Ordinary Differential Equations For Scientists and Engineers
Ordinary Differential Equations For Scientists and Engineers
Ordinary Differential Equations For Scientists and Engineers
Gregg Waterman
Oregon Institute of Technology
c
2015
Gregg Waterman
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported
License. The essence of the license is that
You are free:
to Share to copy, distribute and transmit the work
to Remix to adapt the work
Share Alike If you alter, transform, or build upon this work, you may distribute the resulting
work only under the same or similar license to this one.
With the understanding that:
Waiver Any of the above conditions can be waived if you get permission from the copyright
holder.
Public Domain Where the work or any of its elements is in the public domain under applicable
law, that status is in no way affected by the license.
Other Rights In no way are any of the following rights affected by the license:
Your fair dealing or fair use rights, or other applicable copyright exceptions and limitations;
The authors moral rights;
Rights other persons may have either in the work itself or in how the work is used, such as
publicity or privacy rights.
Notice For any reuse or distribution, you must make clear to others the license terms of this
work. The best way to do this is with a link to the web page below.
To view a full copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/ or send a letter
to Creative Commons, 444 Castro Street, Suite 900, Mountain View, California, 94041, USA.
Contents
0 Introduction to This Book
0.1 Goals and Essential Questions . . . . . . . . . . . . . . . . . . . . . . . . . .
0.2 An Illustrative Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 Functions and Derivatives, Variables and Parameters
1.1 Functions and Variables . . . . . . . . . . . . . . . . .
1.2 Derivatives and Differential Equations . . . . . . . . .
1.3 Parameters and Variables . . . . . . . . . . . . . . . .
1.4 Initial Conditions and Boundary Conditions . . . . . .
1.5 Differential Equations and Their Solutions . . . . . . .
1.6 Classification of Differential Equations . . . . . . . . .
1.7 Initial Value Problems and Boundary Value Problems .
1.8 Chapter 1 Summary . . . . . . . . . . . . . . . . . . .
1.9 Chapter 1 Exercises . . . . . . . . . . . . . . . . . . . .
2 First Order Equations
2.1 Solving By Separation of Variables .
2.2 Solution Curves and Direction Fields
2.3 Solving With Integrating Factors . .
2.4 Phase Portraits and Stability . . . .
2.5 Applications of First Order ODEs . .
2.6 Chapter 2 Summary . . . . . . . . .
2.7 Chapter 2 Exercises . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7
9
15
20
23
27
32
36
39
41
43
45
50
55
59
64
73
74
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
of Order
. . . . .
. . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
77
. 79
. 84
. 88
. 93
. 95
. 99
. 104
. 106
.
.
.
.
.
.
.
.
.
.
108
110
115
119
120
122
.
.
.
.
.
128
129
134
139
144
148
.
.
.
.
.
.
.
.
.
1
1
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
A Formula Sheet
150
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
152
152
157
162
164
166
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
ii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
174
174
178
183
185
188
190
0.1
Differential equations are perhaps the most central mathematical topic of science and engineering. Our quest in those areas is to understand and predict the behavior of some sort
of system consisting of a collection of parts that could be things like electrical or mechanical components, living organisms, or some part of the natural world. We often wish
to construct a mathematical model that describes the behavior of the system reasonably
well; such a model usually consists of one of three things:
An equation or a set of equations (an analytical model).
Snapshots of the state of the system at discrete points in time and/or space (a
numerical model).
The problem is that we generally cant construct directly the equation or equations
making up an analytical model of a system. What we will usually have at our disposal are
pieces of information about how a system is changing and/or how forces are acting on and
within the system. Those pieces of information are combined to form an equation containing
the changes or forces, in the form of derivatives. Such an equation containing derivatives is
called a differential equation. Once a differential equation is obtained, we hope we can
use some mathematical technique to extract a model without derivatives that describes the
behavior of the system.
This book is a fairly straightforward introduction to differential equations, with an applied
emphasis. The student should be aware that this is a huge subject, with lifetimes of study
possible. Our hope is that this collection of explanations, examples and exercises will create
a solid foundation for understanding differential equations when they are encountered in
subject-specific courses, and for further study of differential equations themselves.
In the past an introduction to differential equations has usually consisted of learning
specific techniques for solving a variety differential equations. It should be no surprise that
those techniques are easily forgotten in short order! We will look at techniques for obtaining
solutions - that is an essential part of the subject. However, we will also attend to the
bigger picture, in the hopes of giving the student an overall understanding of the subject
that will hopefully be more lasting than just a bunch of recipes for obtaining solutions.
Our study of the subject of differential equations will be guided by some overarching goals,
and essential questions related to those goals.
Goals
Upon completion of his/her study, the student should understand what differential equations, initial value problems, and boundary value problems are, and what their solutions
consist of. For ordinary differential equations (ODEs) and associated initial value or boundary value problems, the student should understand
where such problems come from,
what their solutions consist of,
how solutions are obtained,
how parameters of a system and initial or boundary conditions influence the nature of
solutions.
1
Our pursuit of these goals will take place through the consideration of some related essential
questions.
Essential Questions:
What are differential equations and why do we need them?
How do parameters differ from variables? What is the role of parameters in differential
equations?
What is a mathematical model? How do differential equations and their solutions
model systems and their responses?
It has been demonstrated experimentally that retaining the things we learn can be enhanced if those things are learned through spaced repetition. To that end, I have attempted
to write this book like a novel in which most of the characters are introduced early on, and
are then developed and fleshed out as the plot unfolds. A large number of the important
concepts of differential equations (including at least a little bit about partial differential
equations) are first seen in Chapter 1, then taken up again at later points in the book, where
they are reinforced and expanded upon.
Those having prior knowledge of ordinary differential equations (in most cases, the instructor) will notice that the focus of this book is more on the important concepts related to
differential equations (both ordinary and partial) rather than techniques for solving a broad
range of types of equations. This is based on my conviction that most students will quickly
forget the specific procedures for solving differential equations unless those techniques are
used in other courses taken shortly after this one. However, a person with a good working
understanding of differential equations, initial value problems and boundary value problems
should be able to go to any of the many resources available and quickly remind themselves
of techniques previously learned, or even techniques not seen in this course!
0.2
An Illustrative Example
height h, in feet
In this section we will use a simple and perhaps familiar problem to illustrate many of
the main ideas of this book. In the process we will begin our quest to answer the essential
questions. Lets consider the following situation and question: A rock is fired straight upward
with a velocity of 60 feet per second, from a height of 20 feet off the ground. What will the
rock be doing after being fired?
We will begin by offering what well call a qualitative
solution to the problem that was posed. Intuitively, we
all know that the ascent of the rock will slow as time goes
on, until at some point the rock will come to a complete
stop at its maximum height. It will then begin to pick up
speed downward, falling until it hits the ground. Letting
h represent the height (above the ground) of the rock and
t represent time, we say the height is a function of the time,
20
and we could graph this behavior to obtain the graph shown
to the right. Note that the horizontal axis is for the variable
of time, so the graph does not indicate the the trajectory of
time t
the projectile. The trajectory is straight up and then straight
down.
This qualitative solution might be fine for our purposes if all we wish to know is the
general behavior of the rock after being fired. Suppose, though, that we wanted more. We
might wish to know how high the rock was at any time after being fired, or the maximum
height it obtains. We can determine those things by finding what well call an analytical
solution to the problem, consisting of an equation that gives the height of the rock at any
time. To obtain such a solution, we will need to recall a few things from differential calculus
and physics:
When the position of an object is changing, the rate at which its position is changing
is its velocity, given by the first derivative of position with respect to time.
The rate at which the velocity of an object is changing with respect to time is the
acceleration of the object, which is also the second derivative of position with respect
to time.
F = ma, the force on an object is its mass times its acceleration.
dh
d2 h
and
. We
dt
dt2
will assume that the only force acting on the rock is the force due to gravity. (Later we will
consider the possibility that air resistance applies a force to the rock as well.) The force of
gravity on an object of mass m has been experimentally determined to be mg, where g is
the gravitational constant. For the surface of the earth, that constant has value g = 9.8
m/sec2 or g = 32 ft/sec2 . So we know two things about the force acting on our rock: it is
d2 h
F = ma = m 2 and it is also mg = 32m. (Later we will discuss why this is negative.)
dt
We set our two expressions for force equal to each other, then divide both sides by the mass
m:
d2 h
d2 h
=
= 32
(1)
m 2 = 32m
dt
dt2
Both the equations above are what we call differential equations, which simply means
that they are equations that contain derivatives. (They are equivalent equations, but well
Going back to our rock, its velocity and acceleration at any time are
use the second because it is simpler. Note that the mass of the rock will not be needed from
here on, and our result will be independent of the mass.)
Our goal now is to determine an equation for h, in terms of t. We know that if the
second derivative is 32, as stated in (1) above, then the first derivative must be
dh
= 32t + C1 ,
dt
(2)
(3)
where C2 is yet another unknown constant. This gives us an equation for the height as
a function of time, but it has the problem that it contains two unknown constants. This
function is a solution to the differential equation (1).
Because the highest (and only) derivative in (1) is a second derivative, the differential
equation is called a second order differential equation. What we see in (3) is typical the solution to a second order differential equation contains two arbitrary (meaning they can
have any value) constants. However, the differential equation is not the only information we
dh
have. We also know that h = 20 when t = 0, and
= 60 when t = 0. These pieces of
dt
information are called initial conditions, and they will usually be written in the function
form
h(0) = 20,
h (0) = 60.
(4)
We can substitute the second initial condition into (2) to get
60 = 32(0) + C1 ,
resulting in C1 = 60. If we substitute that value into (3) along with the first condition of
(4), we get
20 = 16(0)2 + 60(0) + C2 ,
giving us that C2 = 20. Thus
(5)
This is our analytical solution to the problem, which allows us to compute the height of the
projectile at any time (until it hits the ground), as well as other things like its maximum
height.
The solution (5) was determined from three pieces of information, the differential equation
and two initial conditions. Those three things together constitute what we call an initial
value problem consisting of a differential equation and any conditions we know are placed
on our solution. In this case the initial value problem is
d2 h
= 32,
dt2
and its solution is h = 16t2 + 60t + 20. Lets now come back to the issue of the signs in
our differential equation and its solution, and lets see if we can interpret what the solution
is telling us.
When solving any problem that has a spatial component, we need to begin by establishing
a coordinate system. In this case the coordinate system is a vertical number line, with its
origin (zero) at ground level and the positive direction being up. So the fact that the rock
4
starts 20 feet above the ground means that h is positive 20 at time zero. The signs
of velocity and acceleration must be consistent with the coordinate system. When the rock
is going upward its velocity is positive, which is why the initial velocity of 60 is positive,
and when the rock is going downward its velocity is negative. The acceleration is always
negative, because it is caused by gravity. When the rock is travelling upward the acceleration
is working against the velocity, causing the rock to slow down. When the rock is travelling
downward the acceleration is working with the velocity, causing the rock to speed up.
Now think about this: If we had launched the rock from ground level at time zero, and
if there was no gravity, the height of the rock at any time t would be h = 60t (distance
equals rate times time). If there was still no gravity, but we want the rock to start at an
initial height of 20 feet, we need to modify our equation to h = 60t + 20. Finally, the
effect of gravity needs to be added in, which is where the 16t2 term comes in.
Section 0.2 Exercises
1. (a) Determine the height of the rock after three seconds. Round your answer to the
nearest tenth, and include units.
(b) Determine when, to the nearest hundredth of a second, the rock is at a height of
50 feet.
2. The solution (5) of the differential equation gives the height of the rock as a function
of time. Although we could determine h values for all real number values of t, that
would not make sense in the context of the problem. Determine the values of t for
which it really does make sense to find values of h. Well call the set of all those
values the domain of the solution (as opposed to the domain of the function, which is
all real numbers). Give your answer using interval notation.
3. Determine the maximum height of the rock and when it occurs, both to the nearest
tenth. Give your answer as a complete sentence that includes both pieces of information
asked for. (Hint: If you are not sure how to approach this, read about the qualitative
solution again for a hint.)
4. At some time that well call zero you are 380 miles from Klamath Falls, on your way
TO Klamath Falls. You are driving at a constant speed of 58 mph. Let x represent
your distance from Klamath Falls and t the time after time zero.
(a) Draw a graph with time on the horizontal axis and distance from Klamath Falls
on the vertical axis, and draw a graph showing what is happening for you in your
car, starting at time zero.
(b) Write two mathematical statements representing each of the two numerical pieces
of information given. (One of those statements should contain a derivative, the
other wont.)
(c) The two pieces of information you gave in (a) constitute an initial value problem.
(It is first order, so its solution will only have one constant and only one initial
condition is needed in order to determine that constant.) Solve the initial value
problem. That is, find an equation for the distance x from Klamath Falls as a
function of time t.
(d) Your answer to (c) is your analytical solution to the initial value problem. Explain
how you know that this agrees with your qualitative solution from (a). (If it
doesnt agree, fix the problem and then answer the question.)
(e) What are the units of x and t? Give your answer as a sentence.
5
Much of science and engineering is concerned with understanding the relationships between measurable, changing quantities that we call variables. Whenever possible we try
to make these relationships precise and compact by expressing them as equations relating
variables; often such equations define functions. In this chapter we begin by looking at ideas
you should be familiar with (functions and derivatives), but hopefully you will now see them
in a deeper and more illuminating way.
We then go on to introduce the idea of a differential equation, and we will see what
we mean by a solution to a differential equation, initial value problem, or boundary value
problem. We will also learn various classifications of differential equations. This is important
in that the method used to solve a differential equation depends on what type of equation
it is.
It is valuable to understand these fundamental concepts before moving on to learning
techniques for solving differential equations, which are addressed in the remainder of the
text.
1.1
spring
mass
Example 1.1(b): Consider a beam that extends horizontally ten feet out from the side of a building, as shown
to the right. The beam will deflect (sag) some, with the
distance below horizontal being greater the farther out
on the beam one looks. The amount of deflection is a
function of how far out a point is on the beam.
deflection is
exaggerated
Example 1.1(c): Suppose that we have a tank containing 100 gallons of water with
10 pounds of salt dissolved in the water, as shown below and to the left. At some time
we begin pumping a 0.3 pounds salt per gallon (of water) solution into the tank at two
gallons per minute, mixing it thoroughly with the solution in the tank. At the same
time the solution in the tank is being drained out at two gallons per minute as well.
See the diagram below and to the right.
0.3 lb/gal at 2 gal/min
100 gal
water
10 lbs salt
100 gal
solution
A lbs salt
2 gal/min
Because the rates of flow in and out of the tank are the same, the volume in the tank
remains constant at 100 gallons. The initial concentration of salt in the tank is 10
pounds/100 gallons = 0.1 pounds per gallon. Because the incoming solution has a
higher concentration, the amount of salt in the tank will change as time goes on. (The
amount will increase, since the concentration of the incoming solution is higher than
the concentration of the solution in the tank.) We can say that the amount of salt in
the tank is a function of time.
12
Example 1.1(d): Consider the equation y = 2 . For any number other than zero
x
that we select for x, there is a corresponding value of y that can be determined by
substituting the x value and computing the resulting value of y. y depends on x,
or y is a function of x
Example 1.1(f): Different points on the surface of a cube of metal one foot on a
side are exposed to different temperatures, with the temperature at each surface point
held constant. The cube eventually attains a temperature equilibrium, where each
point on the interior of the cube reaches some constant temperature. The temperature
at any point in the cube is a function of the three-dimensional location of the point.
In each of the above examples, one quantity (variable) is dependent on one or more
other quantities (variables). The variable that depends on the other variable(s) is called
the dependent variable, and the variable(s) that its values depend on is (are) called the
independent variable(s).
Example 1.1(g): Give the dependent and independent variable(s) for each of Examples 1.1(a) - (f).
Example 1.1(a): The dependent variable is the height of the mass, and the independent
variable is time.
Example 1.1(b): The dependent variable is the deflection of the beam at each point, and
the independent variable is the distance of each point from the wall in which the beam is
embedded.
Example 1.1(c): The dependent variable is the amount of salt in the tank, and the independent variable is time.
Example 1.1(d): The dependent variable is y, and the independent variable is x.
10
Example 1.1(e): The dependent variable is the height at each point on the drumhead, and
the independent variables are the location (in two-dimensional coordinates) of the point on
the drum head, and time. Thus there are three independent variables.
Example 1.1(f): The dependent variable is the temperature at each point in the b, and the
independent variables are the three coordinates giving the position of the point, in three
dimensions.
When studying phenomena like those given in Examples 1.1(a), (b), (c), (e) and (f),
the first thing we do after determining the variables is establish coordinate systems for the
variables. The purpose for this is to be able to attach a number (or ordered set of numbers)
to each point in the domain, and for different positions or states of the dependent variable:
When position is an independent variable, we must establish a one (for the spring),
two (for the drumhead) or three (for the cube of metal) dimensional coordinate system.
This coordinate system will have an origin (zero point) at some convenient location,
indication of which direction(s) is(are) positive, and a scale on each axis. (The two
space variables would most likely be given using polar coordinates, since the head of
the drum is circular.)
If time is an independent variable, we must establish a time coordinate system by
determining when time zero is. (Of course all times after that are considered positive.)
We must also decide what the time units will be, providing a scale for time.
It may not be clear that there is a coordinate system for the temperature in the cube
of metal, or the amount of salt in the tank. For the temperature, the decision whether
to measure it in degrees Fahrenheit or degrees Celsius is actually the establishing of a
coordinate system, with a zero point and a scale (both of which differ depending on
which temperature scale is used).
The choice of zero for the amount of salt in the tank will be the same regardless of how
it is measured, but the scale can change, depending on the units of measurement.
Once weve established the coordinate system(s) for the variable(s), we should determine
the domain of our function, which means the values of the independent variable(s) for which
the dependent variable will have values. Lets look at some examples.
Example 1.1(h): For Example 1.1(a), suppose that we pull the mass down and then
let it go at a time we call time zero (the origin of our time coordinate system. Time
t is the independent variable, and the values of it for which we are considering the
height of the mass are t 0.
Example 1.1(i): For Example 1.1(b), we will use a position coordinate system consisting of a horizontal number line at the top of where the beam emerges from the wall
(so along the dashed line in the picture), with origin at the wall and positive values (in
feet) in the direction of the beam. Letting x represent the position along the beam,
the domain is 0 x 10.
11
height (inches)
The four functions described in Examples 1.1(a)-(d) are functions of a single variable;
the functions in Examples 1.1(e) and (f) are examples of functions of more than one
variable. The differential equations associated with functions of one variable are called
ordinary differential equations, and the differential equations associated with functions
of more than one variable are (out of necessity) partial differential equations. In this
class we will study primarily ordinary differential equations.
The function in Example 1.1(d) is a mathematical function, whereas the functions from
Examples 1.1(a)-(c), (e) and (f) are not. (We might call them physical functions.) In
your previous courses you have studied a variety of types of mathematical functions, including polynomial, rational, exponential, logarithmic, and trigonometric functions. The main
reason that scientists and engineers are interested in mathematics is that many physical
situations can be mathematically modelled with mathematical functions or equations.
This means that we can find a mathematical function that reasonably well describes the
relationship between physical quantities. For Example 1.1(a), if we let y represent the
height of the mass, then the equation that models the situation is y = A cos(bt), where
A and b are constants that depend on the spring and how far the mass is lifted or pulled
down before releasing it. We will see that the deflection of the beam in Example 1.1(b) can
be modelled with a fourth degree polynomial function, and the amount of salt in the tank
of Example 1.1(c) can be modelled with an exponential function.
Of course one tool we use to better understand a function is its graph. Suppose for
Example 1.1(a) we started the mass in motion by lifting it 1.5 inches and releasing it (with
no upward or downward force). Then the equation giving the height y at any time t would
be of the form y = 1.5 cos(bt), where b depends on the spring and the mass. Suppose that
b = 5.2 (with appropriate units). Then the graph would look like this:
1.5
0.0
0.6
1.8
1.2
time
(seconds)
1.5
When graphing functions of one variable we always put the independent variable (often it will
be time) on the horizontal axis, and the dependent variable on the vertical axis. We can see
from the graph that the mass starts at a height of 1.5 inches above its equilibrium position
(y = 0). It then moves downward for the first 0.6 seconds of its motion, then back upward.
It is back at its starting position every 1.2 seconds, the period of its motion. This periodic
up-and-down motion can be seen from the graph. (Remember that the period T is the
time at which bT = 2, so T = 2
.) Such behavior is called simple harmonic motion,
b
and will be examined in detail later in the course because of its importance in science and
engineering.
Note that even if we didnt know the value of b in the equation y = 1.5 cos(bt) we could
still create the given graph, we just wouldnt be able to put a scale on the time (horizontal)
axis. In fact, we could even create the graph without an equation, using our intuition of
what we would expect to happen. Lets do that for the situation from Example 1.1(c).
Example 1.1(j): A tank contains 100 gallons of water with 10 pounds of salt dissolved in it. At time zero a 0.3 pounds per gallon solution begins flowing into the tank
at 2 gallons per minute and, at the same time, thoroughly mixed solution is pumped
out at 2 gallons per minute. (See Example 1.1(c).) Sketch a graph of the amount of
salt in the tank as a function of time.
12
A
30
10
t
NOTE: We have been using the notation sin(bt) or cos(bt) to indicate the sine or cosine
of the quantity bt. It gets to be a bit tiresome writing in the parentheses every time we
have such an expression, so we will just write sin bt or cos bt instead.
Section 1.1 Exercises
1. Some material contains a radioactive substance that decays over time, so the amount of
the radioactive substance is decreasing. (It doesnt just go away - it turns into another
substance that is not radioactive, in a series of steps. For example, uranium eventually
turns into lead when it decays.)
(a) Give the dependent and independent variables.
(b) Sketch a graph of the amount A of radioactive substance versus time t. Label
each axis with its variable - this will be expected for all graphs.
2. A student holds a one foot plastic ruler flat on the top of a table, with half of the ruler
sticking out and the other half pinned to the table by pressure from their hand. They
then tweak the end of the ruler, causing it to vibrate up and down. (This is roughly
a combination of Examples 1.1(a) and (b).)
(a) Give the dependent and independent variables. (Hint: There are two independent
variables.
(b) Give the domains of the independent variables.
3. Consider the drumhead described in Example 1.1(e). Suppose that the position of any
point on the drumhead is given in polar coordinates (r, ), with r measured in inches
and in radians. Suppose also that time is measured in seconds, with time zero being
when the head of the drum is struck by a drumstick. Give the domains of each of these
three independent variables.
4. Consider the cube of metal described in Example 1.1(f). Suppose that we position the
cube in the first octant (where each of x, y and z is positive), with one vertex
(corner) of the cube at the origin and each edge from that vertex aligned with one of
the three coordinate axes. Each point in the cube then has some coordinates (x, y, z).
Give the domains of each of these three independent variables.
13
5. Consider again the scenario from Section 0.2, in which a rock is fired straight upward
with a velocity of 60 feet per second, from a height of 20 feet off the ground. In that
section we derived the equation
h = 16t2 + 60t + 20
for the ehight h (in feet) of the rock at any time t (in seconds) after it was fired.
Using the equation, determine the domain of the independent variable time.
6. When a solid object with some initial temperature T0 is placed in a medium (like air or
water) with a constant temperature Tm , the object will get cooler or warmer (depending on whether T0 is greater or less than Tm ), with its temperature T approaching
Tm . The rate at which the temperature of the object changes is proportional to the
difference between its temperature T and the temperature Tm of the medium, so it
cools or warms rapidly while its temperature is far from Tm , but then the cooling or
warming slows as the temperature of the object approaches Tm .
(a) Suppose that an object with initial temperature T0 = 80 F is place in a water
bath that is held at Tm = 40 F. Sketch a graph of the temperature as a function
of time. You should be able to indicate two important values on the vertical axis.
(b) Repeat (a) for Tm = 40 F and T0 = 30 F.
(c) Repeat (a) for Tm = 40 F and T0 = 40 F.
7. (a) Suppose that a mass on a spring hangs motionless in its equilibrium position. At
some time zero it is set in motion by giving it a sharp blow downward, and there is
no resistance after that. Sketch the graph of the height of the mass as a function
of time.
(b) Suppose now that the mass is set in motion by pulling it downward and simply
releasing it, and suppose also that the mass is hanging in an oil bath that resists
its motion. Sketch the graph of the height of the mass as a function of time.
8. As you are probably aware, populations (of people, rabbits, bacteria, etc.) tend to grow
exponentially when there are no other factors that might impeded that growth.
(a) The variables in such a situation are time and the number of individuals in the
population. Which variable is independent, and which is dependent?
(b) Using t for time and N for the number of individuals, sketch (and label, of
course) a graph showing growth of such a population.
(c) Often there are environmental conditions that lead to a carrying capacity for
a given population, meaning an upper limit to how many individuals can exist.
Suppose that 500 fish are stocked in a sterile lake (no fish in it) that has a carrying
capacity of 3000 fish. When a population like this starts at well below the carrying
capacity, it experiences almost exponential growth for a while, then the growth
levels off as the population approaches the carrying capacity. Sketch a graph of
the fish population versus time. This sort of growth is called logistic growth.
14
1.2
Performance Criteria:
1. (c) Interpret derivatives in physical situations.
(d) Find functions whose derivatives are given constant multiples of
the original functions.
In the previous sections exercises you graphed the behaviors of some physical systems.
Those graphs are models of the systems themselves; that is, they are human constructed
descriptions of how the systems behave. Such graphical models are good for giving us an
overall qualitative idea of the behavior of a system, but are generally inadequate if we would
like to know precise values of the dependent variable based on a value (or values in the
case of more than one) of the independent variable(s). When we desire such quantitative
information, we attempt to develop an analytical model, which usually consists of an equation
of a function.
Such models for physical situations can often be developed from various principles and
laws of physics. The physical principles do not usually lead us directly to the functions that
model physical situations, but to equations involving derivatives of those functions. This is
because what we usually know is how our variables are changing in relationship with each
other, and such change is modelled with derivatives. Equations containing derivatives are
called differential equations. In this section we will review the concept of a derivative and
see an example of a simple differential equation, along with how it arises.
When you hear the word derivative, you may think of a process you learned in a first
term calculus class. In this course it will be important that you can carry out the process
of finding a derivative; if you need review or practice, see Appendix B. In this section
our concern is not the mechanics of finding derivatives, but instead we wish to recall what
derivatives are and what they mean.
To reiterate what was said in the previous section, a function is just a quantity that
depends on one or more other quantities, one in most cases that we will consider. Again, we
refer to the first quantity (the function) as the dependent variable and the second quantity,
that it depends on, is the independent variable. If we were to call the independent variable
dy
x and the dependent variable y, then you should recall the Leibniz notation
for the
dx
derivative. This notation can be loosely interpreted as change in y per unit of change in x.
Technically speaking, any derivative of a function is really the derivative of the dependent
variable (which IS the function) with respect to the the independent variable. We sometimes
dy
. Obviously it is easier to write y , but that notation
use the notation y instead of
dx
does not indicate what the independent variable is and it does not suggest a ratio, or rate.
Lets consider a couple examples of the meaning of the derivative in physical situations.
Example 1.2(a): Suppose again that we take a mass hanging from a ceiling on a
spring, lift it and let it go, and suppose the equation of motion is y = 1.5 cos 5.2 t. The
dy
derivative of this function is
= 7.8 sin 5.2t, a new function of the independent
dt
variable. This functions value at any time t can be interpreted as how fast the the
15
height of the mass is changing with respect to time, at that particular time. If the
height units are inches and the time units seconds, then the units of the derivative
inches
are seconds
= inches per second, indicating that the derivative of the function y at a
given time is the velocity of the mass at that time. For example, the derivative at time
0.5 seconds is
dy
= y (0.5) = 7.8 sin[(5.2)(0.5)] = 4.02,
dt t=0.5
telling us that the mass is moving downward (indicated by the negative sign) at about
four inches per second at one half second after being set in motion.
Example 1.2(b): Now recall the beam of Example 1.1(b), sticking out from a wall
that it is embedded in. If x represents a horizontal position along the beam and
y represents the deflection (sag) of the beam at that horizontal position, then the
dy
is the change in deflection per unit of horizontal change, which is just
derivative
dx
the slope of the beam at that particular point.
Well now take a break from actual physical situations to ask some questions about
derivatives, in a mathematical sense. After doing so, well see that such questions relate
directly to certain real-life situations.
Example 1.2(c): Find a function whose derivative is seven times the function itself.
Note that the derivative of y = ekt , where k is a constant, is y = kekt . This shows
that exponential functions are essentially their own derivatives, with perhaps a constant
multiplier. If k was seven, the original function would be y = e7t and the derivative
would be y = 7e7t = 7y, seven times the original function y.
Example 1.2(d): Find a function whose second derivative is sixteen times the function itself.
Here we should again be expecting an exponential function, but it will get multiplied twice
because of the chain rule. Note that if y = e4t , then y = 4e4t and y = 16e4t = 16y,
so y = e4t is the function we are looking for. But in fact it is not the function, but only
one such function. The function y = e4t is another such function, as is y = 5e4t .
(You should verify this last claim for yourself.) In fact, y = Ce4t is a solution for any
value of C. We will see later why this is, and what we do about it.
Example 1.2(e): Find a function whose derivative is 16 times the function itself.
The previous example shows that the desired function is not an exponential function, as
the only likely candidates were shown to have second derivatives that are positive sixteen
times the original function. What we want to note here is that if we take the derivative of
sine or cosine twice, we end up back at sine or cosine, respectively, but with opposite sign.
16
However, each time we take the derivative of a sine or cosine of kx, the chain rule gives
us a factor of k on the outside of the trig function. Thus we see that
y = sin 4x
y = cos 4x
y = 4 cos 4x
y = 4 sin 4x
=
=
y = 16 sin 4x = 16y
y = 16 cos 4x = 16y
This shows that y = sin 4x and y = cos 4x are functions whose derivatives are 16 times
the original functions themselves.
Consider Example 1.2(e) above. The words the second derivative is 16 times the
d2 y
= 16y, since the function is the
original function can be written symbolically as
dt2
dependent variable y. This is a differential equation, an equation containing a derivative.
(Differential equations can contain derivatives of any order. The order of a differential
equation is the highest order derivative occurring in the differential equation, so this is a
second order differential equation.)
Now consider the following physical situation: One end of a spring is attached to a ceiling,
as shown to the left below. We then hang an object with mass m (we will refer to both
the object itself and its mass as the mass - one must note from the context which we
are talking about) on the spring, extending it by a length l to where the mass hangs in
equilibrium. This is shown in the center picture below. There are two forces acting on the
mass, a downward force of mg, where g is the acceleration due to gravity, and an upward
force of kl, where k is the spring constant, a measure of how hard the spring pulls
back when stretched. The spring constant is a property of the particular spring. When
the mass hangs in equilibrium these two force are equal in magnitude to each other, but in
opposite directions.
y (pos)
F = kl
F = mg
l
y
F = k(l y)
F = mg
We will put a coordinate system (with a scale in appropriate length units, like inches) beside
the mass, with the zero at the point even with the top of the mass at rest and with the
positive direction being up. If we then lift the mass up to a position y0 , where y0 < l, and
release it, it will oscillate up and down. If we assume (for now) that there is no resistance,
it will oscillate between y0 and y0 forever; as noted in the previous section, this is
called simple harmonic motion. Consider the mass when it is at some position y in this
oscillation, as shown above and to the right. There will be an upward force of k(l y) due
to the spring and a downward force of mg (the negative indicating downward) due to
gravity. Remembering that force is mass times acceleration and that acceleration is the
17
second derivative of position with respect to time, the net force is then
F = ma = m
d2 y
= k(l y) mg = kl ky mg = ky,
dt2
since kl = mg.
d2 y
Extracting the equation m 2 = ky from the above and dividing both sides by
dt
k
k
d2 y
= y. If the values of k and m are such that
= 16, this equation
m gives
2
dt
m
m
d2 y
= 16y, the equation describing the situation from Example 1.2(e)! If we
becomes
dt2
could find a function y that satisfies the differential equation (more about what this means
in Section 1.5), it will model the motion of the mass as it oscillates up and down. This
shows that what seems like a whimsical mathematical question about derivatives (posed in
Example 1.2(e)) is actually very relevant for a practical application.
Section 1.2 Exercises
1. Find the derivative of each function without using your calculator. You MAY use the
course formula sheet. Give your answers using correct derivative notation.
(a) y = 2 sin 3x
(b) y = 4e0.5t
(e) y = te3t
(c) x = t2 + 5t 4
2. Find the second derivatives of the functions from parts (a)-(c) of Exercise 1. Give your
answers using correct derivative notation.
3. The temperature T of an object (in degrees Fahrenheit) depends on time t, measured
dT
= 2.7 when t = 7. Interpret the derivative in a sentence, using
in minutes, and
dt
either increasing or decreasing.
4. The amount A of salt in a tank depends on the time t. If A is measured in pounds
dA
and t is measured in minutes, interpret the fact that
= 1.3 when t = 12.5.
dt
Again, use increasing or decreasing.
5. The height of a mass on a spring at time t is given by y, where t is in seconds and
y is in inches.
dy
= 5 when t = 2.
dt
d2 y
= 3 when t = 2.
(b) Interpret the fact that
dt2
(c) Is the mass speeding up or slowing down at time t = 2? Explain.
(a) Interpret the fact that
7. For this exercise, consider the beam of Examples 1.1(b) and 1.2(b).
(a) Will the value of the derivative
x > 0?
dy
dx
(b) Suppose that 0 x1 < x2 10. Which is greater, the absolute value of the
derivative at x1 , or the absolute value of the derivative at x2 ?
8. (a) Find a function y(x) whose derivative is 3 times the original function. Is there
more than one such function? If so, give another.
(b) Find a function y(t) whose second derivative is 9 times the original function.
Is there more than one such function? If so, give another.
(c) Find a function x(t) whose second derivative is 9 times the original function.
Is there more than one such function? If so, give another.
(d) Find a function y(x) whose second derivative is 5 times the original function.
Is there more than one such function? If so, give another.
9. For each of the situations in Exercises 6, write a differential equation whose solution
is the desired function. (See the paragraph after Example 1.2(e).) Use the given
independent and dependent variables, and give your answers using Leibniz notation.
19
1.3
If you have not recently read the explanation of the spring-mass system at the end of the
last section, you should probably skim over it again before reading this section. Recall that
for the spring-mass system, the independent variable is time and the dependent variable is
the height of the mass. Assuming no resistance, once the mass is set in motion, it will exhibit
periodic oscillation (simple harmonic motion). It should be intuitively clear that changing
either the amount of the mass or the stiffness of the spring (expressed by the spring constant
k) will change the period of oscillation. The mass m and the spring constant k are what
we call parameters, and they should not be confused with the variables, which are time
and the height of the mass. Parameters will show up in three places:
As characteristics of the physical systems themselves, quantified by numerical values.
d2 y
= 8y.
dt2
Here we see the two parameters showing up in the differential equation. If we multiply both
sides by two and subtract the right side from both sides we obtain
d2 y
+ 16y = 0,
dt2
k
, which we often rename as 2 .
where the 16 is the new parameter m
2 = 16 sec1 2 . The most general solution to this equation is
In this case
y = C1 sin t + C2 cos t;
the variables are t and y, and C1 , C2 and
r ! are parameters. The parameter depends
k
and the parameters C1 and C2 depend
on the mass and spring constant
=
m
on how the mass is set in motion, what we will call initial conditions. In Sections 1.4 and
1.7 we will see the significance of C1 and C2 and how they are determined.
We now consider the horizontal beam of Example 1.1(b). One might guess that some
parameters that determine the amount of deflection of the beam would be the material the
beam is made of, the thickness and shape of the beam (square, I-beam, etc.), the length
of the beam, and perhaps other things.
20
Example 1.3(a): The differential equation, and its solution, for the beam of Example
1.1(b) are
w 4
d4 y
=w
and
y=
x + c3 x3 + c2 x2 + c1 x + c0 ,
4
dx
24EI
where E is Youngs modulus of elasticity of the material the beam is made of, I is the
cross-sectional moment of inertia of the beam about the neutral axis, and w is the
weight per unit of length. Give the variables and parameters for both the differential
equation and the solution.
EI
We can see from the derivative in the differential equation that the independent variable is
x and the dependent variable is y. The remaining letters all represent parameters: the
modulus of elasticity E, the cross-sectional moment of inertia I, and the weight per unit
of length w. In the solution we see these parameters again, along with four others, c0 , c1 ,
w
c2 and c3 . We can also think of the quantity 24EI
as a single parameter if we wish.
The last four parameters in the solution will depend on the length of the beam and how it
is supported, in this case by being embedded in the wall at its left end and having no support
at the right end. These things are what are called boundary conditions. Well discuss
them some more in the next section, and look at specific situations involving boundary
conditions in Chapter 5.
In summary, parameters are variables that change from situation to situation, but once
the situation is determined the values of the parameters are constant. At that point, the only
things that change are the variables. In this course we will never again refer to parameters
as variables, and we will consider them distinct from the variables of interest.
Section 1.3 Exercises
1. As mentioned previously, when a solid object with some initial temperature T0 is
placed in a medium (like air or water) with a constant temperature Tm , the objects
temperature T will approach Tm as time goes on. The rate at which the temperature
of the object changes is proportional to the difference between its temperature T and
the temperature Tm of the medium, giving us the differential equation
dT
= k(Tm T ),
dt
where k is a constant dependent on the material the object is made from.
(a) Keeping in mind that parameters are quantities that vary from situation to situation but do not change once the situation is fixed, give all of the parameters.
(b) Give the independent variable(s).
(c) Give the dependent variable.
2. Suppose that a mass on a spring hangs motionless in its equilibrium position. At some
time zero it is set in motion by pulling it downward and simply releasing it, and suppose
also that the mass is hanging in an oil bath that resists its motion. The independent
variable is time, and the dependent variable is the height of the mass. Give as many
physical parameters as you can think of for this situation - there are three or four that
occur to me.
21
3. When dealing with certain electrical circuits we obtain the differential equation and
solution
R
di
E
E
L + Ri = E
e L t .
and
i = + i0
dt
R
R
Give the independent variable, dependent variable, and all the parameters.
4. At some time a guitar string is plucked, and the dependent variable that we are interested in is the displacement of the string from its initial position.
(a) What is(are) the independent variable(s)?
(b) What are some physical parameters of importance?
22
1.4
Performance Criterion:
1. (f) Identify initial value problems and boundary value problems. Determine initial or boundary conditions.
spring
mass
deflection is
exaggerated
and
EI
d4 y
= w,
dx4
and
y=
w 4
x + c3 x3 + c2 x2 + c1 x + c0 ,
24EI
where C1 , C2 , c3 , c2 , c1 and c0 are arbitrary (meaning they can have any values) constants
differing from, and not depending on, the parameters k, m, E, I and w. (Remember
that we are case sensitive in mathematics, science and engineering, so C1 and c1 are not
necessarily the same value.) Note that the number of such arbitrary constants in the solution
of a differential equation is equal to the order of the differential equation. (Again, the order
of a differential equation is the highest order derivative in the differential equation - more
on this in Section 1.6.)
The height of the mass at any time t depends on the amount of the mass and the
stiffness of the spring (given quantitatively by the spring constant k), but it also depends
on how we set the mass in motion. It can be lifted or pulled down and let go, or given a
blow, or some combination of these things. The combination that sets it in motion are what
are called initial conditions. Suppose that we set the mass in motion by simply pulling
23
it down by two units and then letting it go. Calling the moment we let it go time zero, we
would have that
dy
=0
y = 2
and
dt
at time zero. (Remember that the derivative is the velocity, so the second statement says
that the mass has zero velocity at the moment we let it go.) Using function notation and
the fact that the derivative is the function y , this is usually expressed by
y (0) = 0.
y(0) = 2,
The two numbers 2 and 0 are called initial values, a term we will use interchangeably
with initial conditions, even though the concepts are slightly different. We will see later how
these two pieces of information can be used to determine the values of the constants C1 and
C2 in the solution y = C1 sin t + C2 cos t.
Lets now think about the horizontal beam. The independent variable is x, the horizontal
distance along the beam, measured from the wall. Time is not a variable at all; the beam
deflects immediately when put into place, then retains its displacement from then on. The
deflection, though, is dependent on what is going on at the two ends of the beam. At the left
end the beam is what we call embedded. The effect of this is two things: the displacement
of the beam is zero at that point, and the slope of the beam is zero right where it leaves the
wall. We can express these two things by
y(0) = 0
and
y (0) = 0,
which are boundary conditions. The right end of the beam is free, which is described
mathematically by the boundary conditions
y (10) = 0
and
y (10) = 0.
Well discuss the origin of these two conditions a bit more in Chapter 5. Altogether we have
four boundary conditions
y(0) = 0,
y (0) = 0,
y (10) = 0,
y (10) = 0
w 4
x + c3 x3 + c2 x2 + c1 x + c0 .
24EI
The numerical values of zero for all these derivatives are boundary values. As with initial
values/initial conditions, we will blur the distinction between boundary values and boundary
conditions. Lets now look at some more examples of initial and boundary conditions.
Example 1.4(a): Consider the mass on the spring, set in motion by lifting it one
inch and letting it go. Give the height and velocity of the mass at the time it is let go,
using function notation.
Taking up to be positive, at time zero (the moment we set the mass in motion) the height of
the mass is one inch, so we write y(0) = 1. Since we simply release the mass at time zero,
the velocity at time zero is zero. Recalling that velocity is the first derivative of position,
we can describe this by y (0) = 0. The initial conditions are then y(0) = 1, y (0) = 0.
24
Example 1.4(b): Consider the mass on the spring, this time setting it in motion
by hitting it downward at three inches per second from its position at rest. Give the
initial conditions for the height function y.
Because we are forcing the mass from its position at rest, its initial height is zero. This
is given using function notation by y(0) = 0. The fact that it has downward velocity of
three inches per second at time zero gives us the initial condition y (0) = 3.
Example 1.4(c): Suppose that the mass is set in motion by pulling it down two
inches, then giving it an upward velocity of five inches per second to begin. Give the
initial conditions for the height function y.
The initial conditions are y(0) = 2 and y (0) = 5.
Example 1.4(d): Consider a twenty foot beam
that is embedded in walls at both ends, as shown
to the right. The beam will deflect downward
some in the middle; the deflection is exaggerated
in the picture. Give the boundary conditions for
the beam.
The boundary conditions are y(0) = 0, y (0) = 0, y(20) = 0 and y (20) = 0.
Situations in which position along a line is the independent variable will have boundary
conditions.
Situations where a function depends on both position and time will have both initial
conditions and boundary conditions. We will not see these, because they are described
by partial differential equations.
Partial differential equations are also required when working with boundary conditions
only, when the function of interest is a function of more than one space variable. Such
functions would arise when dealing with sheets or solids, rather than beams, which can
be thought of as one-dimensional lines.
We will work primarily with initial conditions, but you will see boundary conditions later in
the course (see Chapter 5).
Section 1.4 Exercises
1. In each of the following, the independent variable is given for a situation (the dependent
variable should be clear), along with initial or boundary conditions, in function form.
For each, give every initial or boundary condition in the form variable = number when
variable = number.
25
y(0) = 7, y (0) = 3
x(0) = 1, x (0) = 5
2. For each of the following, give the initial conditions for a mass on a spring that is set
in motion in the way described. Give the conditions using function notation, as done
in Examples 1.4(a) and 1.4(b). Let the dependent variable in each case be y.
(a) The mass is pulled down five units and let go with no initial velocity.
(b) The mass is not displaced, but it is given an downward velocity of two units per
second.
(c) The mass is lifted by one unit and given an upward velocity of two units per
second.
(d) The mass is pulled down by three units and given an upward velocity of one unit
per second.
3. If a mass on a spring is set in motion are there is no resistance to its vibration, it will
oscillate in the same manner forever. (Resistance to its motion we will call damping,
and well study its effect in Chapters 3 and 4.) Assuming such conditions, sketch the
graph of the displacement of the mass at any time t for each of the sets of initial
conditions listed in Exercise 2. Extend your graph far enough to show at least two full
periods. You will not be able to label a scale on the horizontal axis, but for three of
the cases you should be able to label the vertical axis with a scale. Take care to make
sure that the graph has the correct slope where it leaves the vertical axis.
4. There is one other condition (besides embedded or free) well see at the end of a beam,
called simply supported or pinned. This means that the end is supported but
allowed to pivot freely. In that case the displacement is zero, and the second derivative
of displacement is zero. For each of the following scenarios, give the boundary conditions
for the beam, assuming a dependent variable of y.
(a) A 20 foot beam that is simply supported at its left end and embedded at its right
end.
(b) A 12 foot beam that is simply supported at both ends.
5. Suppose that we have a 70 centimeter metal rod that is perfectly insulated along the
length of the rod, so that no heat can enter or leave along its length, but heat CAN
enter or leave at its ends. We then put the rod horizontally in front of us and consider
a coordinate system that puts zero at the left end of the rod and 70 cm at the right
end, and we let u(x) represent the temperature at any point x along the length of
the rod, using our coordinate system.
Suppose also that we hold an ice cube (temperature 32 Fahrenheit at the left end and
a hair dryer blowing 115 F air on the right end. Because the independent variable is
a space variable x, this situation has boundary conditions. Give them, using function
notation.
26
1.5
Performance Criteria:
1. (g) Determine the independent and dependent variables for a given
differential equation.
(h) Determine whether a function is a solution to an ordinary differential equation (ODE); determine values of constants for which
a function is a solution to an ODE.
dy
+ 3y = 0
dx
Equation 3: y + 9y = 26e2t
Equation 5:
x
dy
=
dx
y
Equation 2: y + 3y + 2y = 0
d4 y
= 1.4
dx4
2u 2u
u
Equation 6:
+ 2 =
2
x
y
t
Equation 4: 15.3
Note that equations 1, 2, 3 and 5 contain not only derivatives of the function y, but the
function itself as well. (We can really think of the function as the zeroth derivative.)
The first five of these equations are all ordinary differential equations, meaning
that they contain ordinary derivatives, which are appropriate when there is only one
independent variable. The last one contains partial derivatives (which are written with the
symbol instead of d) and is called a partial differential equation. (Some of you
may have not yet taken a course in which you learn about partial derivatives.) We often
use the abbreviations ODE for ordinary differential equation and PDE for partial differential
equation.
The order of a differential equation is the order of the highest derivative in the equation.
Equations 1 and 5 above are first order, Equations 2, 3 and 6 are second order, and Equation 4
is fourth order. In this course we will focus almost entirely on ordinary differential equations,
and most of the equations we will work with will be first or second order.
When looking at a differential equation, it is often possible to determine the independent
and dependent variables of interest. Derivatives are always of the dependent variable, and
with respect to the independent variable (or one of the independent variables in the case of
a function of more than one variable). So for Equation 1, the dependent variable is y and
the independent variable is x.
Example 1.5(a): Give the dependent and independent variables for the rest of the
equations.
For Equations 4 and 5 the dependent variable is y and the independent variable is x. For
equation 3 the dependent variable is y, and since the derivative is an ordinary derivative
there must be only one independent variable, and it has to be t, the only other variable
visible in the equation. The dependent variable in Equation 2 is y, and it is not possible
to determine the independent variable in that case. Lastly, u is the dependent variable
27
in Equation 6. There are three independent variables, x, y and t, which is why partial
derivatives are required. Any situation with more than one independent variable will
result in a partial differential equation.
What the above shows us is that a solution to an algebraic equation is a number that, when
substituted for the unknown value, makes the equation true. We should recall that some
equations have more than one solution. For example, both 3 and 3 are solutions to the
equation x2 9 = 0.
In the case of a differential equation, a solution to the equation is NOT a number, it is
a function.
There are some differential equations whose solutions are relations rather than functions;
well solve a few of those, but for all of the applications we will consider, the solutions to the
ODEs modelling the situations will be functions.
When asked to verify that, or determine whether, a function is a solution to an ODE,
you need to show some work supporting whatever your conclusion is. The following example
shows one way to do this.
Example 1.5(b): Show that y = 5 cos 4t is a solution to
d2 y
= 16y.
dt2
We compute the left hand side (LHS) and right hand side (RHS) separately:
dy
= 20 sin 4t
dt
LHS =
d2 y
= 80 cos 4t
dt2
d2 y
= 16y.
Because LHS = RHS, y = 5 cos 4t is a solution to
dt2
28
For the above example the left hand side was just one derivative. When the left hand
side is more complicated, a standard method of verifying a solution is to first calculate any
derivatives that appear on the left hand side of the equation, then substitute them into the
left hand side. If the right hand side is fairly simple, we might be able to simplify the left
side directly to the right hand side, as done in the next example.
Example 1.5(c): Determine whether y = Ce2t , where C is any constant, is a
solution to the differential equation y + 3y + 2y = 0.
First we see that y = Ce2t (2) = 2Ce2t and y = 2Ce2t (2) = 4Ce2t , so
This last example shows that a differential equation can have an infinite number of solutions
(since C can be any real number), and well see the same thing in the next example as well.
Example 1.5(d): Verify that y = C1 sin 3t + C2 cos 3t, where C1 and C2 are any
constants, is a solution to y + 9y = 0.
First we see that y = 3C1 cos 3t 3C2 sin 3t and y = 9C1 sin 3t 9C2 cos 3t.
Therefore
LHS = (9C1 sin 3t 9C2 cos 3t) + 9(C1 sin 3t + C2 cos 3t) = 0 = RHS,
In this last example the function y = C1 sin 3t + C2 cos 3t is a solution regardless of the
values of the parameters C1 and C2 . Because C1 and C2 can take any values, we say they
are arbitrary constants. We will often use the lower case c and upper case C for arbitrary
constants, sometimes with subscripts like above. We call all the functions obtained by letting
the constants take different values a family of solutions for the differential equation. The
solution to every first order equation will contain a constant that can take on infinitely many
values, and solutions to second order equations contain two arbitrary constants, as in the
above example. This may seem to contradict the result of Example 1.5(c), but the most
general solution in that case is y = C1 e2t + C2 et ; the solution verified in that example
is for the case in which C2 = 0. The fact that C1 and C2 are subscripted differently
means that they are probably, but not necessarily, different constants. Other letters will
occasionally be used as constants.
In this next example you will see a situation where a function is a solution only when the
parameter takes a certain value; in this case the constant (parameter) is NOT arbitrary.
Example 1.5(e): Determine any values of C for which y = Ce2t is a solution to
the differential equation y + 9y = 26e2t .
The derivatives of the given function are y = 2Ce2t and y = 4Ce2t . Substituting
the second derivative into the left hand side of the ODE gives
LHS = 4Ce2t + 9Ce2t = 13Ce2t .
29
y = Ce2t is a solution only if LHS = RHS, which requires that 13C = 26. Therefore
y = Ce2t is a solution to the differential equation y + 9y = 26e2t only when C = 2.
(a)
+
+
2y = 0
(b) y y = 0
(c)
=3
dx
t2
x21 x22 x23
(d) y + 9y = 26e2t
(g) L
d2 u
+ g sin u = 0
dx2
(e)
u
2u
= 0.5 2
t
x
(h) EI
d4 y
=w
dx4
d2 x
dx
5 + 6x = 10 sin t
2
dt
dt
2
2u
u 2 u
2
=0
(i)
c
+
t2
r 2
r r
(f)
2. (a) For part (b) of the previous exercise you should not have been able to identify the
independent variable. Given that the solution is y = 3ex 5ex , what is the
independent variable?
30
dy
y = 4e3x .
dx
(a) Is there any value of c for which y = cex a solution to the equation? If so, what
is the value?
(b) Is there any value of c for which y = ce3x a solution to the equation? If so,
what is the value?
(c) Recall that a solution to a differential equation that cannot have an arbitrary
constant in it is called a particular solution to the equation. Give a particular
dy
solution to the differential equation
y = 4e3x .
dx
dy
y = 0?
(d) Is there any value of c for which y = cex a solution to the equation
dx
If so, what is the value?
10. Consider the ODE y + 3y + 2y = 0.
(a) Because exponential functions are their own derivatives, it is conceivable that
y = erx is a solution for some value of r. Substitute it into the ODE.
(b) You should be able to factor er t out of the left side of your result from (a). Now
eu is not zero for any value of u. What does this imply about our situation?
(c) Your answer to (b) should be an equation. Solve to determine what value(s)
r must have in order for y = ert to be a solution.
(d) You now should have at least one solution to the ODE. Check all solutions.
11. Repeat Exercise 10 for the ODE 2y + 3y + y = 0.
31
1.6
There are many different classifications and types of differential equations; we will focus
on just a few classifications here. Lets consider the following examples, most of which we
saw in the previous section.
Equation 1:
dy
+ 3y = 0
dx
Equation 3: y + 9y = 26e2t
Equation 5:
x
dy
=
dx
y
Equation 2: x2 y + xy + x2 y = 0
d4 y
= 1.4
dx4
2u 2u
u
Equation 6:
+ 2 =
2
x
y
t
Equation 4: 15.3
dn y
dn1 y
d2 y
dy
+
a
(x)
+
+
a
(x)
+ a1 (x) + a0 (x)y = f (x),
n1
2
n
n1
2
dx
dx
dx
dx
(1)
where a0 (x), ..., an (x) are functions of x (possibly constants), is called a linear
ordinary differential equation. Equations 1 through 4 are linear ODEs:
Equation 1 is first order linear, with a1 (x) = 1, a0 (x) = 3 and f (x) = 0.
32
dy x
=0
dx y
B. y 2xy = x
C. y + 2y = y 2
D. 5
dy
3y = sin x
dx
Example 1.6(a): Determine the functions F (x, y) for Equations A, C and D above.
Each of the equations can be solved for
A.
dy
x
=
dx
y
C.
dy
to get
dx
dy
= y 2 2y
dx
D.
dy
3
1
= y + sin x
dx
5
5
We can see now that the functions F for the three equations are
A. F (x, y) =
x
y
C. F (x, y) = y 2 2y
3
1
D. F (x, y) = y + sin x
5
5
We can now define two other categories of first order differential equations.
33
dy x
=0
dx y
B. y 2xy = x
C. y + 2y = y 2
D. 5
dy
3y = sin x
dx
We will see the significance of separable and autonomous equations later. For now we
should note a bit of algebra that can be performed with a separable equation. To begin
dy
as being the quotient of the two differentials
with, we need to think of the derivative
dx
dy and dx. Treating each like we would a variable, when we are working with a separable
equation we can get all the x stuff on one side of the equation and the y stuff on the
other side:
dy x
= 0
dx y
dy
x
=
dx
y
x
dy =
dx
y
y dy = x dx
Separable equations are often easy to find solutions for, as well do in Chapter 2, and computations like the above will be part of the process.
34
(a)
2y = 0
(b) y y = 0
(c)
=3
+
+
dx
t2
x21 x22 x23
(d) y + 9y = 26e2t
(g) L
d2 u
+ g sin u = 0
dx2
u
2u
= 0.5 2
t
x
4
dy
(h) EI 4 = w
dx
(e)
(f)
d2 x
dx
5 + 6x = 10 sin t
2
dt
dt
2. Give the order of each of the following ordinary differential equations. Assume that
any letters not used in derivatives represent constants.
(a)
dy
2y = 0
dt
(d) y + 9y = 26e2t
(g) L
d2 u
+ g sin u = 0
dx2
1 dy
+y =1
y dx
(b) y y = 0
(c)
1 dy
+y =1
x dx
d4 y
(h) EI 4 = w
dx
d2 x
dx
5 + 6x = 10 sin t
2
dt
dt
dy
(i)
+ xy = 1
dx
(e)
(f)
3. For each of the first order equations from Exercise 2, give the function F if the the
dy
= F (x, y). (Use the appropriate variables
equation was to be written in the form
dx
for the equation.)
4. For each of the ODEs from Exercise 2 that are linear, give the values of the functions
f , a0 , a1 , a2 , ... (Include the independent variable, like a1 (x) = x2 , for example.) If
the independent variable cannot be determined, use x.
5. For each of the first order equations from Exercise 2 that are separable, give the functions g and h, using the appropriate independent variable.
6. Which of the first order equations from Exercise 2 are autonomous?
35
1.7
dy
+ 3y = 0, for which any function of the form
dx
y = Ce3x is a solution. Suppose we impose the additional condition that y = 4 when
x = 0. This is called an initial condition and we often write such a condition in the form
y(0) = 4. Substituting these values into y = Ce3x gives 4 = Ce3(0) , leading to C = 4.
When we combine a differential equation with one or more initial values, we have what
is called an initial value problem (IVP). The solution to an initial value problem is a
function or equation that satisfies both the differential equation and the initial value(s).
Thus y = 4e3x is a solution to the IVP
Consider again the differential equation
dy
+ 3y = 0 ,
dx
y(0) = 4
Example 1.7(a): Verify that y = 27 e5t + 52 sin t 21 cos t is the solution to the
initial value problem
dy
+ 5y = 13 sin t,
y(0) = 3
dt
dy
e5t + 25 cos t + 12 sin t, so
= 35
2
dt
LHS = 35
e5t + 25 cos t + 12 sin t + 5
2
e5t + 25 cos t + 12 sin t +
= 35
2
=
26
2
7 5t
e
2
35 5t
e
2
+ 25 sin t 21 cos t
25
2
sin t 52 cos t
sin t
= 13 sin t
= RHS
This shows that the function satisfies the differential equation.
We must now show that the function satisfies the initial condition. When t = 0,
y = 72 e5(0) + 25 sin 0 21 cos 0 =
7
2
+0
1
2
6
2
= 3,
so the function satisfies the initial condition also. Therefore it is a solution to the IVP.
We should observe in the above example that the process of checking the initial condition is
easier than checking the differential equation; this is often the case.
36
Lets reiterate that in order to be a solution to an IVP, the function must satisfy BOTH the
ODE and the initial conditions. Since the function in this last example failed to satisfy one
of the initial conditions, it doesnt matter whether it satisfies the ODE or not, it still fails
to satisfy the IVP.
We will now see how initial values or boundary values are used to determine the values
of arbitrary constants for solutions to ODEs for which we also know initial or boundary
conditions.
Example 1.7(c): It can be shown that
x = C1 et + C2 e3t + 2 sin t
(1)
x (0) = 5.
(2)
Next we compute the derivative of (1) to get x = C1 et 3C2 e3t +2 cos t. Substituting
t = 0 and x = 5 into this gives us
5 = C1 3C2 + 2.
(3)
37
x(0) = 2, x (0) = 5.
dy
= 1,
dx
y(1) = 2
y(0) = 1,
y (0) = 4
3. For each of the following, determine whether the given function is a solution to the
initial value problem that is given after it. If it is not, tell why not.
(a) y = 41 ex + 21 e2x + 43 e3x
(c) y = 25 ex
1
2
IVP:
dy
2xy = x, y(0) = 2
dx
IVP:
d2 x
+ 4x = 0, x(0) = 3, x (0) = 4
dt2
4. (a) y = Ce2t +3 cos 2t is the general solution to the ODE y +2y = 6 cos 2t6 sin 2t.
Determine the solution to the initial value problem
y + 2y = 6 cos 2t 6 sin 2t,
y(0) = 5.
d2 x
dx
(b) x = C1 et + C2 e4t + 3t + 1 is the general solution to the ODE
+
5
+ 5x =
dt2
dt
12t + 19. Find the solution to the IVP
dx
d2 x
+ 5 + 5x = 12t + 19,
2
dt
dt
x(0) = 2, x (0) = 1.
y(0) = 3, y (0) = 32 .
(d) y = C1 sin 2t+C2 cos 2t+e3t is the general solution to the ODE y +4y = 14e3t .
Find the solution to the initial value problem
y + 5y = 0 ,
y(0) = 3, y (0) = 32 .
38
1.8
Chapter 1 Summary
For our purposes, a function is a dependent variable that depends on one or more
independent variables. (For most of this course we are concerned only with functions
of one independent variable.)
The values of the independent variable for which values of the dependent variable are
obtained are called the domain of the function.
The graph of a function gives us a quick way to determine the general behavior (sometimes called the qualitative behavior) of the function.
The derivative of a function gives the rate of change of the dependent variable with
respect to the independent variable.
Exponential functions are essentially their own derivatives (of any order). Sine and cosine are essentially their own second derivatives. Essentially means that the function
and the derivative differ only by a factor of (multiplication by) a constant.
Differential equations are equations containing derivatives of a function. When the
function is a function of one variable, the differential equation is an ordinary differential
equation; when the function is a function of more than one variable, the differential
equation is a partial differential equation.
Parameters are values that change from situation to situation, but that do not change
once the situation is set. Variables are values that change once the situation is set.
When studying a situation in which time is the independent variables, we generally
have initial conditions, which describe the state of the dependent variable at time zero
(or some other point in time).
When studying a situation in which position in space (or along a rod, on a surface)
is(are) the independent(s), we have values of the dependent variable on the boundary
of our domain. These are called boundary values.
Even though they are slightly different, we use initial values and initial conditions
synonymously, and the same for boundary values and boundary conditions.
A solution to a differential equation is a function for which the function and its relevant
derivatives can be substituted into the equation to obtain a true statement.
When a solution contains arbitrary constants we call it a family of solutions. A family
that includes all possible solutions to a differential equation is called a general solution;
a solution that contains no arbitrary constants is called a particular solution.
General solutions to first order differential equations contain one arbitrary constant, and
general solutions to second order differential equations contain two arbitrary constants.
To verify that a function is a solution to a differential equation, we substitute the
function and its derivative into the left side and see if the result is the right side, OR
we substitute the function and its derivatives into both sides and see if the results are
equal.
We classify ordinary (and partial) differential equations:
The order of a differential equation is the order of the highest derivative in the
equation.
39
40
1.9
Chapter 1 Exercises
deflection is
exaggerated
(a) The original equation can be written y (4) = 2. Given that y (4) is the derivative
of y (3) = y , what must y (3) look like? Your answer should include a constant.
(b) Use the boundary condition y (10) = 0 to determine the constant from your
answer to (a). Write the function y with that value substituted into your
answer to (a).
(c) Now that you know y , you can use the fact that it is the derivative of y to
find what y looks like. It contains a constant - use the boundary condition
y (10) = 0 to find the value of that constant. Then give y .
(d) Repeat to find y , and then y.
2. In this exercise you will solve what is perhaps the easiest boundary value problem there
is. Remember the situation from Exercise 5 of Section 1.4: A 70 centimeter metal rod
is perfectly insulated along the length of the rod, so that no heat can enter or leave
along its length, but heat CAN enter or leave at its ends. We impose a one-dimensional
coordinate system that puts zero at the left end of the rod and 70 cm at the right end,
and we let u(x) represent the temperature at any point x along the length of the
rod, using our coordinate system. The temperature at the left end of the rod is held
at a constant temperature of 32 Fahrenheit and the right end is held at 115 F.
We now leave the rod in this state for a very long (infinite) period of time. The
temperature at each point in the rod will eventually reach a constant value, called its
equilibrium temperature. Letting u(x), for 0 x 70, represent the function
giving the equilibrium temperature at every point in the rod. Physical principles of
heat flow dictate that the function u must satisfy the following BVP:
d2 u
= 0,
dx2
d2 u
= 0 is called the one-dimensional Laplaces equation or the
dx2
steady-state heat equation, and u(0) = 32 and u(70) = 115 are associated
boundary values. Lets solve the boundary value problem!
The equation
(a) Draw a graph of what you think the equilibrium temperatures will look like. As
we will always do, put the independent variable x on the horizontal axis and
the dependent variable u (the temperature) on the vertical axis. Label each axis
with its variable and the units for that variable.
(b) The ODE can be written as u = 0. Remembering that u is the derivative
of u , what sort of function must u be if it has a derivative of zero? Write an
equation for u .
41
(c) What must u look like in order to have the derivative found in (b)? Write an
equation for u - it should contain two arbitrary constants.
(d) Substitute the first boundary condition into your answer to (b) to get an equation
containing both arbitrary constants but neither of the variables x or u. Repeat
for the second boundary condition.
(e) You now have a system of two equations containing the two unknown constants.
Solve the system to find their values.
(f) Give the function u that is the solution to the BVP. Would the graph of this
function look like what you drew for part (a)?
NOTE: When we have a similar problem, but for temperatures in a two-dimensional plate
of metal rather than a rod, the differential equation will be the two-dimensional Laplaces
equation
2u 2u
+
= 0.
x2 y 2
This is a partial differential equation - we can tell by the facts that the derivatives are partial
derivatives and that there are two independent variables, x and y. You should be able to
guess what the three-dimensional Laplaces equation would look like.
If we were to, instead of waiting for the temperatures to reach equilibrium in our rod,
watch the temperature as time progressed form time zero, then time itself would be an
independent variable as well. The temperature u(x, t) at any point and time in the rod is
then a function of two variables, and it obeys the one-dimensional heat equation
u
2u
= 2.
t
x
Here the constant is a parameter that depends on various properties of the material the
rod is made of. If our object was three dimensional, the corresponding heat equation would
be
2
u
u 2u 2u
=
+
+ 2 .
t
x2 y 2
z
When we work with Laplaces equation or the heat equation in more than one dimension, the
boundary value situation becomes more complicated, as there are infinitely many boundary
points. When working with the heat equation (in any number of dimensions) there will also
be infinitely many initial values. In all of these cases the boundary values and initial values
are given by functions rather than constants (unless the functions happen to be constant
functions!).
There are three important classes of partial differential equations, called elliptic, parabolic
and hyperbolic. Laplaces equation and the heat equation are the standard examples of elliptic and parabolic equations, respectively. The standard example of a hyperbolic equation
is the wave equation. In three dimensions the wave equation is usually written as
2
2u
u 2u 2u
2
=c
+
+ 2 ,
t2
x2 y 2
z
where c is a constant.
42
In the first chapter you found out what what ordinary differential equations (ODEs),
initial value problems (IVPs) and boundary value problems (BVPs) are, and what it means
for a function to be a solution to an ODE, IVP or BVP. You then saw how to determine
43
whether a function is a solution to an ODE, IVP or BVP, and we looked at a few real
world situations where ODEs, BVPs and IVPs arise from physical principles.
Our goal for the rest of the course is to solve ODEs, IVPs and BVPs and to see how the
ODEs, IVPs, BVPs and their solutions apply to real situations. We can solve ODEs (and
PDEs) in three ways:
Analytically, which means paper and pencil methods that give exact solutions in the
form of algebraic equations.
Qualitatively, which means determining the general behavior of solutions without actually finding function values. Results of qualitative methods are often expressed graphically.
Numerical methods which result in values of solutions only at discrete points in time
or space. Results of numerical methods are often expressed graphically or as tables of
values.
In this chapter you will learn how to find solutions qualitatively and analytically for first
order ODEs and IVPs. You will see two analytical methods, separation of variables and
integrating factors.
Separation of variables is the simpler of the two methods, but it only works for separable
ODEs, which you learned about in Section 1.6. It is a useful method to look at because
when it works it is fairly simple to execute, and it provides a good opportunity to
review integration, which we will need for the other method as well.
Solving with integrating factors is a method that can be used to solve any linear first
order ODE, whether it is separable or not. The method of solution is more complicated
than separation of variables, but not necessarily any more difficult to execute once you
learn it.
After learning these two methods we will again look at applications, but only for first order
ODEs and IVPs at this time.
44
2.1
So far you have learned how to determine whether a function is a solution to a differential
equation, initial value problem or boundary value problem. But the following question
remains: How do we find solutions to differential equations? We will spend much of the
course learning some analytical methods for finding solutions. If the ODE is separable,
we can apply the simplest method for solving differential equations, called separation of
variables. The bad news is that separation of variables only works for separable (so
necessarily also first order) equations; the good news is that those sorts of equations actually
occur in some real life situations. Lets look at an example of how we solve a separable
equation.
Example 2.1(a): Solve the differential equation y
Note that we can write the ODE as
6 sin 3x
= 0.
y
1
dy
= 6 sin 3x = g(x)h(y), where g(x) =
dx
y
1
6 sin 3x and h(y) = . (It doesnt really matter where the 6 is.) Therefore the ODE is
y
separable; lets separate the variables and solve:
y
6 sin 3x
= 0
y
dy
6 sin 3x
=
dx
y
dy =
6 sin 3x
dx
y
y dy = 6 sin 3x dx
Z
1 2
y
2
y dy =
6 sin 3x dx
dy
Change to dx
notation and get the term with the
derivative alone on one side.
= 2 cos 3x + C
45
The resulting solution for the above example is not a function, but is instead a relation. In
some cases we will wish to solve for y as a function of x (or whatever other variables we
might be using), but you should only do so when asked to.
In the next example you will see a simple, but a very useful, type of differential equations.
dy
+ 0.5y = 0 by separation of
Example 2.1(b): Solve the differential equation
dt
variables, and solve the result for y.
First lets solve the ODE by separation of variables:
dy
+ 0.5y = 0
dt
dy
= 0.5y
dt
dy = 0.5y dt
dy
= 0.5 dt
y
Z
Z
dy
=
0.5 dt
y
ln |y| + C1 = 0.5t + C2
ln |y| = 0.5t + C3
where C3 = C2 C1 . We now solve for |y|, using the facts that the inverse of the natural
logarithm is the exponential function with base e and if |x| = u, then x = u (the
definition of absolute value):
ln |y| = 0.5t + C3
eln |y| = e0.5t+C3
|y| = C4 e0.5t
y = C4 e0.5t
y = Ce0.5t
The last step above might seem a bit fishy, but it is valid. In most cases we have initial
values, which then determine the constant C, including its sign:
dy
+ 0.5y = 0, y(0) = 7.3.
dt
We already solved the differential equation in the previous example, so we just need to find
the value of the constant by substituting the initial values into the solution y = Ce0.5t :
7.3 = Ce0.5(0)
7.3 = C
The solution to the IVP is y = 7.3e0.5t .
46
Dont assume that the the constant is always the initial value!
Example 2.1(d): Solve the initial value problem y
6 sin 3x
= 0, y(0) = 4.
y
We already solved the differential equation in Example 2.1(a), so we just need to find
the value of the constant. Substituting x = 0 and y = 4 into the solution 12 y 2 =
2 cos 3x + C:
1
(4)2 = 2 cos 3(0) + C
2
8 = 2 + C
C = 10
The solution to the IVP is
1 2
y
2
= 2 cos 3x + 10.
The next, and last, example in this section illustrates something we will see again soon.
Example 2.1(e): Solve the ODE (x2 + 4x 5)y = x + 17.
The derivative y is
dy
. When we separate the variables we get
dx
dy =
x2
x + 17
dx.
+ 4x 5
If we do the partial fraction decomposition of the fraction on the right side (see Example
A.4(a)) we can proceed as follows:
2
3
dx
dy =
x1 x+5
Z
Z
3
2
dy =
dx
x1 x+5
Z
Z
Z
3
2
dy =
dx
dx
x1
x+5
Z
Z
dx
dx
2
y + C1 = 3
x1
x+5
y + C1 = 3 ln |x 1| + C2 2 ln |x + 5| + C3
From here we can combine the constants and apply properties of logarithms to obtain
y = ln |x 1|3 ln |x + 5|2 + C
|x 1|3
y = ln
+ C,
|x + 5|2
which can also be written as
(x 1)3
+ C.
y = ln
(x + 5)2
47
dy
= x sec y
dx
(d) y =
(b) dx + x3 y dy = 0
(c) x2 + y 4
y
2x + 3
(e) x2 dy = ey dx
dy
=0
dx
(f) y =
5x + 3
y
2. Solve each of the following initial value problems. DO NOT solve for y, and give
constants in exact form.
(a) y = xy,
(b) x
y(1) = 3
dx
+ 5t = 3,
dt
(c)
dy ey
= 3,
dx x
y(0) = 2
(d) y = y 4 cos t,
x(2) = 4
y(0) = 2
3. Some of the following initial value problems can be solved by separation of variables.
Solve the ones that CAN be solved by that method. DO solve for y and give constants
in exact form again.
dy
3y = 0 ,
dx
y(0) = 4
(b) x
dy
y = x ,
dx
y(1) = 2
(c) y 4xy = 0 ,
y(0) = 2
(d) y 2x = xy ,
y(2) = 5
dy
y = e3x ,
dx
y(0) = 4
(f)
(a)
(e)
dy
y1
=
,
dx
x+3
y(1) = 3
y(0) = 0.
y(1) = 3. Give the exact
48
6. Solve each of the following ODEs. You will use separation of variables and partial
fraction decomposition for each.
dy
dy
= 2x + 3
(b) (x2 + 3x)
=3
dx
dx
dy
(c) (x2 3x 10)
= 14 (see Exercise 2(b) from Appendix A.4 to check your
dx
partial fraction decomposition)
dy
= 31 y 2 + y with various
7. In this exercise you will solve the differential equation
dx
initial values. This will lead into Sections 2.2 and 2.4, and will illustrate the sort of
calculations that we must perform to solve certain applied problems related to something called the logistic equation. Some of these calculations are not really needed
but make the expressions involved a bit simpler.
(a) Solve this system by separation of variables and partial fraction decomposition.
(Be sure to begin by multiplying both sides by 3 to clear the fraction.) This
situation is a bit different from the other ones youve encountered, in that you will
be doing the partial fraction decomposition with the dependent variable this time.
(b) Check that your answer to (a) agrees with the solution given in the back of the
book. Now e both sides, putting the right side in the form demonstrated in
Example 2.1(b). As in that example, the absolute value can be removed.
(c) Now here comes a bit of algebra: Get rid of the fraction on the left by multiplying
both sides by its denominator. Multiply both sides by ex and then solve for y.
This is the solution to the ODE.
(d) Determine the values of the constant, and then the solution to the corresponding
initial value, for each of the following initial conditions:
y(0) = 12 ,
y(0) = 0,
y(0) = 1,
y(0) = 4
DO NOT give your answers as complex fractions: Multiply the numerator and
denominator both by the same value in order to eliminate smaller fractions within
them.
(e) Use your calculator or a graphing utility like www.desmos.com to graph your solutions. (For y(0) = 12 we want only the part of the solution that goes through
that initial value.) Sketch a single grid with all four solutions on it. We call these
solution curves for the ODE, each corresponding to a different initial value. In
Section 2.4 we will see how to obtain these curves without even solving the ODE!
(f) Remembering that ex 0 as x , give the limit of each of your solutions
from part (d). This should agree with what you see in the graph from (e).
(g) Attempt to determine the value of the constant for the initial condition y(0) = 3.
What happens/what do you get?
49
2.2
Suppose that a tank contains 80 gallons of water with 10 pounds of salt dissolved in it.
Fluid with a 0.3 pounds per gallon salt concentration is being pumped into the tank at a
rate of 7 gallons per minute. The fluid is continually mixed and, at the same time, the fluid
is being drained from the tank at a rate of 7 gallons per minute. (This is similar to the
situation from Example 1.1(c).)
A quick computation reveals that the initial concentration of the solution in the tank
is 0.125 pounds per gallon, less than the concentration of the fluid that is replacing it.
Therefore the concentration of the fluid in the tank will increase, but it can never exceed the
concentration of the incoming fluid. If all of the fluid in the tank had the concentration of
the incoming fluid, there would be (0.3)(80) = 24 pounds of salt. If we were then to graph
the amount of salt in the tank as a function of time we would get the solid curve graphed
below and to the right. The limit of the amount of salt in the tank is 24 pounds.
Now suppose that the tank had A0 = 60 pounds
of salt initially, giving an initial concentration of 0.5
60
pounds per gallon, higher than the concentration
of the incoming fluid. In this case the amount of
40
salt in the tank will decrease, with a limit of 24
pounds again. The dashed line on the graph above
shows the amount of salt as a function of time, for
20
the initial amount of 40 pounds. The dotted curve
is the solution curve for the initial amount of 24
pounds, and it is also the asymptote for the other
Time
solutions.
As we saw before, the situation with the tank can be modelled with a differential equation.
The various curves graphed above are called solution curves for the ODE. Each curve is
associated with a particular initial value. Notice that none of the solutions curves cross each
other; this is not always the case, but will be for most of the ODEs that well look at. For
an initial salt amount of 15 pounds, the solution curve will lie midway between the curves
for initial amounts of 10 and 20 pounds, without crossing either.
In the exercises you will use your calculator or a graphing utility to plot solution curves
for various ODEs.
50
Direction Fields
To obtain a graph like the one on the previous page we need to either find actual solutions
to the differential equation for various initial values, or we have to have a good intuitive idea
of what is happening. What if we dont have either of those two things? Well, for first order
equations it is usually fairly easy to determine what solution curves look like from just the
differential equation itself, as we will now see.
dy
Consider the first order linear differential equation
+ y = x. We can rearrange the
dx
dy
= x y. Now remember that by solving the differential equation we
equation to get
dx
mean finding a function y = y(x) that makes the equation true; there are infinitely many
such functions, with the graph of each representing a particular solution curve. Recall also
that when considering the graph of a function, the derivative of the function at some point
is the slope of the tangent line to the graph of the function at that point. So the second
equation gives us a formula for finding the slope of the tangent line to the unknown function
y(x) at any point (x, y).
To be more specific, consider the point (3, 1). The equation
dy
=xy
dx
tells us that at that point the slope of the tangent line to the
solution curve will be
dy
=31=2
dx (3,1)
4
-2
Example 2.2(a): Find slopes for the remaining grid points, for
It is often easiest to determine slopes not by going point to
point, but to find all points where the slope is the same. For
this equation, the slope will be zero at every point where
x = y, so at (0.0, (1, 1), (2, 2), and so on. Similarly,
the slope will be one at all the points where x is one unit
larger than y; for the above grid those are the points (1, 0),
(2, 1), (3, 2) and (4, 3). Similarly, the slope will be two at
the points where x is two greater than y: (0, 2), (1, 1),
(2, 0), (3, 1) and (4, 2). The slope lines for slopes zero,
one and two are plotted on the grid shown to the right. The
remaining slopes can be seen below.
51
dy
= x y.
dx
4
-2
The graph of the result of what we have been doing is something called a direction
field or slope field; the completed direction field can be seen to the left at the top of the
next page. Direction fields are a way of studying the behavior of solutions to differential
equations without actually solving the equations analytically. The slope lines that we have
drawn in on the direction field to the left below are not all that are possible - such a slope
line exists for every point in the plane. Given a direction field and an initial value, we can
sketch solution curves by drawing curves starting at the initial value point that are tangent
to the imagined slope lines at all points that a curve goes through. Below and to the right
you can see the solution curves corresponding to the initial values y(0) = 2, y(0) = 1 and
y(0) = 3. Each curve is began by sketching a curve that is tangent to the slope line at
the initial value, then continues to be tangent to other slope lines it passes through or near
as the curve is constructed.
-2
-2
(b) y(0) = 0
(c) y(0) = 1
(d) y(0) = 3
2. On your graph from Exercise 1, sketch what you think the graphs for initial conditions
y(0) = 2, y(0) = 1 and y(0) = 3 would look like. Then graph them with a graphing
tool to check yourself. (You will need to find the values of C for each to do this.)
y
3. The graph of some solution curves for a
I
II
differential equation are shown to the right.
3
Give the Roman numeral that corresponds
to each given initial condition.
(a) y(0) = 1
(c) y(1) =
1
2
(b) y( 21 ) = 1
(c) y(0) = 0
x
III
3
52
IV
4. (a) On the grid for Exercise 3, sketch in what yo think the solution curve for the initial
value y(0) = 43 would look like.
(b) The general solution for the ODE for which some solution curves are shown in
Exercise 3 is y = 21 (sin x + cos x) + Cex . Determine the value of C for the initial
value y(0) = 43 and plot the solution curve using technology. Compare with the
curve you sketched for part (a).
5. (a) The graph below shows some solutions to
with its initial value y(0) =
dy
= xy 2 . Label each that you can
dx
2
. Find a point on one of the curves
+C
for which you couldnt find an initial value and substitute it into the solution to
determine the value of C.
x2
(c) Use technology to graph the solution, for the value of C that you found in (b).
Explain what is going on.
y
4
x
-4
-4
6. Suppose that a group of N0 individuals is put in an environment that can only support K individuals, and suppose that the growth rate of the population without any
restrictions would be r percent (in decimal form!) per year. Then the population N at
any time t years is given by
N=
K
1 (1 K/N0 )ert
The value N0 is called the initial population, K is called the carrying capacity.
Suppose that for some population the carrying capacity is 100 and the growth rate is
20%. Graph the functions N for the initial populations below all on the same grid,
for zero to forty years, using technology. Your graph will need to go up to at least 150
individuals. Sketch the graph.
(a) N0 = 20
(b) N0 = 150
(c) N0 = 100
(d) N0 = 0
7. Think about the graph you got in the previous exercise, and make sure that you understand (from a population growth point of view) why each curve looks the way it
does.
53
8. For each ODE given, plot the direction field at integer coordinates over the values given
for each variable.
dx
dt
dy
(b)
dx
dy
(c)
dx
dy
(d)
dt
(a)
1
= xt,
2
1 t 2, 2 x 2
= x2 2x,
0 x 4, 2 y 4
= y 2 2y,
0 x 4, 2 y 4
= y + t.
2 t 2, 2 y 2
9. On the direction field below and to the left, sketch the solution curves going through
the given points.
(a) (4, 5)
(b) (2, 2)
(c) (6, 4)
-1
-4
-2
-6
-3
-4
-2
-2
-6
-3
Exercise 9
-2
-1
Exercise 10
10. On the direction field above and to the right, sketch the solution curves going through
the given points.
(a) (2, 0)
(b) (1.5, 1)
54
(c) (3, 0)
2.3
Lets begin with an example that demonstrates the limitation of separation of variables.
dy
3y = e5x .
dx
Note that if we try to separate the variables we get
dy
3y = e5x
dx
dy
= 3y + e5x
dx
dy = (3y + e5x ) dx
Here we see that there is no way to get the 3y term back over to the left side with dy.
(This is because 3y + e5x cannot be written in the form g(x)h(y).) Therefore this
equation cannot be solved by separation of variables.
The following derivative computation provides the key for solving equations like the one
above.
Example 2.3(b): Suppose that y = y(x) is some function of x. Find the derivative
of ye3x (with respect to x).
Because both y and e3x are functions of x, we must use the product rule:
dy
d 3x
d
3x
3x d
3x
3x dy
3x
(ye ) = y (e ) + e
(y) = 3ye
+e
=e
3y
dx
dx
dx
dx
dx
Notice now that if we were to multiply the left side of the ODE of Example 2.3(a) by e3x we
would get the result of Example 2.3(b). This indicates an idea for solving the ODE:
dy
3y = e5x
dx
e3x
Original equation
dy
3e3x y = e5x e3x
dx
d(ye3x )
= e2x
dx
d(ye3x ) = e2x dx
3x
d(ye
) =
e2x dx
ye3x = 12 e2x + C
y = 12 e5x + Ce3x
dy
3y = e5x is y = 12 e5x + Ce3x . The reason for multiplying both
dx
sides by e3x was to get y alone on the left side.
dy
3y = e5x probably seems a bit
The method just shown for finding the solution to
dx
mysterious, to say the least! This is called the integrating factor method, which we now
summarize. Note that it only applies to linear first order ODEs, which can always be put
dy
into the form
+ p(x)y = q(x).
dx
Thus the solution to
dy
+ p(x)y = q(x),
dx
Note that after integrating both sides of the equation there will be a constant added to
the right side. This constant will beZ multiplied by Z
eu in the solution. For the equation
dy
3y = e5x , p(x) = 3 so u = p(x) dx = 3 dx = 3x and eu = e3x .
dx
Any first order linear ODE can be solved using the integrating factor method, as long
as eu q(x) can be integrated; sometimes you can use either this method or separation of
variables and they both will work. Now lets take a look at executing the above steps with
another example.
dy y
+ = x2 for x > 0 by the integrating factor method.
dx x
1
and q(x) = x2 . Because x > 0, |x| = x and
First we note that p(x) =
x
Z
1
dx = ln x. Therefore eu = eln x = x. We now carry out steps (2) through (4)
u=
x
above:
56
dy y
+
dx x
dy
x +y
dx
d(xy)
dx
Z
d(xy)
= x2
original equation
= x3
= x3
Z
=
x3 dx
xy =
1 4
x
4
+C
y =
1 3
x
4
C
x
1
1
=
u
e
x
dy
3y = e5x , y(0) = 1.
dx
dy 1
y = 0, y(0) = 23 .
dx 2
6. The IVPs from Exercises 3(b) and 3(e) of Section 2.1 couldnt be solved by separation
of variables, but they can be done with integrating factors. You will do them here.
dy
y = e3x ,
dx
dy
(b) Solve the IVP x y = x ,
dx
(a) Solve the IVP
y(0) = 4.
y(1) = 2. Begin by multiplying through by
57
1
.
x
7. In this exercise you will see another method for solving the ODE from Exercise 4. This
method will be used later when we solve second order ODEs.
(a) The equation y 5y = 0 is the homogeneous equation associated with
y 5y = 3 cos 2t. Substitute y = Cert into the homogenous equation to determine
what value r must have in order for y = Cert to be a solution. For that value
of r, y = Cert is called the homogeneous solution to y 5y = 3 cos 2t.
(b) Find the values of A and B for which y = A sin 2t + B cos 2t is a solution to
y 5y = 3 cos 2t. Do this as follows:
Combine the like terms on the left side of the equation to get only one sine
term and one cosine term.
You will need to note that on the right side of your equation 3 cos 2t is the
same as 3 cos 2t + 0 sin 2t. Equate the coefficient of cos 2t on the left side
with the coefficient of cos 2t on the right side to get an equation involving
the unknowns A and B. Then repeat for sine to get another equation with
A and B.
Solve two equations for the two unknowns A and B. The resulting y =
A sin 2t + B cos 2t is called the particular solution to y 5y = 3 cos 2t.
(c) Write the sum of the homogeneous and particular solutions. This is known as the
general solution, and should match what you found in Exercise 4(a).
dy
3y = e5x from Exercise 1 with
dx
this difference: For part (b), find the value of A for which y = Ae5x is a solution to
the ODE.
dy
9. Use the method of Exercise 7 to solve the ODE
+ 3y = t2 + 5t 1 from Exercise
dt
5, but for part (b), find the values of A, B and C for which y = At2 + Bt + C is a
solution to the ODE.
8. Use the method of Exercise 7 to solve the ODE
58
2.4
Recall that a first order ODE is called autonomous if it can be written in the form
dy
= f (y). That is, when we get the derivative alone on the left side of the equation, the
dx
dy
right hand side is a function of only the dependent variable.
= y 2 2y is an example of
dt
dy
= 0 whenever y = 0, so if we had an initial condition
an autonomous ODE. Note that
dt
of y(0) = 0, then the value of y would never change because the rate of change with
respect to time is zero. Therefore the solution to the IVP
dy
= y 2 2y,
dt
y(0) = 0
is the constant function y = 0. We have solved the IVP without doing any calculations!
Now suppose that, for the same ODE, y(0) = 2. We see that when y = 2 we again have
dy
= 0, so the value of y will again not change. This means that y = 2 is the solution
dt
to the IVP
dy
= y 2 2y,
y(0) = 2
dt
The graph to the right shows the two solution curves that
dy
= y 2 2y, for the
we have obtained so far for the ODE
dt
y
initial values y(0) = 0 and y(0) = 2. (The dots represent
the initial values themselves - note the position of zero on the
2
horizontal axis.) We will call constant solutions like those two
y(0) = 2
equilibrium solutions; the word equilibrium essentially mean1
ing unchanging as time goes on. The question that should occur
0
to you is What happens for other initial values of y? With
y(0) = 0
a little thought we should be able to figure that out. There are
three key observations we can make that will help answer the
0
1
2
3 t
question:
The direction field depends only on y alone, so for any given value of y the slope
remains constant.
59
The right hand side of the ODE can be factored to y(y 2). From that we see that
dy
if y < 0 both y and y 2 will be negative, so
= y(y 2) will be positive.
dt
Therefore any solution with an initial value less than zero will be increasing. When
dy
0 < y < 2, y is positive and y 2 is negative, so
is negative and any solution
dt
with an initial value between zero and two will be decreasing. Finally, when y > 2 we
dy
have
> 0 because both y and y 2 are positive when y > 2. Any solution
dt
with an initial value greater than two will be increasing.
dy
approaches zero as y gets nearer
Whether positive or negative, the value of
dt
to either zero or two. Therefore the direction field lines become flatter (closer to
horizontal) for values of y close to zero and two.
dy
= y 2 2y has the
From these observations we can deduce that the direction field for
dt
appearance shown below and to the left. The direction field with solutions corresponding to
four different initial conditions is shown in the center below.
y
y
y(0) = 2.5
y(0) = 2
y(0) = 1
0
y(0) = .5
We will summarize the information in the three bullets above with something called a
phase portrait, shown above and to the right. (Technically it is a one-dimensional phase
portrait. Those of you taking the second term of this course may see two-dimensional phase
portraits.) The vertical line indicates y values, with the two critical points zero and two
indicated. The critical points divide the line into three intervals, and the arrow in each
interval indicates whether y is increasing or decreasing, in the particular interval, as time
increases.
Look again at the direction field in the middle above, with the four solutions curves drawn
in. Note that solutions with an initial y value less than two (including less than zero) all
tend toward the constant solution y = 0, as the phase portrait tells us they will. When
this occurs we say that the solution y = 0 is a stable equilibrium solution. You can
sort of think that if we have an initial condition of zero the solution will be zero, and if we
bump off from zero a bit with our initial condition, the new solution obtained will tend
back toward zero as time goes on.
On the other hand you can see that if y starts with the value two it will remain two,
but if it starts at any value close to, but not equal to, two the solution will diverge away
from two. (Again, look at the phase portrait to see how it indicates this.) For this reason
we call the solution y = 2 an unstable equilibrium solution. Other language you might
hear is that y = 0 is a stable critical point on the phase diagram, and y = 2 is an
60
unstable critical point. On our phase portraits we will indicate stable critical points with
solid dots, and unstable critical points with open circles. (We will also use open circles for
semi-stable equilibria, which you will see later.)
Again, the entire analysis you have just seen is based on the fact that the ODE is
autonomous. Two points of interest here are
Autonomous first order ODEs are not just a curiosity - they occur naturally in many
applications.
Autonomous ODEs can be difficult or impossible to solve. However, an analysis like we
just did can make it very easy to determine how solutions to such an equation behave,
in a qualitative sense.
It is important to be able to recognize autonomous differential equations.
Example 2.4(a): Determine which of the following first order ODEs are autonomous:
dy
= y2 x
dx
dy
= t2 5t + 1
dt
y = 2y 1
dx
=x+1
dt
Clearly the first equation is not autonomous and the second is. The third and fourth
equations might be a bit confusing, as the variables are no longer x and y. In the third
equation the dependent variable is y and the independent variable is t. Because the right
hand side is a function of only the independent variable t, the equation is not autonomous.
In the fourth equation the dependent variable is x, the independent variable is t, and
dx
x+1
the equation can be rewritten as
=
. The right hand side is then a function of
dt
x
x alone, so the equation is autonomous.
If we can determine the phase portrait for an autonomous ODE, we then have a pretty
good idea what all solutions to the ODE look like, without having to go to the trouble of
creating a direction field. The next example shows how this is done.
dy
= y 3 + 6y 2 9y and use it to
dt
sketch solution curves for the initial conditions y(0) = 4, y(0) = 2 and y(0) = 1.
Identify each equilibrium solution as stable or unstable.
We factor the right hand side of the ODE, starting by factoring y out:
dy
= y 3 + 6y 2 9y = y(y 2 6y + 9) = y(y 3)2
dt
From this we can see that the equilibrium solutions are y = 0 and y = 3. Testing values
of y in each of the three intervals (, 0), (0, 3) and (3, ) gives us the following:
When y < 0,
dy
>0
dt
dy
<0
dt
When y > 3,
dy
<0
dt
This gives us the phase portrait shown to the left below, which indicates that there are
equilibrium solutions of y = 0 and y = 3. If y(0) < 0 the solution is increasing, but
will approach the equilibrium solution y = 0; if 0 < y(0) < 3 the solution decreases
61
toward y = 0. If y > 3 the solution also decreases, but toward y = 3. These behaviors
are shown to the right below, for the three solutions with the given initial values. The
solution y = 0 is a stable equilibrium solution, and the solution y = 3 is neither stable
nor unstable, but is what we call a semi-stable equilibrium solution.
y
y(0) = 4
4
3
y(0) = 2
0
y(0) = 1
dy
2y = 0
dt
(d) y = y 2 7y + 10
dy
+ xy = 1
dx
1 dy
+y =1
(e)
x dx
(b)
1 dy
+y =3
y dx
dx
(f)
+ 6x2 = x3 + 9x
dt
(c)
on a single separate graph, sketch in all equilibrium solution curves, and one solution curve with an initial value in each interval of the real line created by the
values of the equilibrium solutions
give all equilibrium solutions and, for each, tell what kind of equilibrium it is
3. In (a) below, the graph of some solution curves to an ODE are given. (b) and (c) are
phase portraits for two other ODEs.
y
y
(b)
(c)
(a) y
6
4
2
0
1
t
5
2
0
(a) For the solution curves in (a) above, identify each equilibrium solution and tell
whether it is a stable equilibrium, unstable equilibrium, or semi-stable equilibrium.
62
(b) Repeat (a) for the phase portrait shown in Exercise 3(b).
(c) Repeat (a) for the phase portrait shown in Exercise 3(c).
dy
= 31 y 2 + y. When working with
dx
it in this exercise, you will find it useful to factor 13 out of the right hand side first.
(a) When we tried to solve the ODE with the initial value y(0) = 3 we were not able
to obtain a solution. What do we now know happens for that initial condition?
(b) Give the equilibrium solutions, and what kind each is.
(c) Sketch the phase portrait and some solution curves for the ODE.
63
2.5
Performance Criteria:
2. (k) Solve an applied problem modelled by a first order ODE using
separation of variables or an integrating factor.
(l) Give an ODE or IVP that models a given physical situation
involving growth or decay, mixing, Newtons Law of Cooling
or an RL circuit.
(m) Sketch the graph of the solution to a mixing or Newtons Law
of Cooling problem, indicating the initial value and the steadystate asymptote.
(n) Identify the transient and steady-state parts of the solution to
a first order ODE.
64
Example 2.5(a): Five hundred rainbow trout are introduced into a previously barren
(no fish in it) lake. Three years later, biologists estimate that there are 1730 rainbow
dN
= rN, detertrout in the lake. Assuming that the population satisfies the ODE
dt
mine the function N = N(t) that gives the number N of rainbow trout in the lake
at time t.
dN
= rN says that we are looking for a function N(t) whose derivative is
dt
r times the function itself, and N = ert clearly is such a function. This function contains
no constant of integration, but we recall that N = Cert is also a solution, for any value
of C. The general solution is then N = Cert . The fact that there are 500 fish in the
lake at time zero gives us 500 = Ce0 , so C = 500 and N = 500ert .
The ODE
To find r we substitute N = 1730 when t = 3 into our solution to get 1730 = 500e3r ,
and solve to find r = 0.414. The equation for the number of rainbow trout at any time
t is then N = 500e0.414t .
Heres how one would solve the differential equation from the last example by separation
of variables, with some of the steps combined:
dN
dt
= rN
dN
= r dt
N
ln N = rt + C
N = Cert
Mixing Problems
Weve discussed this sort of situation previously, so lets go straight to an example:
Example 2.5(b): A tank contains 80 gallons of water with 10 pounds of salt
dissolved in it. Fluid with a 0.3 pounds per gallon salt concentration is being pumped
into the tank at a rate of 7 gallons per minute. The fluid is continually mixed and,
at the same time, the fluid is being drained from the tank at a rate of 7 gallons per
minute. Letting A represent the amount of salt in the tank, in pounds, Sketch a graph
of A as a function of time t. Label any values you can on the A axis.
In the next two examples we will set up and solve the initial value problem for this situation
analytically, arriving at an equation that will give us the amount of salt at any time t.
Before doing that it would be good to have some idea of what the behavior of the solution
would be. We did this previously, in Example 1.1(j), but lets repeat the reasoning here.
65
A (lbs)
24
10
Now let A = A(t) be the amount (in pounds) of salt in the tank at any time t minutes.
A
, so by the same sort of calculation that
Then the concentration of salt in the tank is 80
lbs
= 0.0875A min
. The net rate
we just did the rate at which salt is leaving the tank is 7A
80
of change of salt in the tank is the amount coming in minus the amount going out, or
lbs
(2.1 0.0875A) min
. But this quantity, being a rate of change, is also a derivative. Namely
dA
, giving us the differential equation
it is
dt
dA
= 2.1 0.0875A
dt
We also have the initial value A(0) = 10 pounds, so we have the initial value problem
dA
= 2.1 0.0875A ,
dt
A(0) = 10
We will see differential equations like the one above in several contexts, and it can always
be solved using separation of variables or an integrating factor (and you should be able to
do it either way). However, this form of ODE comes up often enough in practice that we
should use the following formula instead, obtained by u-substitution:
Z
1
1
dx = ln | ax + b | + C
(1)
ax + b
a
We can now use this result to solve the above IVP.
Example 2.5(d): Solve the IVP from Example 2.5(c).
Multiplying both sides of the ODE by dt and dividing by the quantity 2.1 0.0875A gives
us
dA
= dt
2.1 0.0875A
66
To use equation (1) we note that our left side is like the left side of (1), but with x = A,
a = 0.0875 and b = 2.1. The formula then tells us that the left side integrates
1
to 0.0875
ln(2.1 0.0875A) + C. (We dont need absolute value because the quantity
2.1 0.0875A is the rate at which the amount of salt is changing, and this is positive due
to the concentration of the incoming solution being higher than the initial concentration in
the tank.) Thus when we integrate both sides and combine the constants we have
1
0.0875
ln(2.1 0.0875A) = t + C
67
T
88
50
T (0) = 88
dT
= k dt, and the method of Example 2.4(d)
T 50
gives us T = 50 + Cekt before applying the initial condition. Using the initial condition,
we obtain the solution T = 50 + 38ekt . To determine k we apply T (1) = 65 to get
k = 0.930, so the final solution is T = 50 + 38e0.930t .
Electric Circuits
We will work with some basic electrical circuits in this class. The first sort of circuit
well look at consists of a voltage source, a resistor and an inductor. The voltage source
can be constant (so-called direct current, or DC), or it can be variable, usually in an
oscillating manner (alternating current, or AC). The voltage source causes electrons to
move in the circuit, and the flow of electrons is called current. The voltage source provides
an electromotive force which we can think of as sort of pushing current through the
circuit, analogous to a pump pushing fluid through a network of pipes. The units of the
electromotive force are volts. We will use the symbol i for current, and it is measured in
units called amperes. (Amps, for short.)
The resistor has a characteristic called resistance, which
is measured in units called ohms. The inductors characteristic
is called inductance, which is measured in henries. Although
L
resistance and inductance could be variable, they always be conE
stants in our considerations. We will use E = E(t) for the voltage, R for the resistance and L for the inductance. To the
right is a schematic diagram of such a circuit. We will usually
think of our circuits as having a switch as well, which is open
R
(off) until time zero, when it is closed (turned on). From that
Figure 2.5(a)
point on the current is (usually) changing, and is a function of
time: i = i(t).
68
RL Circuit
Consider an electric circuit as described above, with an applied voltage E(t) volts (possibly a function of time) and constant resistance of
R ohms and constant inductance of L henries. The current i = i(t),
in amperes, satisfies the differential equation
L
di
+ Ri = E(t)
dt
Methods for solving ODEs arising from the above are straightforward, so we wont look
at examples of how to solve them. Instead we will examine the case where the voltage
source E(t) is a constant function, to observe why mathematics is so powerful in science
and engineering. Suppose that L = 14 henry, R = 15 ohms, and E = 12 volts; in this case
di
di
= 4 dt or
the ODE is 41 + 15i = 12. If we separate the variables we obtain
dt
12 15i
di
= dt. Lets look at the first of these along with the separated equations arising in
48 60i
Examples 2.5(d) and 2.5(e):
di
= 4 dt
12 15i
Note that these are all of the form
dA
= dt
2.1 0.0875A
dT
= k dt
T 50
dx
= c dt, where a, b and c are constants. This
ax + b
We will see this principle in action again when we study second order ODEs.
Response of a System
Suppose that we have a circuit like that show in Figure 2.5(a), but without a voltage source
E (but with the circuit completed). If there is no current in the circuit initially, then there will
not be any current at any future time. However, if there is some current in the circuit initially
(which can be made to happen by including a voltage source, then removing it and completing the
circuit in its absence), then there will be current in the future as well. Similarly, if we set a mass
on a spring (like shown in Example 1.1(a)) in motion it will continue to oscillate for some time.
We will refer to the circuit without a voltage and the mass on a spring as systems. In the case
of the circuit the variable of interest is the current in the circuit at any time, and for the mass we
are interested in the vertical position at any time. With some initial stimulus (non-zero initial
conditions) in each case the current or vertical position will vary with time. The current or vertical
position will be referred to a s the response of the system to the initial conditions.
Now suppose that we have either an RL circuit with a voltage source or a spring-mass system
with some outside force pushing or pulling on the mass. The outside force or voltage (which is
also sometimes referred to as an electromotive force) we will refer to as a forcing function for
the system. In the presence of a voltage source, the circuit will have current at all future times.
Similarly, a mass on a spring with a forcing function will continue to oscillate
We will revisit the spring-mass system, along with slightly more complex electrical circuits, in
Chapters 3 and 4. For the time being, lets focus on the circuit shown in Figure 2.5(a) and the
69
di
+ Ri = E(t).
dt
(2)
di
Here the left side L +Ri of the equation represents the system, and the right side E(t) represents
dt
the forcing function. Or goal is usually to solve an associated initial value problem for the current
i = i(t). That current, the solution to the IVP, is the response of the system to the forcing function
E(t) and initial current. Another way of thinking about this is that the forcing function and initial
condition(s) are inputs to the system, and the response is the output of the system for that
input.
Let us consider for example an RL circuit where again L = 41 henry and R = 15 ohms, and
for which E(t) = sin 3t. In this case we would not be able to separate the variables, so wed solve
the equation using an integrating factor. In doing so, we would obtain the solution
i=
4
63
sin 3t + Ce60t
where C is a constant that would be determined by an initial condition. Note that the solution
has two parts:
4
sin 3t, which is periodic and goes on forever in the same way. This part of the
The part 63
solution is called the steady-state solution or steady-state response of the system.
The part Ce60t , which approaches zero as time goes on, so it dies out. Such a solution
or part of a solution is called the transient solution or transient response of the system.
To clarify a little, we will define a steady-state solution to be any solution that is either constant
(and, to avoid triviality, not zero) or periodic.
In practice, when a system like a machine with moving parts or an electrical circuit is turned
on, it often exhibits a certain behavior as it starts up, which is the transient response of the
system. Then it will settle in to a steady state behavior or response. (Note that the ideas of
transient and steady state solutions only make sense when the independent variable is time.) In
general only the steady-state response is important in terms of what we want the system to do
from a practical viewpoint, but the transient response might be of interest because it might cause
some sort of stress on the system that could cause a problem.
For the scenario described above but with E having the constant value of 12 volts, the solution
to the differential equation is
12
+ Ce60t ,
i = 15
where C is again determined by the initial current in the circuit. We can see that here the
steady-state solution is i = 12
15 is the resistance;
15 , where 12 is the voltage and
this says that in
V
. This is because
the long run the circuit will exhibit Ohms Law V = IR solved for I =
R
the current approaches a constant value, and the inductor only affects the circuit when there is
change in the current, causing flux in the coil of the inductor.
Example 2.5(f ): In Example 2.5(e) we found that when a solid object with initial temperature 88 F is placed in a medium with ambient temperature 50 F, and after one hour
the object has a temperature of 65 F, the temperature T at any time t is given by
T = 50 + 38ekt . Give the transient and steady-state parts of the solution.
The transient part of the solution is 38ekt and the steady-state part is 50.
70
71
6. The ODE
dA
= 2.1 0.0875A from Example 2.5(c) is autonomous.
dt
(a) Determine all equilibrium solutions, and tell whether each is stable, unstable or semistable.
(b) Sketch some solution curves for A(0) > 0, including some with A(0) greater than the
largest equilibrium solution.
7. When Mrs. W arrives home in the Watermans Subaru Outback, it is at 29 from being
outside all day. She parks it in the geothermally heated garage, which is at a temperature of
73 F.
(a) We wish to determine the temperature T of the Subaru as a function of time t,
assuming it follows Newtons Law of Cooling. Write a differential equation to be solved
to find that function, the solution we are looking for. Also, give any initial condition(s).
(b) Will there be a steady-state solution? If so, what will it be?
(c) Solve the initial value problem. Your answer will still contain a constant, but you should
be able to determine the value of another.
(d) Suppose that two hours later Mrs. W is ready to go back out in the Subaru, and by that
time it had warmed up to 47 F. Determine the function modelling the temperature of
the car in the two hours that it was in the garage.
8. An RL circuit contains an inductor with an inductance of 34 henry and a 15 ohm resistor.
It is driven by a variable voltage E(t) = 6 cos 2t, and the initial current in the circuit is 2
amperes.
(a) Give the initial value problem to be solved, and solve it. Determine exact values for all
constants.
(b) Give the steady-state and transient solutions.
9. A 150 gallon tank contains a 3 pounds (lbs) per gallon (gal) salt solution. At time zero,
solution will begin being pumped out of the tank at a rate of 7 gallons per minute and a 1
pounds per gallon solution will begin being pumped into the tank at the same rate. Assume
that there is constant mixing in the tank, so that it has the same concentration all over in
the tank at any given time. Let A represent the amount of salt in the tank (in pounds) and
let t represent time (in minutes).
(a) Sketch the graph of the solutions as well as possible from the given information.
(b) Give the initial value problem to be solved, and solve it.
(c) Graph your solution using some technology, and compare with your answer to (a).
(d) Give the steady-state and transient solutions.
10. Suppose that we have an RL circuit with no voltage, as shown
to the right. The resistor has a resistance of 8 ohms, and the
inductor has an inductance of 31 henry, and there is an initial
current in the circuit of 5 amperes.
72
2.6
Chapter 2 Summary
73
2.7
Chapter 2 Exercises
dy
1. In Example 2.1(b) the ODE
+ 0.5y = 0 is solved by separation of variables, and it can
dt
also easily be solved using an integrating factor.
(a) Instead of either of these, assume that it has a solution of the form y = Cert and
determine the value of r by substituting this solution into the equation. After finding
the value of r, give the solution to the ODE.
dy
+ 0.5y = 0, y(0) = 4.7
(b) Solve the IVP
dt
2. Consider the situation of Example 2.5(b) with the following change: Suppose that the 0.3
pounds per gallon fluid is coming in at a rate of 5 gallons per minute, rather than 7 gallons
per minute. (The mixed fluid is still being drained from the tank at 7 gallons per minute.)
The goal here is to determine the amount A of salt in the tank at any time t.
(a) Give an expression for the amount of fluid in the tank at any time t.
(b) Give an expression for the concentration of salt in the tank at any time t.
(c) Give the initial value problem to be solved in order to determine the amount of salt in
the tank as a function of time.
(d) The differential equation is linear. What are the functions p(t) and q(t)?
(e) Solve the differential equation, using the integrating factor method. Graph the solution
and make sure it behaves as expected.
Reduction of Order
The term reduction of order refers to any process in which one or more ODEs is turned into one
or more other ODEs of smaller order. This can be done in a variety of ways, the simplest of which
is illustrated in the next exercise.
3. A classic problem in the study of PDEs is the equilibrium distribution of heat in a circular
disk. In the course of solving that problem one obtains the ODE
r2
dR
d2 R
+r
n2 R = 0.
2
dr
dr
(1)
Note that r and R are two different variables! R is the dependent variable, and is a
function of the independent variable r. Later we will see how to solve this equation for
n 6= 0, but here we will solve for n = 0.
(a) Write the equation with n = 0.
d2 R
dR
. What then is
, in terms of S?
dr
dr 2
(c) Substitute what you were given and determined in (b) into your equation from (a) to
obtain a first order ODE with dependent variable S.
(b) Let S =
74
Here you used reduction of order to start with a second order ODE but make a substitution
that gives us a first order equation.
4. Here is another situation which is basically reduction of order, but disguised a little. Recall
(from Section 0.2) that the differential equation governing the motion of a falling object (or
one that has been projected upward) is
d2 h
= 32,
dt2
(2)
where h is the height of the object at any time t. For the value 32 on the right hand
d2 h
is the acceleration due to gravity.
side, h is measured in feet and t in seconds. Now
dt2
Remembering that acceleration is the derivative of velocity, (2) can be rewritten as
dv
= 32.
dt
(3)
All of this is based on the assumption that there is no air resistance, but now lets remove
that assumption. A reasonable assumption is that the air resistance is proportional to the
velocity, but in the opposite direction. Letting the constant of proportionality be k > 0, (3)
then becomes
dv
= 32 kv.
(4)
dt
(a) Equation (4) is autonomous; what is the equilibrium solution? Is it stable, or unstable?
Sketch the phase diagram and and a graph of some solution curves.
(b) Solve (4) using one of the methods from this chapter. Take the limit of your solution as
t and make sure it matches your equilibrium solution from (a).
dh
and solve the resulting differential equation
(c) Replace v in your answer to (b) with
dt
for h.
(d) Give the equation for the height of an object that begins its motion by being projected
upward at 120 feet per second from an initial height of 700 feet.
Logistic Growth
When we assume that a population will increase exponentially for all time, the differential equation
for the number N of individuals at time t is
dN
= rN,
dt
(5)
where r is a constant that represents the growth rate. (See Example 2.5(a).) This model is
somewhat unrealistic, however - usually we expect some upper limit to the population due to the
fact that as it gets large it will begin to be constrained by some factor like the food available or
disease. Thus the growth rate should approach zero at some point, or even become negative if the
population becomes too large. On the other hand, the growth rate should be relatively constant
for very small numbers N of individuals. These conditions can be incorporated into our model
N
for some other constant K > 0:
by including a factor 1
K
dN
N
=r 1
N.
(6)
dt
K
Equation 6 is one of several forms of what is called the logistic equation.
75
5. (a) Equation (6) is autonomous. Determine all equilibrium solutions, and classify each as
stable or unstable. Sketch a phase diagram and some solution curves.
(b) Discuss the significance of the constant K.
(c) What effect should changing r have on solution curves? Be as specific as possible.
(Hint: Think in terms of the direction field for the equation.)
6. (a) Solve equation (6) for K = 3000 and N (0) = 500. You dont need to know the value
of r to do this, but r will appear in your solution.
(b) Suppose that N (4) = 2000, Use this to determine the value of r.
(c) Using your solution to (a) with the value of r determined in (b), determine when the
population will reach N = 2500.
RC Circuits
In Section 2.5 there is a discussion of RL circuits,ones containing a voltage source, resistor
and inductor. Another simple circuit of interest is one containing a voltage source, resistor and
capacitor, called an RC circuit. Capacitors are devices that store something called charge, which
well denote as q. The units of charge are coulombs. The ability of a capacitor to store charge
is quantified by a characteristic called capacitance, denoted by C. The units of capacitance are
farads.
To the right is a schematic diagram of an RC circuit, which we
will also assume has a switch that allows current to begin flowing at
some time. The differential equation that models the charge q (in
C
coulombs) on the capacitor at any time t (in seconds) is
E
1
dq
R + q = E(t),
dt C
where R is resistance in ohms and C is capacitance in farads, and
R
E(t) is the voltage, which may or may not be constant.
dq
Finally, we note that the derivative
, the rate at which the charge on the capacitor is
dt
changing with respect to time, is the current in the circuit.
7. A 12 volt battery is attached to a circuit containing a 0.5F (microfarad, 106 farad) and
an 8k (kilo-ohm, 103 ohm) resistor. At time zero, when the circuit is closed by a switch,
the capacitor has a charge of 109 coulombs.
(a) Give the charge q on the capacitor and the current i in the circuit as functions of
time t in seconds after closing the switch. Note that constants and variables have to be
in terms of volts, ohms, farads in order to get results in terms of coulombs and amperes.
(b) Plot the charge on the capacitor and the current in the circuit as two separate graphs.
Indicate clearly any asymptotes.
8. Consider the same situation as the previous exercise, but with the 12 volt battery replaced
by a variable voltage source E(t) = 10 cos 240t. Repeat parts (a) and (b) of the previous
exercise, but do not expect asymptotic behavior of either the charge or the current.
76
Learning Outcome:
3. Solve second order linear, constant coefficient ODEs and IVPs. Understand the nature of solutions to such ODEs and IVPs.
Performance Criteria:
(a) Solve an Euler equation.
(b) Solve a second order, linear, constant coefficient, homogeneous ODE.
(c) Find a particular solution to a second order linear, constant coefficient
ODE using the method of undetermined coefficients.
(d) Evaluate a differential operator for a given function.
(e) Solve a second order linear, constant coefficient IVP.
(f) Given minimal, but sufficient, information about an ODE of the form
ay + by + cy = f (t) and its general solution, determine whether a
function is a solution to ay + by + cy = f (t) or ay + by + cy = 0.
(g) Determine, from the definition, whether two functions are linearly
independent.
(h) Use the Wronskian to determine whether two solutions to a second
order linear ODE are independent.
(i) Given one solution to a second order ODE, use reduction of order to
find a second solution.
In this course we are focusing on differential equations that can be solved by analytical (penciland-paper) techniques. Many differential equations cannot be solved this way, and numerical
methods must be employed to obtain solutions. (See Appendix B for an introduction to solving
ODEs by numerical techniques.) Our chances of being able to solve an ODE analytically are much
greater if it is linear.
A second order linear ODE has the form
a2 (x)
dy
d2 y
+ a1 (x)
+ a0 (x)y = f (x),
2
dx
dx
(1)
and when f (x) = 0 it is a homogeneous linear differential equation. Here a2 , a1 and a0 are
functions of the independent variable x. In this chapter we will focus almost entirely on second
order linear differential equations in which all the coefficients are constants and the independent
variable is time t, rather than x. So our equations will generally have the form
ay + by + cy = f (t),
(2)
where a 6= 0, b and c are constants and f is a function that we will refer to as the forcing
function. (2) is called a second order, linear, constant coefficient ODE.
The one other type of equation we will see in this chapter is a variety called an Euler equation.
For such equations a2 (x) = ax2 , a1 (x) = bx and a0 (x) = cx, where b and c are constants, and
77
d2 y
dy
+ cy = 0
+ bx
2
dx
dx
(3)
Equations of this form arise when solving certain partial differential equations.
In the first section of the chapter we will solve equations of the forms
ax2
dy
d2 y
+ bx
+ cy = 0
2
dx
dx
and
ay + by + cy = 0,
(4)
both of which are clearly homogeneous. These equations will always have two solutions y1 and
y2 , and the general solution will be a linear combination
y = C1 y 1 + C2 y 2
(5)
of the two solutions. (C1 and C2 are of course constants.) It should at this point be no surprise
that the general solution to equations of either of the forms (4) contain two arbitrary constants!
Our main focus as the chapter goes on will be solving initial value problems of the form
ay + by + cy = f (t),
y(0) = y0 , y (0) = y0 ,
(6)
with a, b and c being constants, a 6= 0. In addition, the initial values y0 and y0 are constants
as well. The method we will use to solve the IVP will consist of four steps:
(1) We will first solve the homogeneous equation obtained by replacing f (t) with zero. This
will give us a solution of the form (5), called the homogenous solution. We will denote it
by yh .
(2) Next well find something called a particular solution, denoted by yp , for the equation
(2). We will do this by a method called the method of undetermined coefficients.
(3) The general solution to the equation (2) is
y = yh + yp = C1 y1 (t) + C2 y2 (t) + yp (t),
(7)
78
3.1
In this section we will solve second order homogeneous equations of the forms
dy
d2 y
+ bx
+ cy = 0
and
ay + by + cy = 0,
(1)
2
dx
dx
where a 6= 0, b and c are constants. In both cases, solutions are obtained by guessing what the
general form of a solution might be and then substitution that guess into the ODE and making
it work. The first equation in (1) is called an Euler equation and the second is a second order,
linear, constant coefficient, homogeneous ODE. The bulk of this chapter is about constant coefficient equations. In this section well first see how to solve Euler equations, then look at constant
coefficient equations, whose solutions take a variety of forms.
Consider the Euler equation
x2 y + 2xy 6y = 0.
ax2
Note that if y was some power of x then y would be one power lower and y two powers
lower. If we then substituted our y, y and y into the ODE we would have three terms of the
same degree due to the multiplication of y by x and y by x2 in the ODE. Thus there might
be some hope that the three terms add up to zero.
Example 3.1(a): Solve the ODE x2 y + 2xy 6y = 0 by guessing a solution of the form
y = xp and determining the value of p.
First we compute y = pxp1 and y = p(p 1)xp2 . Substituting these into the ODE we get
x2 p(p 1)xp2 + 2xpxp1 6xp = 0
p(p 1)xp + 2pxp 6xp = 0
xp [p(p 1) + 2p 6] = 0
xp (p2 + p 6) = 0
Now the quantity xp is only zero when x = 0 and we want x to take on more values than
that, so it must be the case that p2 + p 6 = 0. Solving this gives us p = 3, 2, so both
y = x3 and y = x2 are solutions. The general solution is then y = C1 x3 + C2 x2 .
We now contemplate the second order linear, constant coefficient, homogeneous ODE
y + 5y + 6y = 0.
In this case we are looking for some function y = y(t) (we will be interested in applications
in which time is the independent variable) for which we can multiply the function and its first
and second derivatives by constants, add the results and get zero. This indicates that y must
essentially be equal to its first and second derivatives, which only occurs for exponential functions.
Therefore we guess that the solution has the form y = ert , where r is some constant.
79
(1)
When the auxiliary equation for a second order, linear, constant coefficient homogeneous equation has two real roots r1 and r2 , the solution to the ODE
is
y = C1 er1 t + C2 er2 t .
The most efficient method for finding the roots of the auxiliary equation r 2 + 5r + 6 = 0 from
Example 3.1(b) is to factor the left side, but we could have used the quadratic formula
b b2 4ac
(2)
r=
2a
instead. We will at times want or need to use the quadratic formula, and at other times factoring or
another method may be sufficient for finding the roots of the auxiliary equation. Regardless of what
we might do in practice, an examination of (2) is useful for determining the various possibilities
when obtaining the roots to the auxiliary equation:
When b2 4ac > 0 there will be two real roots, as in Example 3.1(b).
When b2 4ac = 0 there will only be one real root, which is a rather special case.
When b2 4ac < 0 the roots will be complex
conjugates; that is, they will have the form
r = k + i and r = k i, where i = 1.
When b2 4ac < 0 and b = 0, the roots will be two purely imaginary numbers r = i and
r = i. (This is just the previous case, with k = 0.)
80
Each of the above cases results in different forms of the solution to the homogeneous equation
(1), and the methods for solving the auxiliary equations are generally different as well, although
the quadratic formula could be used in all cases. Weve already taken care of the first case with
Example 3.1(b).
We now consider the second case above. Suppose that we were to try the method of Example
3.1(b) for the ODE y + 6y + 9y = 0. We would only get one value for r, 3. By the same
reasoning as used before we then have the solution y = C1 e3t + C2 e3t ; because these are like
terms, this can be written y = Ce3t , where C = C1 + C2 . But (for reasons well go into in
more depth later) the general solution to a second order ODE must be the sum of two different
functions, each multiplied by arbitrary constants, so something is wrong! The following example
shows that there is in fact another solution besides e3t .
Example 3.1(c): Verify that y = te3t is also a solution to the ODE y + 6y + 9y = 0.
Using the chain rule to compute the derivatives,
y = t(e3t ) + t e3t = 3te3t + e3t
y = 3(te3t ) 3e3t = 3(3te3t + e3t ) 3e3t = 9te3t 6e3t
Substituting into the ODE,
y + 6y + 9y = (9te3t 6e3t ) + 6(3te3t + e3t ) + 9(te3t )
= 9te3t 6e3t 18te3t + 6e3t + 9te3t
= 0
When the auxiliary equation for a second order, linear, constant coefficient homogeneous equation has only one real root r, the solution to the ODE is
y = C1 ert + C2 tert .
We now consider the case where b2 4ac < 0 and b = 0, resulting in two purely imaginary
roots r = i. Well consider the homogeneous equation y +9y = 0 that we looked at previously,
which can be rewritten as y = 9y. Through our familiarity with derivatives and the chain rule,
we guessed (correctly) that both y = sin 3t and y = cos 3t are solutions, and it is not hard to
show that y = C1 sin 3t + C2 cos 3t is a solution. Now how would it work to try the method of
Example 3.1(b) for this equation? Letting y = ert we have
y + 9y = r 2 ert + 9ert = (r 2 + 9)ert = 0
so we need to solve r 2 + 9 = 0 r 2 = 9. If we allow complex solutions, the solution to this
equation is r = 3i. To continue we will need the following:
81
ei = cos + i sin
Eulers Formula:
To see where Eulers Formula comes from, see Appendix A.5. We will also use the two basic
trig identities:
cos() = cos ,
sin() = sin
Here A and B are constants, and C1 = A + B, C2 = (A B)i. Now it seems that C2 should
be a complex number, which is a bit disconcerting. However, it is possible that A and B are
complex as well, in such a way that perhaps C2 turns out to be real! You will find that this
method works, regardless. The following summarizes what we have just seen.
When the auxiliary equation for a second order, linear, constant coefficient homogeneous equation has two purely imaginary roots r = i, the solution to
the ODE is
y = C1 sin t + C2 cos t.
Lets use a specific example to examine the final situation, where the auxiliary equation has
two complex roots r = k i.
Example 3.1(d): Solve y + 10y + 28y = 0
Guess y = ert , so y = rert and y = r 2 ert . Then
y + 10y + 28y = r 2 ert + 10rert + 28ert = (r 2 + 10r + 28)ert = 0
Again, ert cannot equal zero, so we solve r 2 + 10r + 28 = 0; this is done in this case using the
quadratic formula:
p
10 102 4(28)
= 5 2i 2
r=
2
Therefore
y = Ae(5+2i
=
=
=
2)t
2t
+ Be(52i
2)t
+ Be5t e2i 2 t
= Ae5t e2i
=
82
When the auxiliary equation for a second order, linear, constant coefficient homogeneous equation has two complex roots r = k i, the solution to the
ODE is
y = ekt (C1 sin t + C2 cos t).
Example 3.1(d) and the discussion before it show how sine and cosine functions arise when
r1 and r2 are complex numbers, but you need not show all those steps when solving. Unless
asked to show more, we will just set up and solve the equation ar 2 + br + c = 0, then give
the solution that arises from the form of the roots. In the Chapter 3 Summary you will find a
flowchart detailing the process of finding the solution to a second order, linear, constant coefficient,
homogeneous differential equation. You will need to memorize the forms of the solution for the
various forms of r1 and r2 , but if you do enough exercises, you should know them by the time
you are done.
Section 3.1 Exercises
1. Solve each Euler equation.
(a) x2 y 4xy + 4y = 0
(b) x2 y + 4xy + 2y = 0
(c) 3x2 y xy + y = 0
2. Solve each ODE, assuming the independent variable is t. Use exact values except for (i) use decimals rounded to the hundredths place there.
(a) y 2y 3y = 0
(b) y + 2y + 10y = 0
(d) y + 6y + 11y = 0
(e) y + 3y + 2y = 0
(f) y + 2y = 0
(g) y + 2y + y = 0
(h) y + 16y = 0
3. (a) Give the auxiliary equations for the ODEs y + 25y = 0 and y + 25y = 0.
(b) Solve the ODE y + 25y = 0.
4. Give the general form of the solution (that is, give one of the forms found in the boxes in this
section) to ay + by + cy = 0 under each of the following conditions:
(a) b = 0
(d) b2 4ac = 0
5. (a) Solve the Euler equation x2 y 3xy + 4y = 0. You should obtain only one solution.
(b) We know that there should be two solutions. Show that y = x2 ln x, x > 0, is also a
solution.
83
3.2
In the previous section we found out how to solve equations of the form
ay + by + cy = f (t)
(1)
when f (t) = 0, but in practice we are usually interested in situations where f (t) 6= 0. For those
situations we will use a technique called the method of undetermined coefficients to find a
solution. To do this we (again!) substitute a guess, which we will call the trial particular solution,
into the ODE. This trial solution will contain one or more unknown constants, whose values can be
determined by finding the result obtained when the trial solution is substituted into the left hand
side of the ODE, and then setting it equal to the known right hand side f (t). Like terms are then
equated, and the constants determined. The resulting solution is called the particular solution
to (1). Later we will see how the particular solution is combined with the homogeneous solution to
give the general solution to (1).
Lets consider the ODE
y + 9y = 5e2t .
If y had the form y = Ae2t for some constant A, then y would have the same form. Perhaps
there is then a choice of A for which y + 9y will equal 5e2t . The next example shows us that
is in fact the case.
Example 3.2(a): Find a value A for which y = Ae2t is a solution to y + 9y = 5e2t .
First we observe that
y = Ae2t
y = 2Ae2t
y = 4e2t .
2t
is a solution to y + 9y = 5e .
y = 13 e
A particular solution to an ODE is any solution that does not contain arbitrary constants, so
5 2t
e
is a particular solution to the ODE y + 9y = 5e2t . We sometimes subscript the
y = 13
dependent variable with the letter p to indicate a particular solution. For the above we would
5 2t
e . This distinction is made because, as we will soon see, there are other
then write yp = 13
solutions as well.
Next we will examine the ODE y + 7y + 10y = 5t2 8. We might guess that the particular
solution will be a fourth degree polynomial, because then y would be a second degree polynomial.
It turns out that there is no harm in trying a fourth degree polynomial (you will try it in the
exercises), but in fact a second degree polynomial is adequate. The next example demonstrates
this.
84
yp = 2At + B
= yp = 2A.
Next we substitute these values into the left side of the ODE and group by powers of t:
yp + 7yp + 10yp = 2A + 7(2At + B) + 10(At2 + Bt + C)
= 2A + 14At + 7B + 10At2 + 10Bt + 10C
= 10At2 + (14A + 10B)t + (2A + 7B + 10C)
Setting this equal to the right hand side of the ODE gives us
10At2 + (14A + 10B)t + (2A + 7B + 10C) = 5t2 + 0t 8,
and equating coefficients of powers of t (including the constant term) results in the three
equations
10A = 5,
14A + 10B = 0,
2A + 7B + 10C = 8.
From the first equation we see that A = 12 . Substituting that value into the second equation
7
. Finally, when we substitute the values we have found for
and solving for B results in B = 10
41
A and B into the last equation and solve for C we get C = 100
. The particular solution
to y + 7y + 10y = 5t2 8 is then
yp = 21 t2
7
10 t
41
100 .
It is important to note that even though the right hand side 5t2 8 contains only a t2 term and
a constant term, we need to include a t term in our guess for the particular solution. If we had
instead guessed a particular solution of the form yp = At2 + B and substituted it into the ODE
we would have obtained the two equations
10A = 5
and
14A = 0,
(1)
Based on what we have seen so far, our first inclination might be that we should consider a particular
solution of the form yp = A sin 2t. Lets try it:
yp = A sin 2t
so
yp = 2A cos 2t
yp = 4A sin 2t
Here we need A = 0 because there is no cosine term on the right hand side of (1), but then wed
have no sine term either!
So what do we do? Well, the trick is to let yp have both sine and cosine terms even though
the right side of the ODE has only a sine term.
85
Example 3.2(c): Determine values for A and B so that yp = A sin 2t + B cos 2t is the
particular solution to y + 4y + 3y = 5 sin 2t.
We see that
yp = 2A cos 2t 2B sin 2t,
so
Thus, in order for the left hand side to equal the right hand side, it must be the case that
8A B = 0 because there is no cosine term in the right hand side, and A 8B = 5, so
that the sine terms are equal. Solving the first equation for B and substituting into the second
8
1
. Substituting this back into the second equation gives B = 13
.
equation results in A = 13
8
1
Our particular solution is then yp = 13 sin 2t 13 cos 2t.
At this point we have the following guesses for a particular solution to an ODE of the form ay +
by + cy = f (t) when using the method of undetermined coefficients:
If f is a polynomial of degree n, then
yp = An tn + An1 tn1 + + A2 t2 + A1 t + A0
Note that all powers of t less than or equal to the degree of f (t) are included.
If f (t) = Cekt, then yp = Aekt .
If f (t) = C1 sin kt + C2 cos kt, then yp = A sin kt + B cos kt. Even if one of C1 or C2 is
zero, the trial yp must still contain both the sine and cosine terms.
In Section 3.5 we will see that there is a bit more to be added, but for now the above summarizes
what we have seen so far.
Section 3.2 Exercises
1. For each of the following, give the form the particular solution must have.
(a) y + 3y + 2y = 5t 1
(b) y + 6y + 9y = 5 cos 3t
(e) 2y + 3y + y = 7
(f) y + 3y = 2t + 3et
(c) y + 9y = 2e5t
(d) y 4y 5y = 6 sin t
4. You would think that the particular solution to y + 3y + 2y = 6et would have the form
yp = Aet , but that is not the case. In Section 3.5 we will see what our guess for the
particular solution should be. For now, try substituting the given particular solution into the
ODE to see what happens.
5. Solve each of the following homogeneous ODEs, assuming the independent variable for each
is t.
(a) y + 4y + 29y = 0
(b) 2y + 11y + 5y = 0
(c) y + 6y + 9y = 0
(d) y + 3y = 0
87
3.3
Differential Operators
Performance Criterion:
3. (d) Evaluate a differential operator for a given function.
An Example
In this section we will begin by exploring a specific second order ODE in order to illustrate
some ideas we will capitalize on in order to solve linear, constant coefficient, second order ODEs.
The ODE that we will be considering is
y + 9y = 5e2t ,
(1)
5 2t
e . (See Example 3.1(a).) This was
which we found to have a particular solution of yp = 13
2t
obtained by substituting a guess of y = Ae
for y in y + 9y and setting the result equal to
5 2t
2t
5e . The following example shows that y = 13 3
is not the only solution to (1).
5 2t
Example 3.3(a): Show that y = C sin 3t + 13
e , where C is any constant, is a solution
to (1).
10 2t
13 3
y = 9C sin 2t +
and
20 2t
.
13 e
20 2t
13 e
+ 9(C sin 3t +
20 2t
13 e
+ 9C sin 3t +
5 2t
)
13 e
45 2t
)
13 e
65 2t
13 e
= 5e2t
Thus y = C sin 3t +
5 2t
13 e
is a solution to y + 9y = 5e2t .
To examine further what is going on here it is convenient to develop some terminology and notation.
Differential Operators
A function can be thought of as a mathematical machine that takes in a number and, in
return, gives out a number. There are other mathematical machines that take in things other than
numbers, like functions or vectors, usually giving out things like what they take in. These sorts
of machines are often referred to as operators. A simple example of an operator that you are
quite familiar with is the derivative operator. When we take the derivative of a function, the result
is another function. To indicate the action of a derivative on a function y = y(t) we will write
dy
as
dt
d
(y),
dt
88
d
which is like the function notation f (x), with dt
taking the place of f and y taking the place
of x. The derivative operator has a very special property that you should be familiar with from
calculus. If a and b are any constants and y1 and y2 are functions of t, then
dy1
dy2
d
d
d
(ay1 + by2 ) = a
+b
= a (y1 ) + b (y2 ).
dt
dt
dt
dt
dt
(2)
Operators that distribute over addition and pass through constants like this are called linear
operators.
You might guess that the second derivative, and other higher derivatives, are linear operators as
well, and that is correct. We are particularly interested in operators that are created by multiplying
a function and some of its derivatives by constants and adding them all together. It is customary
to denote such operators with the letter D, for differential operator. An example would be
d
d2
D = 3 2 + 5 4, whose action on a function y = y(t) is defined by
dt
dt
D(y) = 3
dy
d2 y
+ 5 4y.
dt2
dt
(3)
D(y) = 3
When we combine several mathematical objects by multiplying each by a constant and adding
(or subtracting) the results, we obtain what is called a linear combination of those objects. The
operator D defined by (3) is a linear combination of a function and its first two derivatives. (We
can think of the function itself as the zeroth derivative, making all the things being combined
derivatives.) You have seen linear combinations in other contexts; any polynomial function like
1
f (x) = 3x4 7x3 + x2 x + 5.83
3
is a linear combination of 1, x, x2 , x3 and x4 . Those of you who have had a course in linear
algebra have seen linear combinations of vectors.
t2 3t
Again, we can think of an operator as a machine that takes in a
function and gives out some resulting function that is based somehow
on the input function. This is illustrated to the right for Example
D
3.2(b). The next example demonstrates that a differential operator
formed as a linear combination of derivatives is a linear operator.
Example 3.3(c): Show that the operator D defined by
D(y) = 3
dy
d2 y
+ 5 4y
2
dt
dt
89
8t2 + 22t 9
D(ay1 + by2 ) = 3
= 3a
d2 y2
dy1
dy2
d2 y1
+
3b
+ 5a
+ 5b
4ay1 4by2
dt2
dt2
dt
dt
d2 y1
dy1
d2 y2
dy2
+
5a
4ay
+
3b
+ 5b
4by2
1
2
2
dt
dt
dt
dt
2
2
d y1
dy1
dy2
d y2
= a 3 2 +5
4y1 + b 3 2 + 5
4y2
dt
dt
dt
dt
= 3a
= aD(y1 ) + bD(y2 )
Because D(ay1 + by2 ) = aD(y1 ) + bD(y2 ), D is a linear operator.
At the second line above we have applied the fact that the first and second derivative are linear
operators. With a bit of thought it should be clear that this, along with the distributive property,
is what makes a linear combination of derivatives a linear operator.
Back to the Example
We return now to considering the ODE
y + 9y = 5e2t ,
(1)
5 2t
13 e
and
y = C sin 3t +
5 2t
13 e
(4)
are solutions. If we now let D be the operator defined by D(y) = y + 9y, then (4) says that
5 2t
e ) = 5e2t
D( 13
and
D(C sin 3t +
5 2t
)
13 e
= 5e2t .
d2
(C sin 3t) + 9(C sin 3t) = 9C sin 3t + 9C sin 3t = 0.
dt2
5 2t
)
13 e
5 2t
13 e
is a solution; by linearity of
5 2t
= D(y1 ) + D( 13
e ) = 0 + 5e2t = 5e2t
for any function y1 for which D(y1 ) = 0 (like C sin 3t, for example). This gives us the following:
90
The general form of equation that we will work with in this chapter is
ay + by + cy = f (t),
(5)
where a, b and c are constants. Our goal will be to find a general solution to this equation,
meaning a solution that encompasses all possible solutions. Such a solution will consist of a solution
to (5) that contains no arbitrary constants plus the sum of all possible solutions to
ay + by + cy = 0.
(6)
The solution to (5) without arbitrary constants is called the particular solution to the equation,
and the sum of all possible solutions to (6) is called the homogeneous solution. We will denote
the particular solution by yp and the homogenous solution by yh ; the general solution is then
y = yh + yp . For equations like (5) or (6) the homogeneous solution will always consist of two
different (in a sense that we will discuss further in a later section) functions, each multiplied by an
arbitrary constant and then added together.
As an example, we will find that the ODE
y + 9y = 5e2t
has homogeneous solution yh = C1 sin 3t + C2 cos 3t and particular solution yp =
general solution is
5 2t
e .
y = yh + yp = C1 sin 3t + C2 cos 3t + 13
5 2t
,
13 e
so the
For reasons you will see in Section 3.6, we will find the homogeneous solution first using methods
described in the next section. We will then find the particular solution in the manner that we did
in Example 3.1(a).
Section 3.3 Exercises
1. Let the differential operator D be defined on a function y = y(t) by
D(y) =
d2
d
(y) + 3 (y) + 2y.
2
dt
dt
(b) y = 5e2t
(c) y = 5 cos 2t
(d) y = 25 t
17
4
2. What does your answer to Exercise 1(d) tell us about the ODE y + 3y + 2y = 5t 1?
3. Although you dont realize why at this point, your answer to Exercise 1(b) is somewhat
special.
(a) Does the same thing happen for y = Ce2t , where C is some constant other than 5?
If so, for what value or values of C?
(b) Does the same thing happen if y = ekt , where k is some constant other than 2? If
so, for what value or values of k?
91
d2
d
(y) + 6 (y) + 9y.
2
dt
dt
(b) y = te3t
(1)
where a and b are any constants and f and g are any functions of the sort that S can
act on. (1) is equivalent to the two separate conditions that
S(af ) = aS(f )
and
(2)
where a, f and g are as before. That is, if both conditions in (2) hold for S, then it is a
linear operator. (There are operators that are not linear.) Now lets define a specific operator
S by S(f (t)) = f (t) + 3 for any function f (t).
(a) What is S(cos t)? S(t2 + 5t 1)?
(b) What is S(4 cos t)? What is 4S(cos t)? Are both results the same? What does that
tell us about S, in terms of linearity?
(c) Find S(cos t + e2t ) and S(cos t) + S(e2t ).
92
3.4
Now that we know how to find the general solution to ay + by + cy = f (t) we are ready to
solve initial value problems whose ODEs are of this form. Here is how the process goes:
We have already covered the first three steps of the above. It is very important to remember that
the initial conditions are applied to the general solution to find the values of the arbitrary
constants. A common mistake by students is to find the values of the constants based on only the
homogeneous solution - this is incorrect.
Lets look at an example:
Example 3.4(a): Solve the IVP
y + 4y + 3y = 5 sin 2t, y(0) = 2, y (0) = 1
We must first solve the homogeneous equation y + 4y + 3y = 0. The roots of the auxiliary
equation are r1 = 1 and r2 = 3 (of course it doesnt matter which is which), so the
homogeneous solution is yh = C1 et + C2 e3t . Because the right hand side is not of the same
form as any part of the homogeneous solution, our trial particular solution is yp = A sin 2t +
B cos 2t. This gives us
yp = 4A sin 2t 4B cos 2t.
LHS = (4A sin 2t 4B cos 2t) + 4(2A cos 2t 2B sin 2t) + 3(A sin 2t + B cos 2t)
= (A 8B) sin 2t + (8A B) cos 2t
Thus, in order for the left hand side to equal the right hand side, it must be the case that
8A B = 0 because there is no cosine term in the right hand side, and A 8B = 5, so
that the sine terms are equal. Solving the first equation for B and substituting into the second
93
1
8
equation results in A = 13
. Substituting this back into the second equation gives B = 13
.
8
1
Our particular solution is then yp = 13 sin 2t 13 cos 2t, and the general solution is
y = yh + yp = C1 et + C2 e3t
1
13
sin 2t
8
13
cos 2t
2
13
cos 2t +
16
13
sin 2t
8
13
=2
and
C1 3C2
91 t
26 e
23 3t
26 e
1
13
2
13
= 1
sin 2t
8
13
cos 2t
It is often the case that finding the constants C1 and C2 comes down to solving a system of two
equations in two unknowns, as it did here.
Section 3.4 Exercises
1. You may have found the particular solution to each of the following ODEs in Exercise 2 of
Section 3.2. Give the general solution to each.
(a) y + 3y + 2y = 5t 1
(b) y + 6y + 9y = 5 cos 3t
(c) y + 9y = 2e5t
(d) y 4y 5y = 6 sin t
(e) 2y + 3y + y = 7
(f) y + 3y = 2t + 3et
2. Solve each of the following IVPs by the process described in the box at the start of the section.
(a) y + 9y = 4 sin t ,
y (0) = 4
y(0) = 2 ,
(b) y + 4y + 4y = 5e3t ,
y(0) = 0 ,
y (0) = 0
y(0) = 2 ,
y (0) = 8
94
3.5
Performance Criteria:
3. (a) Find a particular solution to a second order linear, constant coefficient
ODE using undetermined coefficients.
(e) Given minimal, but sufficient, information about an ODE of the form
ay + by + cy = f (t) and its general solution, determine whether a
function is a solution to ay + by + cy = f (t) or ay + by + cy = 0.
At this point you have seen the entire process for solving initial value problems for second order,
linear, constant coefficient differential equations. In this section we see one difficulty that can arise,
and how to handle such situations. We begin with an example.
Example 3.5(a): Determine the values of A and B for which
yp = A sin 3t + B cos 3t
is the particular solution to the ODE y + 9y = 2 sin 3t.
As usual, we begin by finding the derivatives of yp :
yp = 3A cos 3t 3B sin 3t
We then have
yp + 9yp = 9A sin 3t 9B cos 3t + 9(A sin 3t + B cos 3t) = 0,
so there are no values of A and B for which y = A sin 3t + B cos 3t is a solution to
y + 9y = 2 sin 3t.
The problem here is that the homogeneous solution to y +9y = 2 sin 3t is yh = C1 sin 3t+C2 sin 3t.
d2
Thus we cannot hope to obtain 2 sin 3t when applying the operator D = 2 + 9 to y =
dt
A sin 3t + B cos 3t, as the result is always zero. However, we will find that a different guess for
yp will give us the particular solution that we seek:
Example 3.5(b): Determine the values of A and B for which
yp = At sin 3t + Bt cos 3t
is the particular solution to the ODE y + 9y = 2 sin 3t.
We use the product rule carefully to find the derivatives of yp :
yp = 3At cos 3t + A sin 3t 3Bt sin 3t + B cos 3t
and
yp = 9At sin 3t + 3A cos 3t + 3A cos 3t 9Bt cos 3t 3B sin 3t 3B sin 3t.
95
Undetermined Coefficients
Consider the constant coefficient ODE ay + by + cy = f (t), and assume
yh contains no terms that are constant multiples of f (t). The trial particular
solution yp is chosen as follows.
If f is a polynomial of degree n, then
yp = An tn + An1 tn1 + + A2 t2 + A1 t + A0
If f (t) = Cekt , then yp = Aekt .
If f (t) = C1 sin kt + C2 cos kt, then yp = A sin kt + B cos kt. Even if
one of C1 or C2 is zero, yp still contains both the sine and cosine
terms.
When yh contains any term that is a constant multiple of f , yp will be as
above but multiplied by the smallest power of t for which no terms of yp are
of the same form as any terms of yh .
96
We conclude this section by further examining homogeneous and particular solutions to an ODE
ay + by + cy = f (t).
(1)
Lets denote the left side of (1) using the operator notation D(y). We have found that the
homogeneous solution consists of a linear combination of two functions g(t) and h(t) that are
both, by themselves, solutions to D(g) = 0 and D(h) = 0. By a linear combination we mean
yh = C1 g(t) + C2 h(t),
where C1 and C2 are ANY constants. When D is applied to the particular solution yp the
result is D(yp ) = f (t). The general solution is
y = C1 g(t) + C2 h(t) + yp (t).
Applying D to the solution then gives
D(C1 g + C2 h + yp ) = C1 D(g) + C2 D(h) + D(yp ) = 0 + 0 + f (t) = f (t).
Note the use of the fact that D is a linear operator in this computation.
Example 3.5(f ): The general solution to (1) is y = C1 e2t + C2 et + 4 cos 5t. Which of
the following are solutions to ay + by + cy = 0?
(a) y = 5et
The homogeneous solution is the part containing the arbitrary constants, yh = C1 e2t + C2 et .
It is a solution to ay +by +cy = 0 for all choices of C1 and C2 , so the functions in (a) and (c)
are both solutions. The function in (b) is a solution to (1), but not to ay + by + cy = 0 because
D applied to 7e2t is zero, but when applied to the particular solution yp = 4 cos 5t the result
is f (t), not zero.
Example 3.5(g): The general solution to (1) is y = C1 e2t + C2 et + 4 cos 5t. Which of
the following are solutions to (1)?
(a) y = 4 cos 5t
(b) y = 8 cos 5t
We again recognize that the homogeneous solution is yh = C1 e2t + C2 et and the particular
solution is yp = 4 cos 5t. Because the particular solution by itself is a solution to (1), the function
in (a) is a solution. Unlike the homogeneous solution, a constant in the particular solution is not
arbitrary, so the function in (b) is not a solution. The function in (c) is a solution, because it is
simply the general solution with the arbitrary constants having the specific values C1 = 7 and
C2 = 5.
97
(d)
yp = At + b
6y = 7 cos 5t,
(e) y + y 6y = 4e2t ,
yp = A sin 5t + B cos 5t
yp = Ate2t
2. Suppose that when you were finding the particular solution to y + y 6y = 4e2t you didnt
notice that 4e2t was of the same form as one of the terms of yh . Try a particular solution
of yp = Ae2t and see what happens. It will try to tell you that something is wrong!
3. Solve each of the following IVPs by the process described in the box at the start of Section
3.4.
(a) y + y 6y = 1 + 8t 6t2 ,
(b) y + 7y + 10y = 6e2t ,
(c) y + 4y = 3 sin 2t ,
y (0) = 3
y(0) = 2 ,
y (0) = 11
y(0) = 2 ,
y(0) =
1
2
y (0) =
5
2
4. The functions below are solutions to second order linear, constant coefficient initial value
problems. Give the steady-state and transient parts of each.
(a) y = 32 sin 3t + 53 cos 3t
(c) y =
3
5
98
3
4
cos 7t
3.6
(f) Determine, from the definition, whether two functions are linearly
independent.
(g) Use the Wronskian to determine whether two solutions to a second
order linear ODE are independent.
(h) Given one solution to a second order ODE, use reduction of order to
find a second solution.
(1)
If (1) is true only when both c1 and c2 are zero, then f and g are linearly
independent on [a, b].
The expression c1 f (x) + c2 g(x) above is called a linear combination of the functions f and g.
Example 3.6(a): Show that the two functions y = e2t
independent for all values of t.
and y = et
are linearly
Suppose that c1 e2t +c2 et = 0 for some c1 and c2 not equal to zero. Then e2t (c1 +c2 et ) = 0,
but e2t is never zero so it must be the case that c1 + c2 et = 0. This gives us et = c1 /c2 ,
which says that et is constant. Since that is not the case, y = e2t and y = et are linearly
independent.
99
There may be times that it is difficult to tell, using the definition above, whether two functions
are linearly independent. In those cases we can use a new function created from the two functions,
called the Wronskian, to determine whether the functions are linearly independent.
Those of you who have had linear algebra may recognize the Wronskian as the determinant of the
2 2 matrix
"
#
f (x) g(x)
f (x) g (x)
Our interest is in determining whether two solutions to an ODE are linearly independent. Note
that any linear second order homogeneous ODE with independent variable t can (almost always,
and definitely in the case that p and q are constant) be written in the form
y + p(t)y + q(t)y = 0
(2)
The following tells how the Wronskian is used to determine whether two solutions to an equation
of the form (2) are linearly independent.
Example 3.6(b): Show that the two solutions y = sin 3t and y = cos 2t of y +9y = 0 are
linearly independent for all values of t.
The Wronskian of these two functions is
W (t) = (sin 3t)(cos 3t) (cos 3t)(sin 3t) = sin2 3t cos2 3t = 1,
which is clearly not zero for any value of t. Therefore y = sin 3t and y = cos 2t are linearly
independent for all values of t.
100
Note that the above says that the general solution is a linear combination of the two solutions
y1 and y2 .
In the exercises you will show that y = et and y = tet are linearly independent solutions
to y + 2y + y = 0, so the above tells us that the general solution is then y = C1 et + C2 tet .
Its now time to find out where the solution y = tet comes from.
Reduction of Order
Reduction of order is a method for finding a second solution to a second order differential
equation when one solution is already known. Our main interest in this is finding the second
solution when we have repeated roots, so we will not go into the method in excessive detail.
Perhaps the best way to introduce the method is through an example. The two key ideas are these:
We will assume that if y1 = y1 (t) is a solution, then the second solution has the form
y2 (x) = u(t)y1 (t) for some function u. We then substitute y2 into the ODE, which results
in a new ODE for u.
The new ODE for u will contain u and u terms, but no u term. If we let v(t) = u (t) then
v (t) = u (t), and making these two substitutions we get a first order equation in v. (This
is where the name reduction of order comes from - weve reduced a second order equation to
a first order equation.) We solve that for v, then solve u (t) = v(t) to get u.
Now lets get to that example!
Example 3.6(c): Use the solution y1 (t) = et and reduction of order to find a second
solution to y + 3y + 2y = 0.
We begin by assuming y2 = u(t)et . Then (using integration by parts),
y2 = u(t)et + u (t)et
and
so v (t) = u (t) and this last ODE becomes v (t)+ v(t) = 0. This equivalent to v (t) = v(t),
so v(t) = C1 et .
We now replace v(t) with u (t) to obtain u (t) = C1 et . The solution to this is u(t) =
C2 et + C3 ; for reasons to be given later, we can take C2 to be any non-zero value and C3 can
have any value. Well take C2 = 1 and C3 = 0. Therefore y2 (t) = u(t)et = et et = e2t .
Disregarding the constant (because we will replace it when adding this solution to the one given),
we have y2 = e2t .
Forming the linear combination of the given solution and the one that we found using it, we get
y = aet + be2t for constants a and b. We now examine the way that the constants C2 and
C3 were handled in the above. Lets see what would have happened if we had not let C2 = 1 and
C3 = 0. In that case we would have had
y2 = u(t)y1 (t) = (C2 et + C3 )et = C2 e2t + C3 et .
When we then form a linear combination of y1 and y2 using constants A and B, well get
y = Aet + B(C2 e2t + C3 et )
= Aet + BC2 e2t + BC3 et
= (A + BC3 )et + BC2 e2t
= aet + be2t ,
where a = A + BC3 and b = BC2 . If we keep the constants C2 and C3 , they essentially get
absorbed into the constants for the linear combination of the two solutions.
In Example 3.6(c) there was no need to use reduction of order to determine a second solution
y2 = e2t from the first solution y1 = e2t ; we could arrive at both solutions via the auxiliary
equation method. However, the above example demonstrates how the method works. In the
exercises you will encounter exercises where you will again be asked to find a second solution by
this method when it is unnecessary, but you will also use it for situations where the second solution
(and maybe the first as well) cant be obtained by methods we have used so far. You will also
use reduction of order to find the second solution to y + 2y + y = 0, knowing the first solution
y1 = et , which is obtained by the auxiliary equation method.
Section 3.6 Exercises
1. For each pair of functions, determine whether they are linearly independent, using the definition. If they arent, give nonzero constants c1 and c2 for which c1 f (x) + c2 g(x) = 0 for
all real numbers x.
(a) f (x) = 3x2 5, g(x) = 2x + 1
2. Use the facts that cos(x) = cos x and sin(x) = sin(x) for the following.
(a) Give nonzero constants c1 and c2 such that c1 cos x + c2 cos(x) = 0. Are cos x and
cos(x) linearly independent?
(b) Repeat part (a) for sin x and sin(x).
102
3. For each pair of functions in Exercise 1 that you determined were linearly independent, give
the Wronskian and one value of x for which it is not zero.
4. Find the Wronskian for y1 = ekt and y2 = tekt (where k is any nonzero constant) and
give a value of t for which it is not zero. What does this tell us about the functions y1 and
y2 ?
5. Use the Wronskian to determine whether ex and ex are linearly independent.
6. Consider the ODE y + 8y + 15y = 0.
(a) Given that one solution is y1 = e5t , use reduction of order to find another solution.
(b) Use the auxiliary equation to find both solutions, to check your answer to (a).
7. Using the auxiliary equation method with y + 2y + y = 0, we get the single solution
y1 = et . Use reduction of order to obtain the second solution y2 = tet .
8. Given that one solution to 2x2 y + xy 3y = 0 is y1 = x1 , find a second solution y2 .
9. Given that one solution to x2 y + 2xy 2y = 0 is y1 = x, find a second solution y2 .
103
3.7
Chapter 3 Summary
With the exception of the methods for solving Euler equations and finding solutions by
reduction of order, this chapter was primarily concerned with solving initial value problems
of the form
ay + by + cy = f (t),
y(0) = y0 , y (0) = y0 ,
(1)
with a, b and c being constants, a =
6 0. Euler equations and reduction of order will
be addressed at the end of this section. For now we focus on the procedure for solving the
initial value problem (1):
Undetermined Coefficients
Consider the constant coefficient ODE ay + by + cy = f (t), and
assume yh contains no terms that are constant multiples of f (t). The
trial particular solution yp is chosen as follows.
If f is a polynomial of degree n, then
yp = An tn + An1 tn1 + + A2 t2 + A1 t + A0
If f (t) = Cekt , then yp = Aekt .
If f (t) = C1 sin kt + C2 cos kt, then yp = A sin kt + B cos kt. Even
if one of C1 or C2 is zero, yp still contains both the sine and
cosine terms.
When yh contains any term that is a constant multiple of f , yp will be
as above but multiplied by the smallest power of t for which no terms
of yp are of the same form as any terms of yh .
Differential Equation
ay + by + cy = 0
Auxiliary Equation
ar 2 + br + c = 0
Solve by factoring
or by using the
quadratic formula
yes
b2
Is
4ac 0?
no
Are
there two roots
r1 , r2 ?
yes
Solution
to the ODE is
y = C1 er1 t + C2 er2 t
no
Is b = 0?
yes
no
Solve for r = i
by isolating r 2 and
and taking the square
root of both sides
105
3.8
Chapter 3 Exercises
d2 y
dy
+x
y =0
dx2
dx
(b) 6x2
d2 y
dy
+ 11x
+y = 0
dx2
dx
dR
d2 R
+r
n2 R = 0
2
dr
dr
(1)
which arises in the study of the equilibrium distribution of heat in a circular disk. As pointed
out before, r and R are two different variables; R is the dependent variable, and is a
function of the independent variable r. In Chapter 2 we solved the equation for the case
n = 0. Solve it for any integer n 6= 0.
3. The ODE ay + by + cy = 0 is only second order if a 6= 0. We saw in Section 3.3 what the
solution to the ODE looks like when b = 0. In this exercise and the next we will solve, by
two different methods, an equation in which c = 0
(a) In this exercise we will solve 2y + 3y = 0. begin by making the substitution y = x
(so what then is y ?) and solving the resulting first order ODE for x.
(b) To determine y you will now need to integrate your answer to (a). DO that, remembering that your final solution needs to contain two arbitrary constants.
4. Solve 2y + 3y = 0 by assuming y = ert and following a process like that done for the
various scenarios in Section 3.1.
106
107
Our practical applications so far have been focused on first order ODEs and their associated initial value problems. We have seen how such equations model RL electric circuits, mixing problems,
growth of populations and decay of radioactive elements, and the temperature of a solid object in
some constant-temperature medium.
There are other things we would like to model that require not first order ODEs, but second
order ODEs of the form
ay + by + cy = f (t)
(1)
that we learned how to solve in the last chapter. We will proceed through a series of variations on
this equation:
108
In Section 1.2 we saw how a system consisting of a mass on a spring is modelled by the second
order ODE
d2 y
d2 y
m 2 = ky
or
m 2 + ky = 0,
dt
dt
which is (1) with b = 0 and f (t) = 0. When b = 0 we refer to the system as undamped,
and when f (t) = 0 the system is free. We begin our study of applications with undamped,
unforced systems in Section 4.1.
In Section 4.2 we will add damping, but still no forcing function. That is, well have b 6= 0 and
f (t) = 0. In this case well see that there are three possible scenarios, the system is just
one of (a) under-damped, overdamped, or critically damped. In Section 4.2 we will
also introduce an electrical circuit for which the mathematical model is exactly the same as
a mass on a spring.
Next, in Section 4.3, we will consider systems that are both damped (so b 6= 0 and forced
(so f (t) 6= 0).
Finally, well go back to undamped systems, but with nonzero forcing functions f (t), which
are often sine or cosine functions. This will give rise to two phenomena called beats and
resonance. These things will be studied in Section 4.4.
As a specific example of an ODE of the form (1), lets consider
y + 5y + 6y = 5 sin 3t.
(2)
We want to know not only how to solve such ODEs and their associated IVPs, but also to know
what kind of behavior to expect from solutions, based on the original ODE or IVP. Here is a
summary of some of the ideas you will run across, which were previously brought up in Chapter 2:
The left hand side of the ODE represents some sort of system like a mass on a spring
or a simple electric circuit. The function on the right represents some kind of input to the
system, which we call a forcing function, since it forces movement (for a mass on a spring)
or current (for an electrical circuit) in the system.
The solution function y is the response of the system, and will often consist of exponential
or trigonometric functions, or products of the two. The solution may have several terms,
some of which may die out (go to zero) as time goes on, and others that may not. The
parts that die out are called transient parts of the solution, and parts that are constant
or periodic are called steady-state parts. These are also referred to as the transient and
steady-state response of the system.
109
4.1
spring
mass
d2 y
+ ky = 0
dt2
(1)
that governs the motion of the mass. Here y is the height (from equilibrium) of the mass at any
time t after it is set in motion by some means, with positive being upward. m is the mass of
the mass (the first of these being a quantity, and the second being an object) and k is the spring
constant. We will assume that there is no external force acting on the mass once it is set in motion
- in such cases we call the vibration free. We will also consider the system to be undamped,
which means that there is nothing hindering the motion of the mass once it begins moving up and
down.
First lets remind ourselves of how we can solve such ODEs:
Example 4.1(a): Solve Equation 1 for a mass of 10.3 and a spring constant of 28.7 (both
with appropriate units).
The auxiliary equation corresponding to the ODE is 10.3r 2 + 28.7 = 0. Subtracting 28.7 from
both sides and dividing by 10.3 gives r 2 = 2.786.... If we then take the square root of both
sides and round to the nearest hundredth, we have r = 1.67i. Therefore the solution to the
ODE is y = C1 sin 1.67t + C2 cos 1.67t.
110
y (in)
t (sec)
A mass hanging on a spring can be set in motion by doing one of three things:
moving the mass away from its equilibrium position and letting it go
giving the mass some initial velocity upward or downward from its initial position
moving the mass away from its initial position and giving it an initial velocity
If we move the mass upward or downward, that will give y(0) equal to some value other than
zero, and if we impart an initial velocity the value of y (0) will be nonzero.
Example 4.1(b): Suppose that the mass from Example 4.1(a) is set in motion by lifting it
up one inch and giving it an initial velocity of 2 inches per second downward, both at time
zero. Give the initial conditions in function form and sketch a graph of what you expect the
motion to look like. Then determine the solution to the initial value problem and graph it to
check your graph.
y (in)
The condition of raising the mass up by one inch is given
by y(0) = 1, and the initial velocity of two inches per
second downward is given by y (0) = 2, with the
1
negative indicating that the velocity is in the downward
t (sec)
direction. Because the mass starts above its equilibrium
point, we expect the y-intercept of the graph to be at
positive one. The condition of being given an initial
velocity downward means that the slope of the tangent
line to the graph at t = 0 will be negative. We therefore
expect a graph something like that shown to the right.
Applying the first initial condition with the solution from Example 4.2(a) gives C2 = 1. Taking
the derivative of the solution obtained in Example 4.2(a) gives
y = 1.67C1 cos 1.67t 1.67C2 sin 1.67t.
Substituting the initial velocity initial condition into this gives us C1 = 1.20 (the zero indicates
accuracy to the hundredths place), so the solution to the IVP is
y = 1.20 sin 1.67t + cos 1.67t.
The graph of this function agrees with that shown above.
It is sometimes useful to change an expression of the form 1.20 sin 1.67t + cos 1.67t into a
single sine function with a phase shift, and here is what we use to do it:
111
C sin(t + ) Form
A sin t + B cos t = C sin(t + ), where C =
and = tan1
B
A
if
= + tan1
A > 0,
B
A
A2 + B 2
if
A<0
= + tan1
(1.67)2 + 12 = 1.95,
B
= 2.60
A
Note that the amplitude of the motion for the situation from Examples 4.1(a), (b) and (c) is 1.95,
almost twice the initial position of 1 unit. That is due to the initial velocity - you will see in the
exercises what happens if there is no initial velocity.
There are some important parameters associated with a function y = C sin(t + ):
T =
f=
2
is the period
=
is the frequency
T
2
= 2f
Example 4.1(d): For the function y = 1.95 sin(1.67t + 2.60), give the amplitude, period,
frequency and phase shift.
The amplitude is 1.95, the period is T =
the phase shift is 2.60
1.67 = 1.56.
2
1.67
= 0.64 and
With a bit of careful thought, the following should be clear: Only the amplitude and and phase
shift are influenced by the initial conditions. The period, frequency and angular frequency in this
case are all determined by m and k, the parameters of the system.
112
113
2
.
(c) We use the symbol T for the period of the motion, which is obtained from T =
What is the period of motion for Exercises 4 and 5? Give units with your answer they should be the appropriate time units.
1
=
. Find the
(d) The frequency of the motion is denoted by f , and it is given by
T
2
frequency for exercises 4 and 5.
4
kg on a spring with spring constant 4 N/m is given an initial velocity of
7. A mass of 10
9 cm/sec upward from an initial position of 4 cm below equilibrium.
(a) Find the equation of motion of the mass, y(t), in the form y = A sin(t + ).
(b) What are the period and frequency of the motion?
114
4.2
d2 y
dy
= f (t)
ky.
dt2
dt
dy
d2 y
+
+ ky = f (t),
2
dt
dt
(1)
the second order differential equation that models the motion of a spring-mass system with damping
and an external forcing function f .
Now suppose that we have an electric circuit consisting of a
L
resistor, an inductor and a capacitor in series with each other,
along with (perhaps) a voltage source. We will refer to such a
E
C
circuit as an RLC circuit, shown schematically to the right.
The differential equation that models an RLC circuit is derived
from Kirchoffs voltage law which tells us that the sum of the
voltage drops across each of the three components is equal to the
R
voltage imposed on the system by the voltage source. By Ohms
Figure 4.2
law the voltage drop across
the resistor is V = iR, where i is the current and R is the resistance of the resistor. The
voltage drops across the inductor and the capacitor are
115
1
di
and
q,
dt
C
where q is the charge on the capacitor. From Kirchoffs voltage law we now have
L
di
1
+ Ri + q = E(t),
dt
C
(2)
where E(t) is the voltage as a function of time. (It may of course be constant, such as in the case
of DC voltage supplied by a battery.)
dq
But the current i is the rate at which charge is passing by a point in the circuit: i =
Thus
dt
2
d dq
d q
di
=
= 2 and the above equation becomes
dt
dt dt
dt
L
d2 q
dq
1
+R
+ q = E(t) ,
2
dt
dt C
(3),
where L is the inductance of the inductor in henries, R is the resistance of the resistor in
ohms, and C is the capacitance of the capacitor in farads. All are positive quantities; this will be
important! If we would rather work with current rather than charge we can differentiate both sides
of (3) to get
di
1
d2 i
+ i = E (t) ,
(4),
L 2 +R
dt
dt C
Note that both (3) and (4) are completely analogous to equation (1), which models a spring-mass
system. This illustrates the principle we first encountered in Section 2.5 that different physical
situations are often modelled by the same differential equation.
In this section we will consider the situation in which there is no forcing function. That is, the
right hand sides of (1), (3) and (4) are zero. Of course something needs to happen to get the mass
moving or to get current to flow in the circuit, and each can be accomplished in a variety of ways.
For example, the spring and the capacitor both have the ability to store energy by compressing or
stretching the spring, or by storing charge in the capacitor (with the positivity or negativity of that
charge being analogous to compressing or stretching the spring). As that energy is released it will
cause the mass to move or current to flow in the circuit. Here is a summary of initial conditions
we can have for the spring, which we previously gave in Section 4.1:
the mass can be displaced and let go, with no initial velocity
the mass can be given some velocity at its resting position
the mass can be displaced AND given some initial velocity upward or downward
The analogous conditions for the electric circuit are as follows:
there is no current in the circuit, but there is an initial charge on the capacitor
there is no charge on the capacitor, but there is an initial current
there is both an initial charge on the capacitor and an initial current
Our main objective in this section is to understand the behavior of the solution function y(t) of
equation (1), (3) or (4), and how that behavior varies depending on the parameters m, k and or
R, L and C. You will be looking at some exercises to do this. Realize that the values given for
the parameters may not be realistic - they were chosen in such a way as to make the mathematics
reasonable.
116
118
4.3
d2 y
dy
+ ky = f (t)
+
2
dt
dt
or
d2 q
1
dQ
+ q = E(t)
+R
2
dt
dt
C
when none of , R, f (t) or E(t) are zero. This is what we call forced, damped vibration, an
example of which you will see in the following exercise.
Section 4.3 Exercises
1. Consider the second order initial value problem
y + 2y + 10y = 9.4 sin t ,
y(0) = 5 ,
y (0) = 0,
which could model either a spring-mass system or an RLC circuit. In this exercise you will
find the solution to this initial value problem, and you will investigate its behavior. Recall
the process for solving such an equation:
Find the homogeneous solution yh . It will contain two constants.
Use undetermined coefficients (guessing) to find the particular solution yp to the equation.
Apply the two initial conditions to determine the values of the unknown constants. Be
sure to conclude by writing your final solution to the IVP.
(a) Carry out the above process for the given IVP. Give all numbers as decimals,
rounded to the tenths place.
(b) Graph the solution on your calculator, for t = 0 to t = 10. Sketch your graph.
(c) Look carefully at your solution (the equation itself, not the graph). Recall that the
transient part of a solution is any part that goes to zero as time goes on. Any part that
does not go to zero over time is called the steady-state part of the solution, or just the
steady-state solution. Give the transient and steady-state parts of your solution, telling
clearly which is which.
(d) Graph the steady-state solution together with the complete solution that you already
graphed. The complete solution should approach the steady state solution as time goes
on. Add the graph of the steady-state solution to your graph from (b).
119
4.4
Performance Criterion:
4. (a) Set up and solve second order initial value problems modelling
spring-mass systems and RLC circuits.
(g) Determined whether a forced, undamped system will exhibit resonance, beats, or neither.
x(0) = x (0) = 0
(b) Graph the solution using some technology, sketch the graph. Be sure to get a viewing
window that is appropriate. Put a scale on your graph.
(c) Discuss the situation of transient and steady-state solutions. Why should we expect this
before even solving the differential equation?
2. (a) Solve the initial value problem
x + 4.84x = 8 cos 2.2t ,
x(0) = x (0) = 0
(b) Graph the solution using some technology, sketch the graph.
(c) The phenomenon you are observing here is called resonance. In either the mechanical
or electrical case, as the amplitude gets larger and larger, something will fail - the spring
or one of the electrical components. What is it about the situation that is causing this
to happen?
Note that the angular frequency of 2.2 that appears in the solution to the IVP comes from the
homogeneous equation, so it depends only on the spring-mass or LC system, not on the forcing
function. That frequency is sometimes called the natural frequency of the system.
120
x(0) = x (0) = 0
121
4.5
Chapter 4 Summary
Performance Criteria:
4. (h) For a spring-mass system or electric circuit, demonstrate an understanding of the relationships between
the physical situation (presence and type of damping and/or
forcing)
the form of the ODE, including the function f
the the analytic and graphical nature of the solution (in particular, the presence and appearance of transient and steady-state
parts of the solution)
In the previous chapter we saw how to solve second order linear ODEs with constant coefficients,
and their associated initial value problems. In this chapter we have seen two applications, springmass systems and electric circuits, and investigated various combinations of initial conditions and
forcing functions for both undamped and damped systems. In this section we will attempt to
summarize all that we have seen. In particular, we want to recognize from an ODE, the solution
to an ODE, or the graph of the solution to an ODE whether it models a situation
in which the system is undamped, under-damped, critically damped or over-damped
with or without an external forcing function
for which the solution has transient or steady-state parts, or both
resulting in or exhibiting resonance or beats
The type of differential equation that we are talking about here is one of the form
ay + by + cy = f (t)
(1)
where the coefficients a, b and c are constant parameters based on the physical properties of the
system we are considering:
In a spring-mass system a is the mass, b is the coefficient of damping, and c is the spring
constant.
In an electric circuit, a is the inductance, b is the resistance and c is the reciprocal of the
capacitance.
It is clear that without a mass and a spring there is no spring-mass system, so for that situation
neither a nor c can be zero. For the electric circuit situation it is reasonable to consider a system
with only a resistance and inductance, but that can be treated as a first order ODE in a manner
you have already seen. For an RLC circuit of the sort we wish to consider, none of the values a,
b or c are zero, although we will consider the case b = 0 as a theoretical possibility.
122
124
Chapter 4 Exercises
1. Here are some equations of the sort we have been discussing:
(i) y + 3y + 2y = 0
(ii) y + 4y = sin(9t)
(iii) y + 10y + 25y = 0
(iv) y + 25y = 3.1 sin(5t)
(v) y + 9y = 0
(vi) y + 5y + 7y = 7.4 sin(2.4t)
(vii) y + 3y + 5y = 0
(viii) y + 16y = 7 cos(3.8t) 4 sin(3.8t)
(a) Which equations model an undamped system? Which model an under-damped system?
Critically damped? Over-damped?
(b) Which equations will have solutions with a transient part? Which will have solutions
with a steady-state part?
(c) Which equations will have solutions that exhibit resonance? Which will have solutions
exhibiting beats?
2. Consider the following solutions to differential equations of the type we have been discussing
(second order, constant coefficient).
(i) y = C1 et + C2 e3t
(ii) y = C1 e2t + c2 te2t
(iii) y = C1 cos(5.1t) + C2 sin(5.1t)
(iv) y = e1.2t [C1 cos(5t) + C2 sin(5t)]
(v) y = e0.4t [C1 cos(2t) + C2 sin(2t)] 1.3 cos(7t)
(vi) y = C1 cos(3t) + C2 sin(3t) + 0.13 cos(8t) 1.46 sin(8t)
(vii) y = C1 cos(3t) + C2 sin(3t) + 0.13t cos(3t) 1.46t sin(3t)
(viii) y = A sin(0.1t) sin(6.1t)
(a) Identify the transient and steady-state parts of each solution. (Some may not have both.)
(b) Which solutions are for differential equations of the form ay + by + cy = 0?
(c) Which solutions are for undamped systems? Which are for under-damped systems?
Critically damped? Over-damped?
125
3. Below are some graphs of solutions to ODEs of the form ay + by + cy = f (t), where either,
or both, of b or f (t) may be zero.
(i)
(ii)
(iii)
(iv)
(v)
(vi)
(vii)
(a) Which graphs are for solutions to undamped systems? Under-damped systems? Critically or over-damped systems? (You should not be able to tell the graphs for critically
damped or over-damped apart.)
(b) Which graphs are for ODEs of the form ay + by + cy = 0?
4. None of the ODEs in Exercise 1 have a solution equation given in Exercise 2, or solution
graph given in Exercise 3. However, we CAN match up the FORMS of the ODEs, solution
equations, and graphs of solutions. For example, equation (iii) from Exercise 1 matches with
solution (ii) from Exercise 2 and graph (i) from Exercise 3. Find eight other sets of three like
this; one graph will have to be used more than once.
126
127
All of the applications that we have studied so far have involved some quantity that is a function
of time; that is, time has been the independent variable. Arbitrary constants have arisen in the
process of solving the associated ODEs, and we have used given initial conditions to determine the
values of the constants. In this chapter we look at deflection (bending) of horizontal beams and
vertical columns. For horizontal beams the deflection is a function of the distance along the beam
or column. The independent variable is then a one dimensional position variable, as discussed in
Section 1.4. As seen in Section 1.7, we use boundary conditions, rather than initial conditions, to
determine the values of the arbitrary constants.
We will designate the variable x to denote the distance along the beam (or column) from one
end or the other. Due to the weight of the beam there will be some deflection y off of the horizontal
line the beam would follow if it had no weight. The deflection will be different at different points
along the beam, so y = y(x). That is, the amount of deflection depends on where one is looking
along the length of the beam.
The situation will be significantly different for vertical columns, in a way that might be somewhat surprising. Well see that such a column will remain straight as more and more weight is
added to it until, at some weight (called the first critical load), it suddenly deflects (buckles).
It will then either deform or break as a the load is increased. However, if we prevent the middle of
the column from deflecting, each half will deflect at a load (called the second critical load) that
is four times the first critical load.
128
5.1
In this section we will take a look at the differential equations associated with beams that are
suspended horizontally in some way. The beams themselves will not be horizontal over their entire
lengths, because the force of gravity will cause some bending. The first thing to understand is the
mathematical setup. Suppose that we have a beam of length 10 feet. We put the cross-sectional
center of its left end at the origin of an x-y coordinate plane, and the cross-sectional center of its
right end at the point (10, 0). The axis of symmetry of the beam then runs along the x-axis from
x = 0 to x = 10; see the figure below and to the left.
y
y
x
10
10
Now the beam will deflect (a fancy term for sag) in some way, due to any weight it is supporting,
including its own weight. The shape it takes will depend on the manner in which it is supported
(we will get into that soon), but one possibility is shown in the figure above and to the right.
The original axis of symmetry of the beam now follows the graph of a function, which we will call
y(x). Note that the domain of the function is just the interval [0, 10]. Our goal will be to find the
mathematical equation of the function.
y
Let us still consider a 10 foot beam, but we will represent it
with just the curve described by the axis of symmetry. (From
now on, when we talk about the beam, we really mean the axis
x
of symmetry of the beam.) Suppose also that the ends of the
10
beam are what we call embedded. This means that they
are not only supported at both ends, but the ends are also
held horizontal by being clamped somehow. A good image
to keep in mind is a beam that is stuck into two opposing
walls of a structure. See the diagram to the right.
The theory behind obtaining a differential equation to model a horizontal beam is beyond the
scope of this class. Suffice it to say that it involves ideas from the area of statics, like the bending
moment of the beam, and things involving the properties of the material from which the beam is
built. The differential equation itself is fourth order:
EI
d4 y
= w(x)
dx4
129
(1)
Here E is Youngs modulus of elasticity for the material from which the beam is made, I is the
moment of inertia of a cross-section of the beam and w(x) is the load per unit length of the beam.
If the beam has uniform cross-section and the only weight that it is supporting is its own weight,
then w(x) is a constant. We will consider only that condition.
Because the ODE is fourth order, we will need four boundary conditions to determine all of the
constants that will arise in solving it. We first recognize that because the two ends are supported,
there will be no deflection at either end. Therefore y(0) = y(10) = 0. This will be the case for
any horizontal beam that is supported at both ends. Next we consider the fact that the ends of
the beam are embedded horizontally into a wall. The embedding causes both ends to be horizontal
right at the points where they leave the walls they are embedded in, so the slope of the beam is
zero at those points. Mathematically we express this by y (0) = y (10) = 0. When we put the
ODE together with these boundary conditions we get a boundary value problem. Suppose that for
our ten foot beam beam E = 10, I = 5 and w(x) = 100. The boundary value problem that we
have is then
d4 y
y(0) = 0, y (0) = 0,
y(10) = 0, y (10) = 0
50 4 = 100,
dx
Example 5.1(a): Solve the boundary value problem just given.
d4 y
= 2. Our task now is to keep integrating
dx4
d3 y
= 2x + C1 , and integrating
both sides until we find y = y(x). Integrating once gives
dx3
2
d y
dy
again gives
= 13 x3 + 21 C1 x2 + C2 x + C3 , and
= x2 + C1 x + C2 . Next we find that
2
dx
dx
applying the initial condition y (0) = 0 gives C3 = 0. Substituting this value and integrating
1 4
x + 61 C1 x3 + 21 C2 x2 + C4 , and applying the boundary condition
one more time we get y = 12
y(0) = 0 results in
1 4
y = 12
x + 61 C1 x3 + 21 C2 x2 .
(2)
We begin by dividing both sides by 50 to get
1000
100
We now apply the initial condition y(10) = 0 to get 0 = 1000
12 + 6 C1 + 2 C2 , and the initial
1000
100
condition y (10) = 0 to get 0 = 3 + 2 C1 + 10C2 . To solve this system we multiply the
first equation by 12 and the second by 6 to get the system to the left below, which can be
solved in the manner shown in the other steps:
30C1 6C2 =
200
10C1 = 100
C1 = 10
Use your calculator to graph the solution from x = 0 to x = 10, using a y scale that allows
you to actually see the deflection of the beam. Does the result surprise you? (It should!) One
annoying feature of the differential equation is that it is based on taking down to be the positive
direction. To see what the actual shape of the beam will be, multiply your solution by 1, then
graph it. Now the result should look something like the picture near the bottom of page 128.
130
(b)
12 ft
8 ft
131
(c)
(d)
x=0
20 ft
x = 15
2. Suppose that you have a beam that is 8 feet long. For each of the scenarios given, determine
first whether it would make any sense physically to have the beam supported in the manner
given. If not, explain why. If it does make sense, give the four boundary conditions.
(a) Left end embedded horizontally, right end simply supported.
(b) Left end simply supported, right end free.
(c) Both ends free.
3. Find the deflection function y = y(x) for an eight foot beam that is embedded at both ends,
carrying a constant load of w(x) = 150 pounds per foot. Suppose also that E = 30 and
I = 80, in the appropriate units.
(a) Give the appropriate boundary value problem (differential equation plus boundary conditions).
(b) Solve the differential equation and apply the boundary conditions in order to determine
the constants. What is the final solution?
(c) Graph your solution on the appropriate x interval, using a y scale that allows you
to actually see the deflection of the beam. Remember to multiply the right side of your
solution from (b) by negative one so that it appears the same way that the beam will.
(d) Where do you believe the maximum deflection should occur? Find the deflection there you need not give units with your answer, since I have been somewhat vague about the
units of the constants E and I.
4. (a) Sketch a graph of the deflection of a beam that is embedded at its left end and free at
its right end.
(b) Suppose that the beam is 10 feet long, with values of w0 , E and I of 100, 10 and 5,
respectively. Solve the boundary value problem.
(c) Graph your solution and compare with your sketch in part (a). Of course they should
be the same.
(d) What is the maximum deflection of the beam, and where does it occur?
5. Repeat parts (a)-(d) of Exercise 4 for a 10 foot beam that is simply supported on both ends.
132
6. Repeat steps (a)-(c) of Exercise 4 for an eight foot beam, with the same parameters as in
Exercise 3, that is embedded on the left end and simply supported on the right end. Then
do the following:
(d) It was intuitively clear where the maximum deflection occurred for the two previous
situations, but it is not so clear in this case. Take a guess as to about where you think it
should occur for this case. Then use the graph on your calculator, along with the trace
function, to determine where the maximum deflection occurs, and how much it is.
7. The graphs below are those of some fourth degree polynomials. The points labeled A, D, E,
H, I and L are maxima for their respective functions, and the points labeled B, C, F, G, J
and K are inflection points. For each of the following boundary situations, give the endpoints
of a section of graph that has the shape the deflection curve would take. Assume that both
ends of the beam are supported at the same level, and that the dashed lines are horizontal.
(a) Simply supported at the left end, embedded at the right end.
(b) Embedded at both ends.
(c) Simply supported at both ends.
D
B
H
F
133
L
J
5.2
Load
In this section we will examine the behavior of a vertical
Force
column when a load is applied directly downward on its
top, as shown in the first picture to the right. We will begin by considering columns that are pinned; this means
we will allow the top and bottom of the column to be at
angles other than perpendicular to the ground. We will
require the top and bottom of the column to be vertically
aligned with each other - later we will consider a situation
where we will relax this condition. So, for example, when
enough force is applied downward the column will deflect
horizontally as shown in the second picture to the right.
We will set up a coordinate system as shown below
and to the right, with x indicating the distance upward
from the bottom of the beam and y = y(x) representing
the horizontal deflection of the column at any point x.
The differential equation governing the deflection of the
column is
d2 y
EI 2 = P y,
dx
where E and I are as in the last section; they are properties
which could vary along the column (with the variable x), but this
would be unusual. Well only consider columns where they do not
change. P is the (positive) force exerted downward on the top of
the column, and we will look at the effects of different values of P ,
but for purposes of solving the ODE it is a constant. Since the ODE
is second order we will need two conditions to determine the solution.
If the length of the column is denoted by L, we have the boundary
conditions y(0) = 0 and y(L) = 0. If we rearrange the equation
and combine it with the boundary conditions we get the boundary
y (pos)
value problem (BVP)
P
d2 y
y = 0,
+
2
dx
EI
x
x=L
y(L) = 0
y(0) = 0
y(0) = 0, y(L) = 0.
We will see that solving this boundary value problem is somewhat different than solving the kind
of BVPs we saw for horizontal beams, in the previous section. Lets go through that now, with
specific values of E, I and L. The methods and ideas in this section are significantly different
134
than the other things we have done in this course - youll need to put your thinking cap on for this
one!
Example 5.2(a): Solve the boundary value problem
d2 y
P
+
y = 0,
2
dx
EI
y(0) = 0, y(20) = 0,
(1)
(2)
Using our methods from Chapter 3, this equation has auxiliary equation r 2 +
q
P
because P > 0, its roots are r = i 120,000
. The solution to (2) is then
y = C1 sin
P
120,000
x + C2 cos
P
120,000
P
120,000
= 0 and,
Applying
q the boundary condition y(0) = 0 gives us C2 = 0, so the solution is then y =
P
x. (Ive omitted the subscript for simplicity.) Now here is where things start to
C sin 120,000
get interesting! The other boundary condition tells us that
q
P
(20)
(3)
0 = C sin 120,000
q
P
(20) = 0. The first possibility gives
which, in turn, tells us that either C = 0 or sin 120,000
us the trivial solution y = 0 - it satisfies the differential equations and boundary conditions,
but it isnt particularly interesting! Considering the second possibility, that sin = 0 for =
0, , 2, 3, ..., n, ... so, for C 6= 0, (3) will be true for
q
P
(4)
120,000 (20) = 0, , 2, 3, . . . , n, . . . ,
the first of which also gives us the trivial solution y = 0. Therefore, in theory at least (well talk
later bout what all this means from a practical point of view), we can only have nonzero solutions
to the BVP if
q
2 3
n
P
120,000 = 20 , 20 , 20 , . . . , 20 , . . .
This gives us the nonzero solutions
3
n
y = C sin 20
x, y = C sin 2
20 x, y = C sin 20 x, . . . , y = C sin 20 x, . . .
This isnt really the end of the story, but we need to pause to catch our breath and develop some
terminology before resuming. At this point what we know about the boundary value problem (1)
is that y = 0 is a solution, called the trivial solution (because it is mathematically uninteresting),
and we only have nontrivial solutions for discrete values of P ; those solutions are the ones the
example concluded with. The non-trivial solutions are called buckling modes (for reasons youll
soon see). The first non-trivial solution is called the first buckling mode, the second is the
135
second buckling mode, and so on. The only values of P for which we can have nontrivial
solutions are those that satisfy (4). Solving for P gives us
2
2
2
n 2
2
3
P = 120000
, 120000
, ...
, 120000
, . . . , 120000
20
20
20
20
= 300 2 , 300(2)2 , 300(3)2 , . . . , 300(n)2 , . . .
= 300 2 , 4(300 2 ), 9(300 2 ), . . . , n2 (300 2 ), . . .
These values of P are called critical loads. Like the buckling modes, they are numbered, so
300 2 is the first critical load, 4(300 2 ) is the second critical load, etc. The word load
refers to the load held up by the column. Note that the second critical load is four (two squared)
times the first critical load, the third critical load is nine (three squared) times the first critical
load, and so on.
What happens physically is this: When there is no load on the column it is perfectly straight
(the solution y = 0), and it remains that way until the first critical load is reached. At that point
the column will deflect sideways, taking the shape of the curve y = C sin 20
x, shown at the left
below. In reality, as the load increases beyond the first critical load, the deflection will remain the
same shape, but with increasing amplitude, until the column fails.
20
20
20
2
3 (20)
10
1
3 (20)
If we were able to prevent the middle point of the column, at x = 10, from deflecting,
the column would be able to support the second critical load. Because the column is held with
y(10) = 0, the deflection of the column will take the shape of a full period of the sine function,
as shown in the middle picture above; this is the second buckling mode. The act of preventing
deflection is sometimes called pinning. If we pin the column at points one-third and two-thirds
of the way along its length, the column would be able to support the third critical load, and the
shape of the deflection would be given by the third buckling mode, shown to the right above.
Lets revisit the boundary conditions y(0) = y(L) = 0, where L is the length of the column.
We should first note that the only requirement we really had was that the top and bottom of the
column were vertically aligned. We could just as well have put y(0) = y(L) = 2, as shown in the
diagram to the left at the top of the next page; however, the mathematics involved are a bit simpler
if we instead use y(0) = y(L) = 2, as we did. Physically we must still insist that the top and
bottom be aligned. Without this restriction, any horizontal shifting of the ceiling would result in
hinging and a collapse. The beginning of such a collapse is indicated by the three diagrams to the
right at the top of the next page.
136
x
L
Load
Force
y (pos)
y(0) = y(L) = 2
Now suppose the top and bottom were embedded, rather than being pinned. If the column had
length L, we would then have the boundary conditions
y (0) = y (L) = 0.
(5)
This gives us two boundary conditions, which is what we should need in order to solve the second
order ODE
P
d2 y
y = 0,
+
2
dx
EI
and leaves us without any conditions on y(0) and y(L). In this situation, we could conceivably
allow the ceiling to drift laterally without collapse, because the column being held in a vertical
alignment at the bottom and top would provide enough rigidity to prevent collapse. The first three
buckling modes for this situation are shown below; if the ceiling were prevented from drifting, the
second mode would then become the first, the fourth would become the second, and so on. You
will investigate this situation, along with the corresponding critical loads, in the exercises.
137
138
5.3
P
d2 y
+
y = 0, which can be rewritten as
2
dx
EI
!2
r
d2 y
P
y = 0.
+
dx2
EI
q
P
Just to make things simpler we will replace EI
with the symbol (lambda). Since P, E and
I are all positive quantities, is a positive real number also. This will be important soon.
So we wish to solve the differential equation y + 2 y = 0, which we have done several times
before (with specific numbers in place of ). We can rewrite this as
d2
(y) = 2 y
dx2
(1)
d2
(y) is used to indicate that the second derivative
Remember that y is a function. The notation
dt2
with respect to x is an operator, which means it is a function that acts on functions, and the
result of that action is another function. (You may wish to revisit Section 3.2 for a review of
operators.) Equation (1) says that we are looking for a function for which the action of the second
d2
really just amounts to multiplying the function by the constant 2 .
derivative operator
dx2
This sort of situation happens often enough, and is important enough, that mathematicians have
developed some language relating to it. We will introduce that language, as well as some that leads
up to it, in this section. Lets begin with a review of operators, and we will return to our discussion
of equation (1) later.
Operators
A function can be thought of as a mathematical machine that takes in a number and, in
return, gives out a number. There are other mathematical machines that take in things other than
numbers, like functions or vectors, usually giving out things like what they take in. As stated
above, these sorts of machines are often referred to as operators.
d2
is an operator. You should
Example 5.3(a): As just mentioned, the second derivative
dx2
be quite familiar with its action:
d2
(5x3 + 7x2 2x + 4) = 30x + 14
dx2
139
Example 5.3(b): Of course the first derivative is an operator as well, and we can create new
operators by forming something called a linear combination of derivatives. As an example,
we can define an operator
d2
d
D =3 2 +5 4
dt
dt
by its action on a function y = y(t):
D(y) = 3
dy
d2 y
+ 5 4y.
2
dt
dt
d2 2t
d
(e ) + 5 (e2t ) 4e2t = 12e2t 10e2t 4e2t = 2e2t
dt2
dt
You saw such operators when we studied second order linear ODEs.
Example 5.3(c): You may know that when we multiply the matrix A =
1
we get
the vector u =
3
Au =
4 6
3
5
1
3
(4)(1) + (6)(3)
(3)(1) + (5)(3)
22
18
4 6
3
5
times
2
,
1
Av =
4 6
3
5
2
1
2
1
We can think of A as an operator that acts on vectors with two components to create other
vectors with two components.
You may recall that the derivative operator is what we call a linear operator. What this
means is that if f and g are functions, and c is a constant, then
df
dg
d
[f (x) + g(x)] =
(x) +
(x)
dx
dx
dx
and
d
df
[cf (x)] = c (x)
dx
dx
This behavior is not unique. If A is a matrix, u and v are vectors, and c is a scalar (constant),
A(u + v) = Au + Av
and
A(cu) = cAu
Linear operators have these two properties, of distributing over addition and passing through
constants. (This is where the linear in linear algebra comes from.) Many operators used in
applications are linear operators.
140
d2
, and note that
dx2
d2
(sin 5x) = 25 sin 5x.
dx2
The effect of the derivative on sin 5x is to simply multiply the function by 25.
Example 5.3(e): Note that in Example 5.3(c), the result of A times v was simply
1 times v.
Example 5.3(f ): Based on Example 7.4(d), sin 5x is an eigenfunction for the operator
d2
, with corresponding eigenvalue 25.
dx2
2
1
141
Example 5.3(h): Because the derivative of a constant is zero, which is also zero times
d
,
the function, every constant function is an eigenfunction of the first derivative operator
dx
with corresponding eigenvalue zero. This emphasizes that even though the zero function isnt
allowed as an eigenfunction, eigenfunctions are allowed to have eigenvalues of zero.
(2)
showing that an operator can have more than one eigenfunction (beyond just constant multiples)
with the same eigenvalue. This also holds for matrices and eigenvectors.
Finally, we return to the equation
d2
(y) = 2 y.
dx2
(1)
2 is just a constant, so the y that we are looking for is an eigenfunction of the equation (1).
Given what we saw in Examples 5.3(d) and (f) and equation (2) above, we might conjecture that
y = sin x and y = cos x are both eigenfunctions of (1). Indeed, they are, and they are the
only eigenfunctions (excluding constant multiples of those two). The corresponding eigenvalue is
2 in both cases, which shows us that different eigenfunctions (or eigenvectors) can have the
same eigenvalue.
Section 5.3 Exercises
1. In this exercise you will be considering the first derivative operator
d
.
dx
(a) The function y = e3x is an eigenfunction for the operator. What is the corresponding
eigenvalue?
(b) Give the eigenfunction of the operator that has eigenvalue 5.
(c) Based on your answers to parts (a) and (b), what is the general form of any eigenfunction
of the first derivative operator and what is the corresponding general eigenvalue?
d2
2. Now consider the second derivative operator
. There are three general forms of the
dx2
eigenfunctions for this operator, depending on whether the eigenvalues are positive, negative
or zero.
(a) Give a specific function that has eigenvalue zero; that is, the second derivative of the
function is zero.
(b) Give the most general form of function whose eigenvalue is zero.
142
(c) Weve already seen and worked with functions that have negative eigenvalues. Give one
general form of functions that have negative eigenvalues for this operator, and give the
corresponding eigenvalue for the general form. Then give the other general form and its
corresponding eigenvalue.
(d) Give the eigenfunctions that will have eigenvalue nine; there are two of them!
(e) Give the general form of the eigenfunctions of the operator that have positive eigenvalues,
and give the general eigenvalue.
3. Let D be the operator D =
by Dy =
d2 y
dy
+ 5 + 3y.
2
dt
dt
d
d2
+ 5 + 3, whose action on a function y = y(t) is defined
2
dt
dt
(a) Show that y = e2t is an eigenfunction for this operator, and determine the corresponding eigenvalue.
(b) In general, any function of the form ekt is an eigenfunction for D. Determine the
general eigenvalue.
143
5.4
In this section, we will consider the physically impossible but mathematically convenient situation:
We have a metal rod of length L (see picture below) that is perfectly insulated along its length,
so that no heat can enter or escape along its length, but
for which heat can enter or leave the ends. At any time
t greater than zero and any position x along the rod,
x=0
x=L
the function u(x, t) gives the temperature at that point
x and time t. (We will think of the rod as beinginfinitely thin, so that the rod has only one
point at each x position. If you are not happy with this, an alternative is to think that if the
rod had some thickness the temperature at every point in a cross-sectional slice at some x is the
same, so we need not consider the other two space dimensions.) Suppose that at time zero there is
some distribution of temperatures along the rod, given by a the function f (x) for 0 x L, and
suppose also that the ends of the rod are held at temperature zero at all times t 0. The function
f gives an initial condition for each point x along the length of the rod, and the conditions that
the ends are held at temperature zero are boundary conditions.
What we would like to know is whether, and how, we can determine the temperature at any
point x with 0 < x < L (we know the temperatures at x = 0 and x = L are always zero), at
any time t > 0. Here the dependent variable u depends on the two independent variables x and
t. Some physical principles concerning heat give us a differential equation for this situation and,
due to there being two independent variables, it is a partial differential equation. The equation
(called the heat equation), and the conditions given above can all be stated as
2u
u
= k 2,
t
x
u(0, t) = u(L, t) = 0,
u(x, 0) = f (x)
(1)
Here the conditions u(0, t) = u(L, t) = 0 are boundary conditions and u(x, 0) = f (x) is
essentially an initial condition (for every point along the rod). Thus we have a problem that is
a sort of combination initial value/boundary value problem. But we can really think of it as a
boundary value problem for this reason: If we were to think of the
t
Cartesian plane as representing position x along the horizontal axis
and time t along the vertical axis we get a picture like the one to the
right, where each point in the shaded region represents a point x in
the rod and some time t. Our goal is then to find the temperature
at each of those points; in this way we can think of trying to find
function values in a region that is bounded by the line from zero
to L on the x-axis and the two half-infinite lines from zero to
infinity in the t direction at x = 0 and x = L. The condition
u(x, 0) = f (x) can be thought of as a boundary condition along
L x
the bottom, and the conditions u(0, t) = u(L, t) = 0 are boundary
conditions along the two sides.
f
we simply
Recall that if we have the function f (x, y) = x2 y 3 , to get the partial derivative
x
2
3
3
take the derivative of x y , treating y (and therefore y ) as a constant. Similarly, we get
f
by treating x as a constant, so we have
y
f
= 2xy 3
x
and
144
f
= 3x2 y 2
y
u
u
and
for u(x, t) = e2t sin 3x.
x
t
u
we consider t to be a constant, so e2t is as well. The derivative is then
x
u
= 3e2t cos 3x.
x
When finding
u
, sin 3t is essentially a constant, so
t
u
= 2e2t sin 3x.
t
u
2u
= k 2.
t
x
We already have
u
= 3e2t cos 3x
x
and
u
= 2e2t sin 3x.
t
from the previous example. Taking the partial derivative of the first of these with respect to
x again gives us
2u
= 9e2t sin 3x
x2
so
u
9
2u
= 2e2t sin 3x =
k = 92
9e2t sin 3x = k 2 ,
t
2
x
We will soon see that u(x, t) = e2t sin 3x is not the most general solution to the equation.
Those whove had a multivariable calculus course will perhaps recall that the computation of
partial derivatives can be significantly more complicated (and therefore difficult) than the ones
done above, but if we understand the two examples just given we are ready to understand how (1)
is solved. The method for doing it is called separation of variables, which is similar in execution,
up to a point, to the method of the same name we used to solve separable first order ODEs up to
a point. After that point we must proceed differently.
To begin, we assume that the function u(x, t) is actually a product of a function of x alone
and another function of t alone. There is no practical reason to think this might be the case, but
the method works, so well use it! (This method was invented/discovered in the 1700s by Daniel
Bernoulli, part of a family of a number of accomplished mathematicians and scientists. Daniel was
also involved in the derivation of the ODE we used for horizontal beams.) If we let g be the
function of x and h be the function of t, then u(x, t) = g(x)h(t). Now remember that if we
are taking the derivative of g(x)h(t) with respect to x, h(t) is treated as a constant, and when
taking the derivative with respect to t, g(x) is treated as a constant, so
u
= g(x)h (t)
t
and
145
2u
= g (x)h(t).
x2
The differential equation in (1) then becomes g(x)h (t) = kg (x)h(t). If we divide both sides by
kg(x)h(t) we get
h (t)
g (x)
=
.
(2)
kh(t)
g(x)
Here is where we deviate from the procedure for solving first order separable equations. (2) needs
to be true for all values of x and t, and this is likely only the case if both sides of (2) are equal
to some constant (again, well see that it works!) that we will call 2 . For reasons we wont go
into here, is positive. Setting each side equal to 2 and multiplying by the denominators we
get
d2
d
h(t) = k2 h(t)
and
g(x) = 2 g(x).
(3)
dt
dx2
In addition to this, we also have the boundary conditions g(0) = g(L) = 0 for the second equation.
The first equation in (3) tells us that h(t) is an eigenfunction for the first derivative operator, with
2
eigenvalue k2 , and we know that h(t) is then any constant multiple of ek t . The second
equation says that g(x) is an eigenfunction of the second derivative operator with eigenvalue
2 . Because is positive, the eigenfunctions are constant multiples of sin x and cos x, as
determined in the previous section. The general solution to the heat equation then looks like
2
(4)
Lets focus a bit more on the second ODE in (3) and its boundary values g(0) = G(L) = 0. The
general solution to the ODE is g(x) = A sin x + B cos x. Applying the condition g(0) = 0 gives
us B = 0, so the solution is g(x) = A sin t. (At this point this story should be starting to
feel familiar!) We now consider the boundary condition g(L) = 0, which gives us 0 = A sin L.
As before, when considering vertical columns, we dont want to let A = 0, so we must have
sin L = 0. This implies that
L = 0, , 2, 3, ...
= 0,
2 3
,
,
, ...
L L L
and the solutions to the boundary value problem (disregarding constants and the zero solution
arising from = 0) are
2
3
x, sin
, ...
sin x, sin
L
L
Lx
k 2
x, e
u(x, t) = e L2 t sin L
4k 2
t
L2
4k 2
t
L2
, e
9k 2
t
L2
sin 2
L x, e
9k 2
t
L2
sin 3
L x, ...
(5)
Recall that when solving an ODE like y + 3y + 2y = 0 we assumed y = ert for some constant
r. From this we obtain y = et or y = e2t , but we saw that the sum of constant multiples
of these two, y = C1 et + C2 e2t is the most general solution. By the same reasoning, the most
general solution to the PDE were looking at is an infinite sum of the solutions in (5):
k 2
4k 2
t
L2
sin 2
L x+A3 e
9k 2
t
L2
sin 3
L x+ +An e
n2 k 2
t
L2
sin n
L x+ (6)
This story goes on quite a bit longer, but lets end it with the following. In order to try to meet
the condition u(x, 0) = f (x) we must have
3
n
f (x) = A1 sin L x + A2 sin 2
L x + A3 sin L x + An sin L x +
The right hand side of (7) is something called a Fourier series. This brings up the question
146
(7)
In what way (or ways) do we interpret the equal sign in (6), and for what
functions f can such interpretation(s) be made?
Attempts to answer this question gave birth to a large amount of mathematics over many years,
starting with Joseph Fouriers work in the early 1800s, and with a major result proved as late as
1966. Perhaps some of you will investigate this subject more in later coursework.
147
5.5
Chapter 5 Summary
Boundary value problems arise when the independent variable of an ODE is length. Applications include the deflection of horizontal beams and vertical columns along their lengths.
The differential equation for a horizontal beam is fourth order, and the solution is a fourth
order polynomial with four arbitrary constants.
There are two boundary conditions at each of the two ends of a horizontal beam, giving four
conditions used to determine the values of the constants.
There are three possible end conditions for each end of a horizontal beam:
Embedded: This is when the end of the beam is clamped horizontally. The mathematical conditions for such an end are y = 0 and y = 0.
Simply Supported (Pinned): This is when the beam is held up but allowed to pivot.
Mathematically, y = 0 and y = 0.
Free: This is when an end is completely unsupported, and the other end must be
embedded. The mathematical conditions for such an end are y = 0 and y = 0.
The ODE for a vertical column is second order, and the solution is either a sine function or
a cosine function, depending on the end conditions:
When the ends are pinned (hinged) the solution is a sine function.
When the ends are embedded the solution is a cosine function.
Mathematically, there are infinitely many solutions for a vertical column that is pinned at its
ends.
Each is some multiple of a half period of a sine function beginning at x = 0.
The first solution, called the first buckling mode, is a single half-period of the sine
function. This occurs physically when the column is allowed to deflect over its entire
length.
Each additional solution (buckling mode) consists of n2 periods of the sine function for
n = 2, 3, 4, 5, .... Physically, the solution consisting of n2 periods of the sine function
occurs when the the column is pinned along its length at n 1 equally spaced points.
Mathematically, there are infinitely many solutions for a vertical column that is embedded at
its ends.
Each is some multiple of a half period of a cosine function beginning at x = 0.
In the case that the ends of the column are embedded, it is physically possible that the
ceiling can float (move laterally).
If the ceiling is allowed to float the first buckling mode is a single half-period of the
cosine function. Each additional buckling mode consists of n2 periods of the cosine
function for n = 2, 3, 4, 5, ....
If the ceiling is NOT allowed to float the first buckling mode is a single period of the
cosine function. Each additional buckling mode consists of n periods of the cosine
function for n = 2, 3, 4, 5, ....
148
The load that causes the first buckling mode is called the first buckling load, and the nth
buckling load leads to the nth buckling load.
The nth buckling load is n2 times the first buckling load.
Let A be an operator that operates on functions and y a nonzero function. If there is a
constant such that
Ay = y
then y is an eigenfunction of the operator A, with corresponding eigenvalue .
149
Formula Sheet
xn =
(xa )b = xab ,
loga
u
w
1
,
xn
= loga u loga w
eln u = u,
xn =
log(uc ) = c log u
ln eu = u
Trigonometric Functions:
tan u =
sin u
cos u
cot u =
cos u
sin u
sec u =
1
cos u
csc u =
1
sin u
Trigonometric Identities:
sin2 + cos2 = 1
sin 2 = 2 sin cos
tan2 + 1 = sec2
A2 + B 2 and = tan1
Eulers Relations:
ei = cos + i sin
1 + cot2 = csc2
or
B
B
if A > 0, = + tan1
if A < 0
A
A
ei = cos i sin
d[uv]
dv
du
=u
+v
dx
dx
dx
150
[cos(u)] = [ sin(u)](u) ,
(eat ) = aeat
[ln(u)] =
1
(u)
u
A Few Integration Formulas: All formulas should include an arbitrary constant, which I have
left off here to keep things a little cleaner.
Z
Z
Z
Z
dx = x
cf (x) dx = c f (x) dx
k dx = kx
Z
Z
eat (at 1)
t e dt =
a2
at
Z
Z
g(x) dx
du =
Z
e du = e
t2 eat dt =
sin u du = cos u
Z
cos u du = sin u
1
du = ln |u|
u
eat dt =
1
sin at dt = cos at
a
cos at dt =
1
sin at
a
eat cos bt dt =
eat
(a cos bt + b sin bt)
a 2 + b2
dy
+ p(x)y = q(x) Using an Integrating Factor:
dx
Compute u =
d[eu y]
. Multiply both sides by dx and integrate.
dx
151
1 at
e
a
eat
e sin bt dt = 2
(a sin bt b cos bt)
a + b2
at
Solving
1
un+1 as long as n 6= 1
u du =
n+1
1
1
dx = ln | ax + b |
ax + b
a
f (x) dx
[f (x) g(x)] dx =
B
B.1
Outcomes:
Apply the rules of differentiation, along with provided formulas, to find
the derivatives of functions.
The purpose of this appendix is to remind you of how to find the derivatives of some functions
of the sorts that we will encounter as we go on. For this we will usually take the independent
variable to be x and the dependent variable y; the statement y = f (x) indicates that y is
dy
for the derivative. Lets begin by
a function of x; recall also the notations y = f (x) =
dx
giving the derivatives of some common functions, and some basic rules of taking derivatives. The
notation ( ) means the derivative of the quantity in the parentheses, and k represents an
arbitrary constant.
Derivatives of Some Functions
(k) = 0
(x) = 1
(sin x) = cos x
(xn ) = nxn1
(cos x) = sin x
(ex ) = ex
(ln x) =
1
x
For the following, k again represents an arbitrary constant, and f (x), g(x), u = u(x) and
v = v(x) are functions of the independent variable x.
Derivative Rules
[kf (x)] = kf (x)
(uv) = uv + vu
[u v] = u v
u
vu uv
v2
Youll recall the third and fourth items above as the product rule and quotient rule, and the
last item is the chain rule. The first two rules above tell us that the derivative is something called
a linear operator. This is the same idea as a linear transformation, for those of you who have
had linear algebra. Now well see how to use some of the rules and derivatives of common functions
to find derivatives of some combinations of those functions.
152
Derivatives of Polynomials
Youll recall that to find the derivative of a polynomial like f (x) = 5x3 34 x2 7x + 2 we
simply use the various basic derivatives and rules as follows: Multiply the coefficient of each term
by the corresponding power of x, and decrease the power by one, remember that the derivative
of a constant is zero.
Example B.1(a): Find the derivative of f (x) = 5x3 43 x2 7x + 2.
3
f (x) = 5x3 x2 7x + 2
4
3
f (x) = 15x2 x 7
2
Second Derivatives
For certain applications we will need to take the derivative of the derivative of a function. Youll
recall that such a derivative is called the second derivative (so the kind of derivative you did
above is called a first derivative). The notation is this: If the first derivative is y or f (x), then
dy
the second derivative is y or f (x). If the first derivative is
, then the second derivative is
dx
2
d y
. Note the different placement of the exponents in the numerator and denominator of this
dx2
expression - there is a reason for this, but its not too exciting, so I wont go into that here.
Other Applications of the Power Rule
5
= 5x3
x3
5
x3
dy
15
= 15x4 = 4
dx
x
x
4
1 2
1
x 3 =
3
12
12 x2
Note that the four in the denominator is a constant, but it is really 14 , not 4.
f (x) =
x
1 1
= x3
4
4
f (x) =
In the notation of the rule in the box on the previous page, g(x) = x5 + 2x 4 and f (u) = u8 .
Example B.1(e): Find the derivative of f (t) = 5.6 sin(4.2t 1.3).
f (t) = 5.6[cos(4.2t 1.3)] (4.2t 1.3) = 5.6[cos(4.2t 1.3)] (4.2) = 23.6 cos(4.2t 1.3)
In this last example there is the additional constant 5.6 that is just multiplied by the result of
the derivative of the rest of the function. Note that the result of the derivative of 4.2t 1.3 is
brought to the front and multiplied by the original constant at the end of the process. More on
this in a bit.
In light of what you have seen, we can revise some of our basic derivatives some. Again, u is
a function of x: u = u(x).
du
d u
(e ) = eu
dx
dx
d
du
(cos u) = sin u
dx
dx
d
du
(sin u) = cos u
dx
dx
and
154
So how do we find the derivative of a product of two functions? Well, we use the product rule,
which says that if u = u(x) and v = v(x) are two functions of x, then
(uv) = uv + vu
or
d(uv)
dv
du
=u
+v
dx
dx
dx
v u u v
v2
e2x
x5
3
x2
(b) g(t) =
6
t2
2
6
t
(c) y =
1
5 x
4. Find the second derivatives of the functions from 3(a) and (b).
155
4
(d) h(x) =
3
x
5. Find the derivative of each of the following, utilizing the suggestions provided.
(a) g(x) = 16 x2 - If you write the square root as an exponent you will have something
very much like the first example above.
t
(b) y = 3 cos
2
(c) A(t) = 500e0.3t - This is a two-step process (again, ignoring the constant of 500),
with the first step being 0.3t and the second step being the exponential. Remember
that the derivative of ex is ex .
6. Find the derivative of each of the following.
(a) y = 5e2x
(b) x = 4 sin(3t)
(e) y = ex
(c) g(x) =
(x2
2
4x)7
7. For each of the following, put the product in the correct order.
3
8. For each of the following, multiply and put the final product in the correct order.
(a) 4 sin(5t + 3) (2e2t )
(c) (3 ln x)(4x3 )
9. Use the product rule to find the derivative of each of the following.
(a) f (t) = t2 e7t
(b) y = 3x cos 2x
3 sin 2t
e7t
(b) y =
4e5t
t2
11. Use the quotient rule and the fact that tan x =
156
(c) g(x) =
cos 6x
2x3
sin x
to determine the derivative of tan x.
cos x
B.2
In this course we will often need to solve systems of two equations in two unknowns, and it is
important to be able to do this quickly and correctly. There are two methods, the addition
method and the substitution method. Each has its own advantages and disadvantages; both
methods will be demonstrated here.
The Addition Method
Well begin with the addition method, going from the easiest scenario to the most difficult
(which still isnt too hard).
Example B.2(a): Solve the system
3x y = 5
.
2x + y = 15
The basic idea of the addition method is to add the two equations together so that one of the
unknowns goes away. In this case, as shown below and to the left, nothing fancy need be done.
The remaining unknown is then solved for and placed back into either equation to find the other
unknown as shown below and to the right.
3x y = 5
3(4) y = 5
2x + y = 15
12 y = 5
5x = 20
12 = y + 5
x = 4
7 = y
What made this work so smoothly is the y in the first equation and the +y in the second;
when we add the two equations, the sum of these is zero and y has gone away. In the next two
examples we see what to do in slightly more difficult situations.
Example B.2(b): Solve the system
3x + 4y = 13
.
x + 2y = 7
We can see that if we just add the two equations together we get 4x + 6y = 20, which doesnt
help us find either of x or y. The trick here is to multiply the second equation by 3 so
that the the first term of that equation becomes 3x, the opposite of the first term of the first
equation. When we then add the two equations the x terms go away and we can solve for y:
3x + 4y = 13
x + 2y = 7
=
=
times 3
3x + 4y =
13
3x 6y = 21
2y = 8
y =
4
x + 2(4) = 7
x+8 = 7
x = 1
The solution to the system of equations is (1, 4). Note that we could have eliminated y first
instead of x:
3x + 4y = 13
x + 2y = 7
3x + 4y = 13
2x 4y = 14
x = 1
times 2
157
1 + 2y = 7
2y = 8
y = 4
4 =
3A 5B + 1
.
0 = 5A 3B 2
In this case we will chose to eliminate A because its coefficients already have opposite signs
in the two equations. Well multiply the first equation by 5 and the second by 3 so that the
coefficients of A will be the same, but with opposite signs. When we then add the two equations
the A terms go away and we can solve for B:
4 =
3A 5B + 1
0 = 5A 3B 2
times 5
times 3
20 =
0
20
21
21
34
=
=
=
=
15A 25B + 5
15A 9B 6
34B 1
34B
B
Ordinarily we would substitute this value into one of the original equations and solve for A at
this point. However, when one of the unknowns is a messy fraction it is often easier to repeat the
same procedure, but eliminate the other unknown. Lets multiply the first equation by 3 and the
second by 5:
4 =
3A 5B + 1
0 = 5A 3B 2
times 3
times 5
12 =
0
12
1
1
34
=
=
=
=
9A 15B + 3
25A + 15B + 10
34A + 13
34A
A
21
1
, B = 34
.
The solution to the system is A = 34
Lets summarize the steps for the addition method, which youve seen in the above examples.
158
Example B.2(d):
tution method.
x 3y = 6
2x + 5y = 5
Solving the first equation for x, we get x = 3y + 6. We now replace x in the second equation
with 3y + 6 and solve for y. Finally, that result for y can be substituted into x = 3y + 6 to
find x:
2(3y + 6) + 5y = 5
6y 12 + 5y = 5
x 3(7) = 6
y 12 = 5
x + 21 = 6
y = 7
x = 15
y = 7
159
When solving a system of three equations in three unknowns, matrix methods are usually employed.
However, we will encounter certain systems of three equations in three unknowns for which the
substitution method yields a solution rather easily. The next example illustrates this.
Example B.2(e):
9A
12A + 8B
3A + 6B + 12C
= 3
= 0
= 8
using the
substitution method.
We divide both sides of the first equation by 9
second equation we get
12( 31 ) + 8B
4 + 8B
8B
B
We can now substitute A =
1
3
0
0
4
12
= 8
= 8
= 10
= 56
2x + y = 13
5x + 3y = 6
(b)
2x 3y = 6
3x y = 5
(c)
x+y = 3
2x + 3y = 4
7x 6y = 13
6x 5y = 11
(b)
5x + 3y = 7
3x 5y = 23
2x 3y = 4
4x + 5y = 3
(c)
5x 3y = 11
7x + 6y = 12
(b) Ordinarily you would substitute your answer to (a) into either equation to find the other
unknown. However, dealing with the fraction that you got for part (a) could be difficult
and annoying. Instead, use the addition method again, but eliminate x to find y.
4. Consider the system of equations
such a system of equations.
1
2x
1
4x
13 y = 2
+ 23 y = 6
160
(a) Multiply both sides of the first equation by the least common denominator to kill off
all fractions.
(b) Repeat for the second equation.
(c) You now have a new system of equations without fractional coefficients. Solve that
system by the addition method.
5. Solve each of the following systems of equations.Each is of the sort that arise in solving various
initial value and boundary value problems.
(a)
4 =
C1 + C2
3 = 2C1 5C2 + 7
(b)
7A 2B = 4
2A 7A = 0
(c)
8A
= 4
10A + 8B
= 3
4A + 5B + 8C = 10
(d)
3 =
C1 + C2 3
2 = 4C1 C2 + 1
(e)
3A 8B = 2
8A 3B = 1
(f)
161
0 =
0 =
800
12
800
3
+
+
800
80
6 C1 + 2 C2
80
2 C1 + 8C2
B.3
Ax + B
, where either
(x x1 )(x x2 )
(but not both) of A or B might be zero, and find two expressions
There are times when we wish to take an expression of the form
C
x x1
such that
and
D
x x2
(1)
Ax + B
C
D
=
+
.
(x x1 )(x x2 )
x x1 x x2
(2)
The sum to the right of the equal sign in (2) is called the partial fraction decomposition of the
expression to the left of the equal sign there. The process of obtaining the right hand side is also
called partial fraction decomposition, and we illustrate it in the following example.
Example B.3(a): Find two expressions of the form (2) whose sum is
First we note that
x2
so we are looking for
x + 17
.
x2 + 4x 5
x + 17
x + 17
=
,
+ 4x 5
(x 1)(x + 5)
C
D
and
such that
x1
x+5
D
x + 17
C
+
= 2
.
x1 x+5
x + 4x 5
But
C
D
C(x + 5)
D(x 1)
Cx + 5C + Dx D
+
=
+
=
x1 x+5
(x 1)(x + 5) (x 1)(x + 5)
(x 1)(x + 5)
x + 17
, then the numerators of both fractions
(x 1)(x + 5)
must be equal (because they both have the same denominator). Note that both fractions have
the same denominator, so the two fractions will be equal only if their numerators are equal:
Now if this last expression is to equal
Cx + 5C + Dx D = x + 17
By grouping like terms, this can be rewritten (be sure you see how) as
(C + D)x + (5C D) = 1x + 17,
and these will be equal only if C + D = 1 and 5C D = 17. Now we have two equations
in two unknowns, which we know how to solve (see Appendix B.2). If we add the two equations
together we get 6C = 18, so C = 3. Substituting this into the first equations gives D = 2.
Thus
3
2
3
2
x + 17
=
+
=
.
2
x + 4x 5
x1 x+5
x1 x+5
The method of partial fractions has many complications that can arise when the expression to
be decomposed has other forms. Those complications are dealt with by varying the above process
slightly, but for our needs the above method is sufficient.
162
1. Add
4x + 7
x2 + 5x + 6
(b)
14
x2 3x 10
11 x
x2
(d)
4x 10
x2 1
x2
163
B.4
Try the following: Set your calculator for radians (as you always should in this course) and find
sin(0.05), sin(0.1) and sin(0.5). You should get (to five places past the decimal) 0.04998,
0.09983 and 0.47943, respectively. These numbers are fairly close to the numbers that you were
finding the sine of, with the one closest to zero being the best approximation. This seems to indicate
that
sin x x
for values of x near zero. Since of course sin(5) must be a number between 1 and 1 this
will certainly not hold in that case!
1 5
x . Here we find that if we round to five
Now consider the function f (x) = x 16 x3 + 120
places past the decimal we get f (0.5) = 0.47943, the value of sin(0.5) when rounded to the same
number of decimal places. Lets try something bigger:
f (1.3) = 0.96477,
sin(1.3) = 0.96356
f (2.0) = 0.93333,
sin(2.0) = 0.90930
It appears that this function f comes pretty close to approximating the sine function, especially
for values of x nearer to zero.
The fraction coefficients in the equation for the function f appear to be a bit mysterious, but it
turns out that 6 = 321 and 120 = 54321. We call the quantity n(n 1)(n 2) 321 n
factorial, denoted by n! So 6 = 3!, 120 = 5! and
sin x x
x3 x5
+
3!
5!
It turns out that (in a sense that you really must take a course in sequences and series to understand)
sin x = x
x3 x5 x7 x9
+
+
3!
5!
7!
9!
This is called the (infinite) series representation of the sine function. As you have seen, for
values of x near zero, very good approximations of sin x can be obtained by using just the first
term or few terms of this series. For values farther from zero, more terms must be used to get a
good approximation.
Cosine and the exponential function have series representations as well:
cos x = 1
and
ex = 1 + x +
x2 x4 x6 x8
+
+
2!
4!
6!
8!
x2 x3 x4 x5 x6
+
+
+
+
+
2!
3!
4!
5!
6!
Recall now i, the imaginary unit (j for you electronics folks), for which i2 = 1.
also compute things like
i1 = i,
i3 = i2 i = i,
i4 = i3 i = i i = i2 = 1,
164
We can
i5 = i4 i = 1i = i
and so on. Because we can compute these, we can now find something like ei by using the series
representation of ex :
ei = 1 + i +
= 1 + i +
= 1 + i
=
=
2
2!
2
2!
+ + i i + i +
4!
6!
3!
5!
4 6
3 5
+
+ + i
+
+
4!
6!
3!
5!
= cos + i sin
A similar computation can be done for ei . The result of the two of these is
Eulers Relations:
ei = cos i sin
ei = cos + i sin
165
The point of numerical methods is that they can be used to solve IVPs that cannot be solved
analytically. (We will sometimes use numerical methods with IVPs that we CAN solve analytically,
so that we can compare numerical solutions with analytical solutions.) So numerical methods are
needed when we want a function y = y(t) that we cant use algebra and calculus methods to
find an equation for. The use of numerical methods (both for solving IVPs and for other things)
is a huge part of mathematics, science and engineering. Here you will be introduced to some
methods for finding solutions to first order initial value problems. The simplest method is called
Eulers method; the name comes from Leonhard Euler (pronounced oiler), a very prolific Swiss
mathematician who lived in the 1700s. You will see that this is not a very good method, but we
will study it because it is simple to implement, it illustrates what are called iterative techniques,
and the ideas behind it are the same ones that make some better methods work.
Before actually finding a numerical solution to an IVP, lets go over what a solution will consist
of, and the special notation well be using in this section. Suppose that we have an IVP
dy
= f (t, y),
dt
y(t0 ) = y0 ;
the solution to the IVP is some unknown function y = y(t). Consider the sequence of times
t = 0, 0.2, 0.4, 0.6, 0.8, ... seconds, which well denote as t0 , t1 , t2 , t3 , t4 , .... Note that these times
are spaced apart by the increment of 0.2 seconds - we could just as well have used an increment
of 0.1 seconds, which would give us instead
t0 = 0,
t1 = 0.1,
t2 = 0.2,
t3 = 0.3, ...
The increment used is denoted by the letter h, so for the first sequence of times listed above
h = 0.2 and for the second sequence h = 0.1. We will refer to the value of h as the step size,
meaning that time advances in steps of size h. Regardless of the value of h, we always start with
a known value of t0 (usually zero) and h, then we use the relations
t1 = t0 + h,
t2 = t1 + h = t0 + 2h,
t3 = t2 + h = t0 + 3h,
t4 = t3 + h = t0 + 4h, ...
y2 y(t2 ) = y(0.4),
Each value in the sequence must be determined before the following term can be found, so we use
the known value of y0 to find y1 , then use that value of y1 to find y2 , use that to find y3 ,
166
and so on, as far as we wish to have values for y. It should be noted that to find any value in the
y sequence we always use the previous value of y, but might also use other prior y values and
perhaps some time values as well.
Any method that proceeds sequentially like this is called an iterative method. To carry it
out we need an iteration formula that tells us how to obtain a y value from previous y values
and perhaps previous time values. The simplest such formula, as mentioned before, is the one for
Eulers method:
dy
= yn + hf (tn , yn )
(1)
yn+1 = yn + h
dt (tn ,yn )
Lets think about what this tells us. It says that to get the next y value yn+1 we take the
current y value yn and add the product of the increment h and the derivative evaluated at
dy
our current t and y values. But because the ODE can be put in the form
= f (t, y), we use
dt
the final expression above to compute each new y value.
You may (understandably) be totally confused by now! Lets look at a specific example.
Example C(1): Use the Eulers method equation (1) to find y0 , y1 , y2 , y3 and y4 for the
initial value problem
dy
= t y,
y(0) = 2,
dt
using a step size of h = 0.2.
First we identify t0 = 0 and y0 = 2 (from the initial value), t1 = 0.2, t2 = 0.4, ... from t0 and
h, and f (t, y) = t y (from the ODE). By (1),
y1 = y0 + hf (t0 , y0 ) = 2 + (0.2)(0 2) = 1.6
We then use this value and (1) again to obtain
y2 = y1 + hf (t1 , y1 ) = 1.6 + (0.2)(0.2 1.6) = 1.32
Continuing, we get
y3 = y2 + hf (t2 , y2 ) = 1.32 + (0.2)(0.4 1.32) = 1.136,
y4 = y3 + hf (t3 , y3 ) = 1.136 + (0.2)(0.6 1.136) = 1.0288.
More y values, if desired, could be obtained by continuing in the same way.
dy
works. Recall the definition
= yn + h
dt (tn ,yn )
dy
y(t + h) y(t)
= lim
.
h0
dt
h
This says that as h approaches zero, the quantity on the right approaches the derivative. Thus,
for small values of h,
dy
y(t + h) y(t)
.
dt
h
Multiplying both sides by h and solving for y(t + h) gives us
y(t + h) y(t) + h
167
dy
,
dt
But y(tn+1 ) yn+1 and y(tn ) yn , and our expression f (t, y) for the derivative may contain
both t and y, so we must evaluate the derivative at (tn , yn ), giving us the final result
dy
yn+1 = yn + h
= yn + hf (tn , yn ).
(1)
dt (tn ,yn )
This formula can also be derived using something called Taylor polynomials, which many of
you have not yet seen. For this reason we will not go into it here but, if you are interested, you can
see the derivation of Eulers method from a Taylor polynomial in most any introductory differential
equations book. Using Taylor polynomials has a couple advantages:
Their use can lead to better formulas than the one for Eulers method.
Their use can give us a bound for the error that occurs when using a numerical method.
The distinction is a bit subtle and often blurred, but when we just use an iterative formula,
as we will, we are performing a numerical method. When we try to analyze our method for how
accurate it is, we are doing something called numerical analysis.
There is also a geometric explanation for how and why Eulers method works. Suppose that we
have the initial value problem
dy
= f (t, y),
dt
y(t0 ) = y0
with solution y = y(t). Figure C(1) on the next page shows the solution curve (which is really
unknown) through (t0 , y0 ) and the points on that curve corresponding to two times t1 and t2 as
described previously. We can think of generating the approximate numerical solution consisting of
a sequence y0 , y1 , y2 , y3 , ... as follows:
We are given the initial value y0 .
We construct a line through (t0 , y0 ) that is tangent to the solution curve - its slope is the
derivative at (t0 , y0 ). See Figure C(2). The point on that line with t-coordinate t1 is
y1 (Figure C(3)).
We construct a line through (t1 , y1 ) that is tangent to the solution curve passing through
that point. Its slope is the derivative at (t1 , y1 ), and the point on the line with t-coordinate
t2 is y2 . See Figure C(4).
Continue in this manner to obtain the values y3 , y4 , ....
168
(t0 , y0 )
(t0 , y0 )
solution y = y(t)
(t1 , y(t1 ))
y(t1 )
(t1 , y(t1 ))
(t2 , y(t2 ))
y(t2 )
(t2 , y(t2 ))
dy
slope =
dt (t0 ,y0 )
| {z } | {z }
t1
t2
h
h
t1
Figure C(1)
t2
Figure C(2)
(t0 , y0 )
(t0 , y0 )
(t1 , y(t1 ))
(t1 , y(t1 ))
(t1 , y1 )
(t2 , y(t2 ))
(t1 , y1 )
(t2 , y2 )
t1
t2
dy
slope =
dt (t1 ,y1 )
t1
Figure C(4)
Figure C(3)
169
t2
Eulers Method
dy
= f (t, y), y(t0 ) = y0 for some x increment
To solve the first order IVP
dt
h, use the recursive relation
dy
= yn + hf (tn , yn )
(1)
yn+1 = yn + h
dt (tn ,yn )
as follows:
1) y0 = y(t0 )
2) Use the above relation with n = 0 to find y1 .
3) Use the relation with n = 1, remembering that t1 = t0 + h, to find
y2 .
4) Repeat as long as desired, remembering that tn = t0 + nh.
There other methods besides Eulers method for generating approximate solutions to IVPs. The
general method for applying them is the same as Eulers method, but the iteration formula (1) is
different for them. You will see some of those other methods in the exercises.
Appendix C Exercises
For Exercises 1 and 2 you will be considering the ODE
dy
= y t.
dt
1. In this exercise you will use Eulers method to construct a numerical solution for an initial
value of y(0) = 2 and a step size of h = 0.2
(a) Build a table with three columns, the first for n = 0, 1, 2, ..., the second for tn , and
the third for yn . Complete the second column for t = 0 to t = 1, and finish the top
row by putting in the initial value y0 = y(0).
(b) Use the recursion formula yn+1 = yn + hf (tn , yn ) and the values in the first row of your
table to find y1 . Put the result in your table. Then continue computing yn values
up to time one. Do not round any of your answers and use all digits of each
yn to find yn+1 .
(c) Solve the IVP, using the integrating factor method to solve the ODE.
170
(d) Add a fourth column to your table for the exact solution values y(t0 ), y(t1 ), y(t2 ), ...
Compute those values, rounding to four places past the decimal using your answer to
(c), and add them to your table.
100 | approximate value actual value |
(e) The percent error is percent error =
. Find
| actual value |
the percent error for each of your approximations, and add them to the table in a fifth
column. Round to the nearest hundredth of a percent.
2. (a) Find a recursion formula for the given ODE and a step size h = 0.05 as follows: For
the formula given in 1(b), insert 0.05 for h and apply the function f (t, y) = y t to
tn and yn . Distribute the 0.05 and combine like terms. Your final result should be
something of the form yn+1 = ayn + btn , for some constants a and b.
(b) Repeat Exercise 1, but for an initial value of y(0) = 1.4, a step size of h = 0.05, and
from t = 0 to t = 0.2, using your formula from part (a) to find each successive yn .
Round both the yn and y(tn ) values to four places past the decimal, and
the percent error to two places past the decimal.
3. Consider the IVP
dy
+ ty = 0, y(0) = 2.
dt
(a) Using Eulers method, which is the recursion formula given in Exercise 1(b), find the
first four approximate solutions (including y0 , which is really exact) to the IVP using
a step size of h = 0.1. Compile your results in a table, like you have been doing.
(b) Use separation of variables to solve the IVP. Use your answer to add a column to your
table for the exact values y(tn ).
So far you have approximated solutions to an ODE using an iterative method called Eulers
method. Now you will use two other methods that give better approximations. For any first order
dy
= f (t, y), Eulers method uses the recursion formula given in Exequation, written in the form
dt
ercise 1(b) to find each successive solution approximation from the previous approximation.bEulers
method is the simplest of a variety of methods called Runge-Kutta methods. The derivation
of these methods is beyond the scope of our course, but well use two of them without necessarily
understanding their derivations. The first method well look at is called the midpoint method,
which has a fairly straightforward geometric interpretation that well go over in class. Here is the
formula for the method:
h
h
(1)
yn+1 = yn + hf tn + , yn + f (tn , yn )
2
2
4. Lets use the midpoint method to obtain an approximate solution to the initial value problem
dy
= y t, y(0) = 2,
dt
with a step size of h = 0.2, from t = 0 to t = 1. As in Exercise 2, you will first derive a
recursion formula specific to this problem.
h
f (tn , yn ) for our particular ODE and step size, in the
2
manner you did for Exercise 2(a).
171
(b) Insert your answer from (a) into the appropriate place in (1) and simplify (1) for our
ODE. Go until you obtain something of the form yn+1 = ayn +btn +c for some constants
a, b and c.
(c) Use your answer to (b) to fill out a table of values like you did for Exercise 1, up to
time t = 1. As before, include a column at the right for exact values of the solution,
rounded to four places past the decimal. (Remember that the exact solution for this
IVP is y = t + 1 + et .)
(d) How do the percent errors for this method compare with those for Eulers method?
The most commonly used Runge-Kutta method uses a set of equations to obtain each successive
approximation. These could all be combined into one equation like (1), but it would be very
cumbersome. Instead we use
1
yn+1 = yn + (k1 + 2k2 + 2k3 + k4 ) ,
6
where
k1 = hf (tn , yn ),
k2 = hf
h
1
tn + , yn + k1 ,
2
2
k3 = hf
h
1
tn + , yn + k2 ,
2
2
k4 = hf (tn+1 , yn +k3 )
The values k1 , k2 , k3 and k4 must be computed each time you iterate to find a new yn+1 .
dy
= yt, y(0) =
5. Now well use the above method to approximate a solution to the same IVP
dt
2.
(a) Find simplified forms of k1 through k4 for our particular ODE, again with a step size
of h = 0.2.
(b) Build a table of values with columns for n, tn , yn , k1 , ... , k4 and fill it out up through
time t = 1. As before, round to four places past the decimal when necessary.
(c) Compare your approximate values with the exact values in your table for Exercises 1
and 4.
172
173
Solutions to Exercises
D.1
Chapter 1 Solutions
2. (a) The independent variables are time t and the distance x out from the edge of the
table, and the dependent variable is the deflection of the ruler at any point and time.
(b) 0 t, 0 x 6, where x is measured in inches. (Substitute 0.5 for 6 if measuring
in feet instead of inches.)
3. 0 r 5 inches, 0 2, 0 t Note that, unless we are doing right triangle
trigonometry, angles will be measured in radians.
height (feet)
4. If measuring in feet, 0 x 1, 0 y 1, 0 z 1.
5. (a) See graph to the right.
(b) The shape is a parabola, opening downward.
(c) We would model the height with a quadratic function.
5
0 time (seconds)
T ( F)
(b)
(c)
T ( F)
Exercise 5(a)
T ( F)
80
40
40
40
30
t
7. (a) y
(b) y
174
dy
= 6 cos 3x
dx
(b) y = 2e0.5t
(e)
(c) x = 2t + 5
dy
= 3te3t + e3t
dt
d2 y
= 18 sin 3x
dx2
(b) y = e0.5t
(c) x = 2
dy
= 3y
dx
(b)
d2 y
= 9y
dt2
(c)
d2 x
= 9x
dt2
(d)
d2 y
= 5y
dx2
4. (a) The independent variables are time and the distance along the string.
(b) Some parameters would be physical properties of the material the string is made of,
the thickness and cross-section of the string (which is probably circular, but could be
different) and the tension in the string.
Section 1.4 Solutions
1. (a) y = 7 when x = 0, y = 3 when x = 0
(b) x = 1 when t = 0, x = 5 when t = 0
3. (a) 5
0
5
(d)
(c)
3
0
3
1
1
Dependent variable: y
Dependent variable: y
Dependent variable: u
Dependent variable: y
Dependent variable: u
Dependent variable: u
Dependent variable: y
Dependent variable: y
Dependent variable: u
Dependent variable: y
Dependent variable: y
3. no
9. (a) no
4. yes
7. A = B = 1.
(c) y = 2e3x
(c) F (x, y) = y y 2
(e) F (x, y) = x xy
(i) F (x, y) = 1 xy
1
x
(d) Solution.
t + 5 e4t + 3t + 1
(b) x = 14
3 e
3
2
(c) y = 3
sin 5t 3 cos 5t
5
177
D.2
Chapter 2 Solutions
(c)
1 5
5y
(b)
= 13 x3 + C or
(d) ln |y| =
3y 5 = 5x3 + C
(e) ey = x1 + C or
ey =
1
x
1 2
1
y = 2 + C or
2
2x
1
y2 = 2 + C
x
(f)
1 2
2y
1
2
ln |2x + 3| + C
= 52 x2 + 3x + C or
y 2 = 5x2 + 6x + C
+C
2. (a) ln |y| = 21 x2 + ln 3
1
2
1 2
2x
(b)
= 25 t2 + 3t + 12 or
x2 = 5t2 + 6t + 24
(c) ey = 23 x2 + e2
(d)
3. (a) y = 4e3x
(d) y =
4. (a) y = ln(x2 + x + 1)
(b) y = ln(x2 + x + e3 2)
(c) y = 7et
5
2
5. (b) y = Cet
x+2
x5
7 12 x2
e2 e
(f) y = 21 x +
1
1
= sin t
3
3y
24
2
+C
x
+C
(b) y = ln
x + 3
y
|y|
y
=x+C
7. (a)
(b)
= x + C or
= Cex
|y 3|
y3
y3
3C
(c) y =
(d) In order, the constants are C = 71 , 0, 12 , 4 and
C ex
3
12
3
, y = 0, y =
, y=
the solutions are y =
x
x
1 7e
1 + 2e
4 ex
(f) y 3 as x in all cases
(g) The value of the constant cannot be determined - the equation to be solved has no
solution.
178
(a) C = 3
(b) C = 1
(c) C = 0
(d) C = 2
-2
3. (a) II
(b) I
(c) III
(d) IV
4. (a) The top, U-shaped curve is for initial value y(0) = 1. The next curve down has initial
value y(0) = 1, and the one below that has initial value y(0) = 2.
(b) Using the point (2, 1), the value obtained for C is 2.
point (2, 1), the same value of C is obtained!
(c) When the solution is graphed for C = 2, the graph includes the U-shaped curve as
well as the two curves in the lower left and lower
right. Theyare all parts of the same
graph, which has vertical asymptotes at x = 2 and x = 2.
8.
(a)
(b)
4
-1
-1
-2
-2
-1
(c)
(d)
-1
-1
-2
-2
0
-2
179
-1
(c) y = 32 e 2 x
6
29
sin 2t
15
29
cos 2t + Ce5t
5. (a) y = Ce3t + 13 t2 +
6. (a) y = 12 e3x + 27 ex
7. (a) r = 5
(b) y =
13
9 t
(b) y =
22
27
(b) y =
6
29
sin 2t
76 3t
27 e
15
29
+ 31 t2 +
cos 2t
13
9 t
101 5t
29 e
22
27
(b) y = x ln x + 2x
6
29
sin 2t
15
29
cos 2t
(c) y =
6
29
sin 2t
15
29
cos 2t + Ce5t
(c)
t
t
y = 0 is an unstable equilibrium
y = 3 is a stable equilibrium
y = 0 is an unstable equilibrium
(d)
(f)
y = 2 is a stable equilibrium
y = 5 is an unstable equilibrium
y = 0 is an unstable equilibrium
y = 3 is a semi-stable equilibrium
180
4.
(a)
(b)
(c)
5
4
2
2
1
0
5.
6. (c)
dy
(a)
= y(y 2)(y 6)
dt
dy
= (y 4)(y 1)2
(b)
dt
dy
= y(y 2)(y 5)
(c)
dt
0
1
dy
= 13 y(y 3).
dx
(b) A = 80ekt
dA
3
= A, A(0) = 87
dt
10
3. (a) T = 70 + Cekt
(b) i =
6
6 20t
or
5 5e
5
101 (10 cos 2t +
(b) A = 87e 10 t
(c) k = ln 19
6 1.15
4. (a) i =
(d) Steady-state: 70
i = 56 (1 e20t )
50 20t
101 e
50 20t
101
e
and
sin 2t)
or i =
50
101
Transient: 38e ln
cos 2t +
5
101
sin 2t
19
t
6
38e1.15t
50 20t
101 e
5
101 (10 cos 2t
+ sin 2t).
5. (a) If T0 = Tm the temperature will not change because the initial temperature of the
object will be the same as the temperature of the medium; the solution will be completely
steady-state.
A (lbs)
(b) The steady-state solution is Tm .
(c) The transient solution is (T0 Tm )ekt
24
181
dT
= k(T 73), T (0) = 29
dt
(b) There will be a steady state solution of T = 73.
7. (a)
(c) T = 73 44ekt
8. (a) Initial value problem:
Solution: i =
(b) Steady-state:
(d) T = 73 44e0.263t
3 di
4 dt
2
101 (20 cos 2t
+ 2 sin 2t) +
2
101 (20 cos 2t
+ 2 sin 2t)
162 20t
101 e
Transient:
dA
7A
= 7
, A(0) = 450
dt
150
162 20t
101 e
182
D.3
Chapter 3 Solutions
(b) y = C1 x1 + C2 x2
2. (a) y = C1 e3t + C2 et
(c) y = C1 x + C2 x 3
(e) y = C1 et + C2 e2t
(g) y = C1 et + C2 tet
(b) y = C1 + C2 e25t
(b) y = C1 er1 t + C2 er2 t
5. (a) y = Cx2
Section 3.2 Solutions
1. (a) yp = At + B
(d) yp = A sin t + B cos t
2. (a) yp = 52 t
(c) yp = Ae5t
(e) yp = A
(f) yp = At + B + Cet
17
4
9
sin t +
(d) yp = 13
(b) yp =
6
13
cos t
5
18
sin 3t
(e) yp = 7
(c) yp =
(f) yp =
1 5t
17 e
2
3 t
3t + 4e
1
(b) y = C1 e5t + C2 e 2 t
(b) D(y) = 0
17
4
(d) D(y) = 5t 1
(b) S(4 cos t) = 4 cos t + 3, 4S(cos t) = 4(cos t + 3) = 4 cos t + 12 These are not the same,
so S is not linear.
(c) S(cos t + e2t ) = cos t + e2t + 3, S(cos t) + S(e2t ) = cos t + 3 + e2t + 3 = cos t + e2t + 6
183
1. (a) y = C1 e2t + C2 et + 25 t
(c) yh = C1 sin 3t + C2 cos 3t +
17
4
1 5t
17 e
7
6
3
5
(b) y = C1 x 2 + C2 x 2
3. (a) y = C1 x2 + C2 x3
Section 3.5 Solutions
1. (a) yp = 56 t
13
36
(b) yp =
35
986
13
8
sin 2t +
1
2
cos 2t 34 t cos 2t
(c) Steady-state:
3
4
3
5
cos 7t
5
3
sin 5t
217
986
cos 5t
(c) yp = 45 te2t
cos 3t
Transient: none
sin 5t 56 cos 5t
7 2t
2e
7e5t + et
Transient:
Transient: 3te5t
(b) c1 = 1, c2 = 2
(c) c1 = 1, c3 = 1
8. y2 = x 2
sin 3t
9
6
(d) yh = C1 et + C2 e5t 13
sin t + 13
cos t
(e) yh = C1 e 2 t + C2 et + 7
2. (a) y =
5
18
9. y2 =
1
x2
184
D.4
Chapter 4 Solutions
2.
3.
2.5
2
t
t
2
2.5
y(0) = 2.5, y (0) = 0
y(0) = 0, y (0) = 3
y(0) = 2, y (0) = 8
3
4y
5
t)
+
2.0
cos(2
5
t)
if
working
in
centimeters,
y
=
0.018
sin(2
5 t) +
5. (a) y = 1.8 sin(2
11 2t
3 e
35 e8t
(c)
(d)
t
t
t
t
1
C < 0.
105 cos 1530t)
coulombs
105 e1670t
185
(c) We should expect no transient solution because there is no damping. Note that whenever
there is damping the homogeneous solution is the transient part of the solution. When
there is no damping, the homogenous solution will not be transient.
Beats: viii
Under-damped: iv, v
Over-damped: i
Under-damped: iv, vi
(b) i, vi
(c)
Equation
Solution
Graph
ii
vi
ii
iii
ii
i
186
Critically or over-damped: i, vi
Equation
Solution
Graph
iv
vii
vii
iii
iii
vi
iv
vii
iv
vi
viii
viii
187
D.5
Chapter 5 Solutions
3. (a) (30)(80)
(d) The maximum deflection should appear at the middle of the beam (x = 4).
deflection there is 23 .
4. (b) y =
1 4
12 x
10 3
3 x
The
+ 50x2
(d) The maximum deflection is 2500 at x = 10, the right hand end of the beam.
5. (b) y =
1 4
12 x
35 x3 +
250
3 x
1 4
384 x
5 3
96 x
3125
12
at x = 5
+ 41 x2
(b) E and H
(c) F and G
3
4
1. (a) y = C sin 12
x, y = C sin 2
12 x, y = C sin 12 x, y = C sin 12 x, . . .
(b) P =
6250 2
9 ,
6250 2
6250 2
2
4( 6250
9 ), 9( 9 ), 16( 9 ), . . .
(c) The third critical load is nine times the first critical load.
3
4
2. (a) y = C sin 6 x, y = C sin 2
6 x, y = C sin 6 x, y = C sin 6 x, . . .
25000 2
25000 2
25000 2
2
P = 25000
9 , 4( 9 ), 9( 9 ), 16( 9 ), . . .
(b) Each critical load is four times the corresponding critical load for the twelve foot column.
(c) The first buckling mode for the six foot column is the same as the second buckling mode
for the twelve foot column, but is half as high, so it only includes half a period of the
sine function, whereas the twelve foot columns second buckling mode has a full period
of the sine function.
4
6
3. (a) y = C cos 2
12 x, y = C cos 12 x, y = C cos 12 x, . . .
6250 2
6250 2
2
(b) P = 4( 6250
9 ), 16( 9 ), 36( 9 ), . . .
(c) The third critical load is nine times the first critical load.
(d) Each critical load is four times the corresponding critical load for the column with pinned
ends.
188
3
4
4. (a) y = C cos 12
x, y = C cos 2
12 x, y = C cos 12 x, y = C cos 12 x, . . .
(b) P =
6250 2
9 ,
6250 2
6250 2
2
4( 6250
9 ), 9( 9 ), 16( 9 ), . . .
(c) The third critical load is (again!) nine times the first critical load.
(d) The critical loads are the same as those for the pinned ends.
3
4
5. (a) y = C sin L x, y = C sin 2
L x, y = C sin L x, y = C sin L x, . . .
(b) P =
EI 2
L2 ,
EI 2
EI 2
2
4( EI
L2 ), 9( L2 ), 16( L2 ), . . .
3
4
5. (a) y = C cos L x, y = C cos 2
L x, y = C cos L x, y = C cos L x, . . .
(b) P =
EI 2
,
L2
4( EI
2 ), 9( EI
2 ), 16( EI
2 ), . . .
L2
L2
L2
Eigenvalue: k
(b) y = Ax + B
Eigenvalue: k2
Eigenvalue: k2
Eigenvalue: k2
189
D.6
(c) y =
10 x
4. (a) f (x) =
18
x4
1
4
(d) s (t) = 6t 4
(b) g (t) =
x
5. (a) g(x) =
16 x2
1 36
4
3
t
(b) y = 3
2 sin
6. (a) y = 10e2x
(d) s (t) = 52 sin
(b) h (t) = 32
1
12
(b) g (t) = t + 3
3
t
4
(d) h (t) =
3
3 x4
6
x3
3. (a) f (x) =
2t
(b) x = 12 cos 3t
2
5t
(e) y = 2xex
(b) 15 sin(5x 2)
(c) 12x3 ln x
(b) y =
(b) (3, 4)
2. (a) (1, 1)
(b) (1, 4)
(c) (2, 31 )
5
29
, 11
)
3. ( 22
5. (a) C1 =
4. (8, 6)
10
3 ,
C2 =
2
3
(c) A = 12 , B = 1, C =
(e) A = 14
73 , B =
(b) A =
13
8
8
B = 53
(d) C1 = 13 , C2 =
13
73
(f) C1 = 19, C2 =
+ 5x + 6
x+3 x+2
x+3 x+2
14
2
2
(b)
=
+
x2 3x 10
x5 x+2
1. (a)
28
53 ,
x2
190
1
3
185
3
4
3
11 x
=
+
x2
x+1 x2
4x 10
7
3
7
3
(d)
=
+
=
2
x 1
x+1 x1
x+1 x1
(c)
x2
Appendix C Solutions
1.
tn
yn
y(tn )
% error
0
1
2
3
4
5
0.0
0.2
0.4
0.6
0.8
1.0
2
2.4
2.84
3.328
3.8736
4.48832
2
2.4214
2.8918
3.4221
4.0255
4.7183
0
0.88
1.79
2.75
3.77
4.87
3. (a)
tn
yn
y(tn )
% error
0
1
2
3
0.0
0.1
0.2
0.3
2
2
2.02
2.0604
2
2.01
2.0404
2.0921
0.00
0.50
1.00
1.52
tn
yn
y(tn )
% error
0
1
2
3
4
0.00
0.05
0.10
0.15
0.20
1.4
1.47
1.541
1.6131
1.6863
1.4
1.4705
1.5421
1.6147
1.6886
0.00
0.03
0.07
0.10
0.14
1 2
(b) y = 2e 2 t
tn
yn
y(tn )
% error
0
1
2
3
4
5
0.0
0.2
0.4
0.6
0.8
1.0
2
2.42
2.8884
3.4158
4.0153
4.7027
2
2.4214
2.8918
3.4221
4.0255
4.7183
0
0.06
0.12
0.18
0.25
0.33
(d) The error using the midpoint method are much smaller than those obtained using Eulers
method with the same step size.
5. (a) k1 = .2yn .2tn ,
k2 = 0.22yn 0.22tn 0.02,
k4 = 0.2444yn 0.0444tn 0.044 0.2tn+1
Note the appearance of both tn and tn+1 in k4 .
(b)
tn
yn
k1
k2
k3
k4
0
1
2
3
4
5
0.0
0.2
0.4
0.6
0.8
1.0
2
2.4214
2.8918
3.4221
4.0255
4.7182
0.4
0.4443
0.4984
0.5644
0.6451
0.42
0.4687
0.5282
0.6009
0.6896
0.422
0.4712
0.5312
0.6045
0.6941
0.4444
0.4985
0.5646
0.6453
0.7439
(c) The approximations obtained using this method are very close to the exact values.
191
Index
ambient temperature, 67
ampere, 68
amplitude, 112
analytical solution, 3
angular frequency, 112
autonomous equation, 34, 59
auxiliary equation, 80
beats, 109
bending moment, 129
Bessel equation, 33
boundary conditions, 21, 24, 130
boundary value, 24
boundary value problem, 130
buckling mode, 135
capacitance, 76
carrying capacity, 14
charge, 76
coulombs, 76
critical load, 136
critical point, 60
critically damped, 109
current, 68
damping, 26
deflection of a horizontal beam, 129
dependent variable, 10
differential equation, 3, 15, 17, 27
differential operator, 88, 89
direction field, 52
domain, 5, 11
eigenfunction, 141
eigenvalue, 141
electric circuit, 68
electromotive force, 68, 69
elliptic partial differential equation, 42
embedded, 24, 129
equilibrium, 10, 41
equilibrium solution, 59
Euler equation, 77, 79
Eulers formula, 82
eulers method, 166
family of solutions, 29
farad, 76
first critical load, 128
forced, 109
ohms, 68
one-parameter family of solutions, 30
operator, 88, 139
order of a differential equation, 17, 23, 27, 32
ordinary differential equation, 12, 27, 32
overdamped, 109
overdamped vibration, 117
parabolic partial differential equation, 42
parameter, 20, 112, 116
partial differential equation, 12, 25, 27, 32
partial fraction decomposition, 162
particular solution, 84
particular equation, 30
particular solution, 78, 84, 91
period, 12, 112
phase portrait, 60
phase shift, 111, 112
pinned, 26
pinned boundary condition, 134
population growth, 64
qualitative solution, 3
radioactive decay, 64
reduction of order, 74, 101
relation, 46
resistance, 68
resistor, 68
resonance, 109
response, 69
RLC circuit, 115
Runge-Kutta method, 171
second critical load, 128
semi-stable equilibrium, 62
separable equation, 34
separation of variables, 44, 45
simple harmonic motion, 12, 17
simply supported, 26, 131
slope field, 52
solution curve, 49, 50
solution to a differential equation, 28
spring constant, 17
stable critical point, 60
stable equilibrium, 60
steady-state heat equation, 41
steady-state response, 70
steady-state solution, 70
step size, 166
system, 69
system response, 69, 109
193