is lim u(3, t)?
t→∞
6. Consider the following two heat conduction initial-boundary value problem

uxx = ut , 0 < x < 10, t>0
u(0, t) = 0, u(10, t) = 0, I.
u(x, 0) = 200.
uxx = ut , 0 < x < 10, t>0

ux (0, t) = 0, ux (10, t) = 0, II.
u(x, 0) = 200.
(a) Which problem (I or II) models the temperature distribution of a rod with both ends
perfectly insulated?
(b) Suppose u1 (x, t) is the solution of problem I and u2 (x, t) is the solution of problem II.
Compare the temperatures at the middle of the rod (that is, at x = 5) as t → ∞.
i. lim u1 (5, t) > lim u2 (5, t)
t→∞ t→∞
ii. lim u1 (5, t) < lim u2 (5, t)
t→∞ t→∞
iii. lim u1 (5, t) = lim u2 (5, t)
t→∞ t→∞
iv. There is not enough available information to compare them.
7. Find the steady-state solution, v(x), of the heat conduction problem
6uxx = ut , 0 < x < 4, t>0
u(0, t) = 5, ux (4, t) = 2,
u(x, 0) = f (x).
(a) v(x) = 34 x − 5
(b) v(x) = −2x − 5
(c) v(x) = 2x + 5
(d) v(x) = − 34 x + 5
8. . Find the steady-state solution, v(x), of the heat conduction problem with nonhomogeneous
boundary conditions:
α2 uxx = ut , 0 < x < 6, t>0
u(0, t) = 100, u(6, t) − 2ux (6, t) = 20,
u(x, 0) = f (x).
(a) v(x) = −20x + 100
(b) v(x) = 40
3 x + 100
(c) v(x) = − 40
3 x + 100
(d) v(x) = 20x + 100
Answers
36 2 2 441 2
1. u(x, t) = e− 25 π t sin 2πx
5 − 3e
−9π t
sin πx + 13e− 25 π t sin 7πx
5
2 45 2
2. u(x, t) = 20 + 10x + 5e−10π t sin πx − 10e− 2 π t sin 3πx2
3. b)
4 2 2 2
4. a) u(x, t) = 2e− 9 π t sin πx
3 + 4e
−4π t
sin πx − 10e−9π t sin 3πx
2 , b) lim u(x, t) = 0
t→∞
2 2
5. a) u(x, t) = 2 − e−9π t cos πx − 7e−225π t cos 5πx, b) lim u(3, t) = 2
t→∞
6. a) II, b) (ii)
7. c)
8. a)
8.4 Wave Equation

Suppose that an elastic string of length L is tightly stretched between two supports at the same
horizontal level so that the x-axis lies along the string.
Figure 8.3: A vibrating string.
Let U (x, t) be the vertical displacement experienced by the string at the point x at time t. If
damping effects, such as air resistance, are neglected and if the amplitude of the motion is not too
large, then U (x, t) satisfies
a2 Uxx = Utt ,
where a, horizontal propagation speed of the wave motion, is a positive constant. The equation
is known as wave equation. Since the ends are assumed to remain fixed and therefore the
boundary conditions are
U (0, t) = 0, U (L, t) = 0, t > 0.
In addition, the equation also come with two initial conditions,
U (x, 0) = f (x) (initial displacement)
Ut (x, 0) = g(x) (initial velocity).
Alternatively, it can be considered as a boundary value problem in the semi-infinite strip 0 < x <
L, t > 0 of the xt-plane. One condition is imposed at each point on the semi-infinite sides, and
two are imposed at each point on the finite base.
Figure 8.4: Boundary value problem for the wave equation.
So the IBVP we would like to solve is
Wave Equation
2
(Wave equation) a Uxx = Utt , 0 ≤ x ≤ L, t>0
(Initial conditions) U (x, 0) = f (x), Ut (x, 0) = g(x) 0 < x < L,
Let’s use the method of separation of variables by writing
U (x, t) = X(x)T (t).
Thus Uxx = X 00 T, Ut = XT 00 . Substituting these into the wave equation,

X 00 T 00
a2 Uxx = Utt ⇒ a2 X 00 T = XT 00 ⇒ = 2 = −λ,
X a T
where λ is constant. Now consider them separately,
X 00
= −λ ⇒ X 00 + λX = 0.
X
T 00
= −λ ⇒ T 00 + α2 λT = 0.
α2 T
For boundary conditions, again we use U (x, t) = X(x)T (t),
U (0, t) = 0 ⇒ X(0)T (t) = 0 ⇒ X(0) = 0 or T (t) = 0

U (L, t) = 0 ⇒ X(L)T (t) = 0 ⇒ X(L) = 0 or T (t) = 0
After we discard the case of T (t) = 0, as we did before, we obtain
X(0) = 0, X(L) = 0.
Thus our problem changes to
X 00 + λX = 0, X(0) = 0, X(L) = 0 (8.24)

00 2
T + α λT = 0. (8.25)
Recall once again that the result of BVP (8.24) is that the eigenvalues are
nπ 2
λn = ( ) ,
L
nπx
Xn (x) = sin ,
L
where n = 1, 2, . . .
nπ
Now let’s substitute λn = ( )2 into Equation (8.25),
L
α 2 n2 π 2
T 00 + α2 λT = 0 ⇒ T 00 + T =0
L2
anπt anπt
⇒ Tn (t) = An cos + Bn sin , n = 1, 2, . . .
L L
anπ
Notice that the above equation is the second order homogeneous equation with r = ± i, as
L
the roots of its characteristic equation.
So, for each n = 1, 2, . . .,

anπt anπt nπx
Un (x, t) = Xn (x)Tn (t) = An cos + Bn sin sin .
L L L
Then the linear combination of Un form the general solution of the heat equation,
∞
X anπt anπt nπx
U (x, t) = An cos + Bn sin sin , (8.26)
n=1
L L L
where An and Bn are not yet determined.

To find the value of An , apply initial condition U (x, 0) = f (x) into Equation (8.26),
∞
X nπx
f (x) = An sin .
n=1
L
Hence f needs to be a Fourier sine series and so we expand it into its odd periodic extension to
receive,
2 ∞
Z
nπx
An = bn = f (x) sin dx.
L 0 L
Note here that
∞
X anπ anπt anπ anπt nπx
Ut (x, t) = −An sin + Bn cos sin . (8.27)
n=1
L L L L L
Plug in another initial condition Ut (x, 0) = g(x) into (8.27),

∞
X anπ nπx
g(x) = Bn sin .
n=1
L L
So g needs to be also a Fourier sine series. By expanding it into odd periodic extension,
∞ ∞
X nπx X anπ nπx
bn sin = g(x) = Bn sin .
n=1
L n=1
L L
It immediate follows that

!
Z L Z L
L L 2 nπx 2 nπx
Bn = bn = g(x) sin dx = g(x) sin dx.
anπ anπ L 0 L anπ 0 L
We conclude what we get here:
Wave Equation
2
(Wave equation) a Uxx = Utt , 0 ≤ x ≤ L, t > 0,
(Initial conditions) U (x, 0) = f (x), Ut (x, 0) = g(x) 0 < x < L,
∞
X anπt anπt nπx
The solution is U (x, t) = An cos + Bn sin sin ,
n=1
L L L
2 ∞
Z
nπx
where An = f (x) sin dx,
L 0 L
Z L
2 nπx
and Bn = g(x) sin dx.
anπ 0 L
( The steady state which is lim U (x, t) doesn’t exist. )
t→∞
We take a look here in two particular cases which make our solution in much simpler form.
Special case 1: If f (x) 6= 0, g(x) = 0, then all Bn = 0. Hence
Wave Equation: Special Case 1

(Wave equation) a2 Uxx = Utt , 0 ≤ x ≤ L, t > 0,
(Initial conditions) U (x, 0) = f (x), Ut (x, 0) = 0 0 < x < L,
∞
X anπt nπx
The solution is U (x, t) = An cos sin ,
n=1
L L
2 ∞
Z
nπx
where An = f (x) sin dx.
L 0 L
( The steady state lim U (x, t) doesn’t exist. )
t→∞
Special case 2: If f (x) = 0, g(x) 6= 0, then all An = 0. Hence
Wave Equation: Special Case 2

(Wave equation) a2 Uxx = Utt , 0 ≤ x ≤ L, t>0
(Initial conditions) U (x, 0) = 0, Ut (x, 0) = g(x) 0 < x < L,
∞
X anπt nπx
The solution is U (x, t) = Bn sin sin ,
n=1
L L
Z L
2 nπx
where Bn = g(x) sin dx.
anπ 0 L
( The steady state lim U (x, t) doesn’t exist. )
t→∞
Example 146 (Summer 2010, Final Exam).

Solve the one-dimensional wave problem
Utt = 4Uxx , 0 ≤ x ≤ 5, t>0

U (0, t) = 0, U (5, t) = 0,
U (x, 0) = 2 sin πx − 3 sin 2πx, Ut (x, 0) = 0.
Solution We have a2 = 4(a = 2), L = 5, f (x) 6= 0 and g(x) = 0. This is a wave equation
special case 1. So the general solution is
∞ ∞
X anπt nπx X 2nπt nπx
U (x, t) = An cos sin = An cos sin .
n=1
L L n=1
5 5
Notice that by substituting t = 0 into the above equation, we have

∞
X nπx
U (x, 0) = An sin . (8.28)
n=1
5
But it is provided that

U (x, 0) = 2 sin πx − 3 sin 2πx. (8.29)
nπx
= πx ⇒ n=5 ⇒ A5 = 2,
5
nπx
= 2πx ⇒ n = 10 ⇒ A10 = −3.
5
And An = 0 where n 6= 5, 10. Substitute all An back into the general solution to get a
5πx 10πx
U (x, t) = A5 sin + A10 sin
5 5
= 2 sin πx − 3 sin 2πx.
Exercises 8.4
1. Which of the following functions below is a solution of the wave equation
utt = 4uxx ?
2
(a) e−4π t sin(πx)
(b) sin(x − 2t)
(c) x2 + t2
(d) 1 + 4 cos(t) + x2
2. Consider the initial-boundary value problems for the wave equation
9uxx = utt , 0 < x < 3, t>0

u(0, t) = 0, u(3, t) = 0,
u(x, 0) = f (x), ut (x, 0) = 0.
In what specific form will its general solution appear?

∞
X nπx
(a) u(x, t) = An cos(nπt) sin( )
n=1
3
∞
X nπx
(b) u(x, t) = An cos(3nπt) sin( )
n=1
3
∞
X nπx
(c) u(x, t) = Bn sin(nπt) sin( )
n=1
3
∞
X nπx
(d) u(x, t) = Bn sin(3nπt) sin( )
n=1
3
3. Suppose the displacement u(x, t) of a piece of flexing string is given by the initial-boundary
value problem
9uxx = utt , 0 < x < 2, t>0

u(0, t) = 0, u(2, t) = 0,
u(x, 0) = 0
ut (x, 0) = 4 − x2 .
(a) What is the physical meaning of the boundary conditions?

(b) What is the propagation speed of the standing waves?
(c) When t = 0, what is the velocity of the vibrating string at x = 12 ?
(d) In what specific form will its general solution appear?
∞
X 3nπt nπx
i. u(x, t) = Cn sin sin
n=1
2 2
∞
X 3nπt nπx
ii. u(x, t) = Cn sin cos
n=1
2 2
∞
X 3nπt nπx
iii. u(x, t) = Cn cos cos
n=1
2 2
∞
X 3nπt nπx
iv. u(x, t) = Cn cos sin
n=1
2 2
(e) True or False: The coefficients of the solution in part (d) above can be found using the
integral Z 2
2 nπx
Cn = (4 − x2 ) sin dx.
3nπ 0 2
(f) True or False: As t → ∞, the solution u(x, t) in (e) will reach a limit.
4. Suppose the displacement u(x, t) of a piece of flexing string that has both ends firmly fixed
in places is given by the initial-boundary value problem
9uxx = utt , 0 < x < 6, t>0

u(0, t) = 0, u(6, t) = 0,
u(x, 0) = 0
ut (x, 0) = 0.5.
(a) State the general form of its solution. Then find the particular solution of the initial-
boundary value problem.
(b) True or False: As lim u(1, t) = 0.5. Explain your answer.
t→∞
5. Suppose the displacement u(x, t) of a piece of flexing string that has both ends firmly fixed
in places is given by the initial-boundary value problem
16uxx = utt , 0 < x < 4, t>0

u(0, t) = 0, u(4, t) = 0,
√ √ 5
u(x, 0) = 5 sin(πx) − 2 7 sin( πx),
2
ut (x, 0) = 0.
(a) State the general form of its solution. Then find the particular solution of the initial-
boundary value problem.
(b) True or False: As t → ∞, the solution u(x, t) will go to 0 for all x.
(c) True or False: If f (x) = 0, then u(x, t) = 0 is the unique solution.
Answers
1. b)
2. a)
3. a) Two ends of the string are clamped in fixed positions at the horizontal level so they are held
motionless at all time. b) a = 3, c) 15
4 , d) i, e) T,Pf) F
∞
4. a) The general form of the solution is u(x, t) = n=1 Bn sin nπt nπx
2 sin 6 . The particular solution
P∞ 4 (2n−1)πt (2n−1)πx
is u(x, t) = n=1 (2n−1)2 π2 sin 2 sin 6 P, b) F
∞
5. a) The general form of the solution is u(x, t) = n=1 Cn cos nπt sin nπx
4 . The particular solution
√ √
is u(x, t) = 5 cos(4πt) sin(πx) − 2 7 cos(10πt) sin( 25 πx), b) F, c) T

Math 251 Notes

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Math 251 Notes

Hochgeladen von

Copyright:

Verfügbare Formate

Math 251: Ordinary Differential Equation

and Partial Differential Equation

July 14, 2014

2 First Order Differential Equations 12

3 Second Order Differential Equations 43

4 Higher Order Linear Equation 95

5 Laplace Transform 104

6 Systems of First Order Linear Equations 138

6.3.1 Real and Distinct Eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . . . 150

7 Nonlinear Differential Equations and Stability 173

8 Partial Differential Equation 184

Example 1 (Newton’s Second Law of motion). If an object of mass m is moving with

• We write first, second and third derivatives of y as y 0 , y 00 and y 000 respectively.

Example 2. Consider the following differential equations.

Differential equations Order ODE/PDE Linear/Nonlinear

y 000 + 2 sin(t)e3t y 00 + cos(t)y = 3 3 ODE linear

Differential equations Order Linear/Nonlinear

y 0 − cos(2t)y = t2 − e−5t 1 linear

y 000 + ty 0 + (cos2 y)t = t3 3 nonlinear

y 0 (t) = ty 2 (t) 1 nonlinear

y 00 (t) = ln(t)y(t) 2 linear

ty(t)y 0 (t)y 00 (t) − 1 = 0 2 nonlinear

Example 4 (Fall 2009 Midterm Exam I).

Which function is a solution of the differential equation

Solution Let’s consider each choice separately,

(c) Since ((−t)0 )2 − 5t(−t) = 5t2 + 1 ⇒ −t is a solution.

Example 5 (Summer 2003 Midterm Exam I).

Which equation has y1 = e−2t and y2 = e3t as two solutions?

Solution Again let’s take a look each choice one by one,

(c) For y = e−2t ⇒ −(e−2t )00 + (e−2t )0 + 6(e−2t ) = 0 ⇒ e−2t is a solution.

• The first order differential equation in the form

Example 6. Draw the direction field of y 0 = y − t.

9. Determine the differential equation whose direction field is given below.

First Order Differential Equations

2.1 Solving First Order Differential Equations

2.1.1 Separable Equation

N (y)dy = M (t)dt. (separate y’s and dy to t’s and dt)

Before we see the examples, let’s learn some new terms.

Example 8 (Fall 2005 Midterm Exam I).

So the general solution (in implicit form) to the above equation is

From the above example, if we simplify further, we have

Example 9 (Summer 2006 Midterm Exam I).

Simplify a few more steps to get the solution in explicit form,

After a few examples, we get one useful observation here:

Example 10 (Fall 2002 Midterm Exam I).

2.1.2 Method of Integrating Factor

µ(t)p(t) = µ0 (t). (2.2)

Notice here that

Hence the formula for the integrating factor µ(t) is

In conclusion, we have the following:

The process of using the method of integrating factor,

2. Find the integrating factor, µ(t), by using

Example 11 (Summer 2006 Midterm Exam I).

Solution Rewrite the differential equation in canonical form,

t−2 y 0 − 2t−3 y = cos t.

Then apply the given initial condition y(π) = 2,

Therefore the particular solution is

So we have the shortcut here,

Example 12 (Summer 2010 Midterm Exam I).

2y 0 (t) + y(t) = 3t.

(a) What is the integrating factor µ(t)?

(b) What is the general solution of the above differential equation?

Example 13 (Fall 2000 Midterm Exam I).