MTH312 Notes

Differential Equations and Vector Calculus
(MTH 312): Notes
Adam Szava
Fall 2021
Introduction
This is my compilation of notes from Differential Equations and Vector Calculus
(MTH 312) from Ryerson University. All information comes from my professor’s
lectures, the textbooks Elementary Differential Equations with Boundary Value
Problems, OpenStax: Calculus Volume 3, and online resources.
In this text, definitions are boxed in red, theorems are boxed in blue, and
any examples are boxed in green.
Introductory Topics and Definitions

The following section is a collection of definitions and discussions which I feel
need to be properly defined before you can being this course, however they were
not included in the actual course material. All information comes from the text-
book Advanced Engineering Mathematics by Dennis G. Zill. I also recommend
this textbook to be used for further reading.
We begin with the most important definition of the course, the formal defi-
nition of a differential equation:
Definition 0.1.1 Differential Equations
An equation containing derivatives of one or more depen-

dent variables with respect to one or more independent vari-
ables is a differential equation (DE).
A DE with only one independent variable is called an ordi-
nary DE (ODE).
A DE with multiple independent variables is called a partial
DE (PDE).
The order of a DE is the order of the highest order derivative
in the equation.
1
There are a few forms you can write a differential equation.
1. General Form for nth order ODEs:
F (x, y, y 0 , y 00 , . . . , y (n) ) = 0
... where F is a function of n + 2 variables. We assume in this course that

it is always possible to rearrange general form into normal form...
2. Normal Form for nth order ODEs:
y (n) = f (x, y, y 0 , y 00 , . . . , y (n−1) )
3. Differential Form for first order ODEs:
M (x, y)dx + N (x, y)dy = 0
These two equations are the key to solving DEs:

dy
= y0
dx
dy = y 0 dx
... this may seem obvious if you think of the differential as a fraction, which is
fine to do for our purposes.
You can convert between differential form and normal form by using the
previous equations:
M (x, y)dx + N (x, y)dy = 0
N (x, y)dy = −M (x, y)dx
M (x, y)
dy = − dx
N (x, y)
dy M (x, y)
=−
dx N (x, y)
... and making the substitution f (x, y) = − M (x,y)
N (x,y) :
y 0 = f (x, y)
... which is exactly the definition of a first order ODE in normal form.
Definition 0.1.2 Solutions to a DE
Any function φ, defined on some interval I possessing n

derivatives that are continuous on that interval I, which
when substituted into the DE it becomes an identity, is a
solution to the DE on I.
2
Any function that can be substituted into the DE and the resulting equation
is true, is a solution. You must define what interval the solution is defined on,
because it does matter.
Definition 0.1.3 Implicit Solutions
A relation G(x, y) = 0 is an implicit solution to an ODE

on I if there exists a function y = φ(x) that satisfies both
the relation, and the ODE on I. In this course we assume
that that function φ(x) always exists.
You can implicitly differentiate the relation G(x, y) = 0 to find the deriva-
tives of y with respect to x, and substitute that into the DE.
A DE can have a family of solutions which come from an arbitrary parameter
in it’s explicit or implicit form. This corresponds to an infinite set of solutions.
An exact solution is a solution without a parameter. Singular solutions do
not belong to any families and are usually trivial.
Review of Differential Equation Content from MTH

240
4.3 Separable Differential Equations
This section begins with a lot of definitions to familiarize with terminology
relating to differential equations.
Definition 1.1.1 Terminology
A Differential Equation (DE) is an equation in variables

x, y, y 0 , y 00 , y 000 , etc, where y is a function of x.
A solution to a DE is a function f (x) where subbing in
y = f (x) makes an equation which is always true.
The order of a DE is n where the nth derivative of y is
the highest order term which appears.
The general solution to a DE is a parametrized family
of functions.
An initial value problem consists of a DE with a specific
condition. A solutions to this type of problem is a solution
to the DE that satisfies the conditions.
3
Definition 1.1.2 Seperable D.E.
A DE is separable if it can be rearranged into the form:
y 0 = f (x) g(y)
As in you can separate x and y.
In order to solve a separable differential equation, you follow this formula:

Z Z
1
dy = f (x) dx
g(y)
Do the integration, and rearrange for y if possible.
4.5 Linear Differential Equations
Definition 1.2.1 Linear D.E. (L.D.E.)
A linear differential equation (LDE) is a DE which can

be written in the form:
y 0 + p(x)y = q(x)
Notably the coefficient on the y 0 term is 1. In order to find a solution to this

type of DE, we must first define an Integrating Factor.
Definition 1.2.2 Integrating Factor
In general (not just for LDEs) an Integrating Factor is a

function usually denoted I(x) which, when multiplied on
both sides of a DE, allows both sides to be integrated. For
LDEs, the Integrating Factor will always be in the form:
R
p(x) dx
I(x) = e
Where p(x) is the same p(x) as in the LDE.
This helps us solve the LDE because of the following:
y 0 + p(x)y = q(x)
R R R
p(x) dx 0 p(x) dx p(x) dx
e y +e p(x)y = e q(x)
4
Z R R
Z R
p(x) dx 0 p(x) dx p(x) dx
e y +e p(x)y dx = e q(x) dx
Notice the the left side will always be the result of a product rule differentiation.
So: R
Z R
e p(x) dx y = e p(x) dx q(x) dx
Z
1 R
p(x) dx
y= R e q(x) dx
e p(x) dx
And now we have the solution to the LDE.
Chapter 2: First Order Equations

Differential equations model the real world. Solving a differential equation pro-
vides and interpretation of a real phenomenon. The first part of this course is
concerned with beginning the study of solving these differential equations.
2.2 Separable Differential Equations

This course begins with another section on separable D.E.
Definition 2.1.1 Seperable DE
A first order DE in the form:
y 0 = f (x, y)
Is seperable if and only if f (x, y) can be written as:
f (x, y) = h(x)g(y)
So the DE can be written in the form:
y 0 = h(x)g(y)
In this course we often abuse the differential notation, treating it as a frac-

tion. For our purposes you can.
To solve a separable DE, we look at the general form and solve for y(x):
y 0 = h(x)g(y)
dy
= h(x)g(y)
dx
1
dy = h(x)dx
g(y)
5
Z Z
1
dy = h(x)dx
g(y)
G(y) = H(x) + c
If you can rearrange the following equation for y then do so and you will have
an explicit formula for y(x). If you cannot, then leave the equation in the form:
G(y) − H(x) = c
P (x, y) = c
... which is a function which implicitly solves the DE. This means that the
function P (x, y) can be implicitly differentiated to show that it solves the DE.
2.5 Exact Differential Equations

Given some first order differential equation in the form:
P (x, y)
y0 =
Q(x, y)
... we can choose to write it as the following:
dy P (x, y)
=
dx M (x, y)
M (x, y)dy = P (x, y)dx

P (x, y)dx − M (x, y)dy = 0
... now let’s just make the substitution Q(x, y) = −M (x, y), to obtain the form:
P (x, y)dx + Q(x, y)dy = 0
Given a differential equation in this form, we can check if it is exact.
Definition 2.2.1 Exact DE
A first order DE in the form:
P (x, y)dx + Q(x, y)dy = 0
Is exact if and only if:

∂P ∂Q
=
∂y ∂x
6
In the case that the DE is exact, we have a method to solve for an implicit
solution.
Assuming there exists some function u(x, y) which implicitly solves the DE:
P (x, y)dx + Q(x, y)dy = 0
We can use the following theorem to find it:
Theorem 2.2.1 Exactness Solutions
If u(x, y) has continuous partial derivatives ux (x, y) and

uy (x, y) then:
u(x, y) = c
... is an implicit solution to the DE:
ux (x, y)dx + uy (x, y)dy = 0
We do have to check that the function will behave properly, and actually
be the solution to the DE that we want. Notably, the partial derivatives need
to be continuous. Recall that uxy = uyx for continuous partial derivatives, or
more illustratively:
∂ ∂
(ux ) = (uy )
∂y ∂x
Which is equivalent to our condition for exactness in the definition (recall ux =
P (x, y), uy = Q(x, y)):
∂P ∂Q
=
∂y ∂x
Now we have a new question, when solving:
P (x, y)dx + Q(x, y)dy = 0
What function u(x, y) has partial derivatives of P (x, y) and Q(x, y)... meaning
we can from the following two equations to solve for the exact solution of the
DE:
∂u
= P (x, y)
∂x
∂u
= Q(x, y)
∂y
We must now use the first equation to solve for u(x, y):
∂u
= P (x, y)
∂x
Z
u(x, y) = P (x, y)dx
7
u(x, y) = M (x, y) + h(y)
Note that the integration constant is not just c, but h(y). This is because we
are integrating with respect to x and so any function of y would not change the
derivative with respect to x of the anti-derivative.
Now note that we have an expression for u(x, y) with an unknown integration
constant and we also know that:
∂u
= Q(x, y)
∂y
Compute the left hand side from what we calculated for u(x, y) and set it equal
to Q(x, y). From this we can solve for h(y)
Finally, we can write the implicit function u(x, y) in the form:
u(x, y) = c
... by moving all the constant to the right. If we can solve for y(x) explicitly
then do so.
Method to Solve a Exact DE

Given a DE in the form:
P (x, y)dx + Q(x, y)dy = 0
1. Check for exactness:

∂P ∂Q
=
∂y ∂x
If true, then continue. If false, then try another method.
2. Write the equation:
∂u
=P
∂x
3. Integrate both sides with respect to x. You will have some equation in
the form:
u(x, y) = M (x, y) + h(y)
4. Write the equation:

∂u
=Q
∂y
∂u
5. Calculate ∂y and substitute it into the equation above.
6. Solve for h(y) by integrating h0 (y)
8
Chapter 5: Linear Second Order Equations
Theory of LDEs
Definition 5.1.1 nth Order LDE
A nth order linear differential equation (LDE) can be written

in the form:
d(n) y dn−1 y dy
an (x) + an−1 (x) + ... + a1 (x) + a0 (x)y = g(x)
dx dx dx
Characteristics which identify an LDE:
The dependent variable is always of first degree.
Coefficient functions (a0 , a1 , . . . , an ) are all functions of only the indepen-

dent variable.
Definition 5.1.2 Initial Value Problems
An initial value problem (IVP) is a differential equation

subject to:
y(x0 ) = y0 , y 0 (x0 ) = y1 , y 00 (x0 ) = y2 , ..., y (n−1) (x0 ) = yn−1
Essentially this means that we limit the solutions to the differential equation
to those that have the particular ”initial” values. You are given all the derivative
values for the ”initial” time x0 .
Usually you substitute the values you know into the DE at any point and
solve for constants with that assumption. From there you can just replace the
constance with what you solved for and continue rearranging.
9
Theorem 5.1.1 Existence and Uniqueness of IVP
Given some LDE in the form:

d(n) y dn−1 y dy
an (x) + an−1 + ... + a1 (x) + a0 (x)y = g(x)
dx dx dx
Assume an (x), an−1 (x), ..., a1 (x), a0 (x), g(x) are all continu-
ous on some interval I, and that ai (x) 6= 0 for any value of
x ∈ I. Also pick some number x0 ∈ I, then:
The LDE subject to the following initial conditions:
y(x0 ) = y0 , y 0 (x0 ) = y1 , y 00 (x0 ) = y2 , ..., y (n−1) (x0 ) = yn−1
...has a unique solution.
This means that if you have a DE where the ”coefficient functions” (the
an (x) functions) are equal to 0 at any point, then the IVP will not have a
unique solution.
Definition 5.1.3 Homogeneous DE
A LDE in the form:

d(n) y dn−1 y dy
an (x) + an−1 (x) + ... + a1 (x) + a0 (x)y = 0
dx dx dx
Is homogeneous since g(x) = 0
Definition 5.1.4 Non-Homogeneous DE
A LDE in the form:

d(n) y dn−1 y dy
an (x) +an−1 (x) +...+a1 (x) +a0 (x)y = g(x), g(x) 6= 0
dx dx dx
Is non-homogeneous since g(x) 6= 0
10
Theorem 5.1.2 Superposition of Solutions
Let y1 , y2 , ..., yk be solutions to a homogeneous LDE on I.

Then any linear combination of those solutions is also a
solution to that same homogeneous LDE on I.
Recall from linear algebra that a linear combination of things is where you
scale each thing by a real number and then add them all together.
Definition 5.1.5 Wronskian
The Wronskian (W ) of a set of functions f1 , f2 , ...fn all

with (n − 1)th order derivatives is defined as:
f1 (x) f2 (x) ··· fn (x)

f10 (x) f20 (x) ... fn0 (x)
W (f1 , f2 , . . . , fn ) = .. .. .. ..
. . . .
(n−1) (n−1) (n−1)
f1 (x) f2 (x) · · · fn (x)
Recall that the vertical bars means determinant, which works the same as the
determinant of matrix with real entries, it’s just now the entries are functions.
Recall as well that the determinant of a 2 × 2 matrix is:
a b
det(A) = = ad − bc
c d
Theorem 5.1.3 Non-Zero Wronskian
Let y1 , . . . , yn be solutions to a homogeneous DE on I. Then

the set of solutions is linearly independent on I if and only
if:
W (y1 , . . . , yn ) 6= 0
Definition 5.1.6 Fundamental Set
The fundamental set of solutions to a DE is the set of

solutions which are linearly independant.
Recall that linearly independent means none are constant multiples of each
other.
11
Theorem 5.1.4 General Solution of a HDE
If y1 , . . . , yn is a fundamental set of a homogeneous DE, then

the general solution to that DE is:
y = c1 y1 + c2 y2 + · · · + cn yn
This makes sense since we know that a linear combination of solutions to

a homogeneous is itself a solution, and since the fundamental set is linearly
independent, combining all the of the functions in a linear combination gives
you the most general form of the solution.
Theorem 5.1.5 Solutions to a Non-HDE
Given some non-homogenous DE in the form:
d(n) y dn−1 y dy
an (x) +an−1 (x) +...+a1 (x) +a0 (x)y = g(x), g(x 6= 0)
dx dx dx
We can find the general solution of it by finding the general
solution to the corresponding homogenous DE:
d(n) y dn−1 y dy
an (x) + an−1 (x) + ... + a1 (x) + a0 (x)y = 0
dx dx dx
Say y1 , . . . , yn is a fundamental set of the homogeneous
DE on I and yp is any particular solution to the non-
homogenous DE on I, then:
y = c1 y1 + c2 y2 + · · · + cn yn + yp
Is the general solution to the non-homogeneous DE on I.
What this means is that if we have a non-homogeneous LDE, then we can find
the general solution to it by finding the general solution to the homogeneous
LDE (called the complementary solution yc ) and add on any particular solution
to the non-homogeneous LDE. Every second order LDE has two families of solu-
tions, and so the general form will have two terms at least (if it is homogeneous).
12
Theorem 5.1.6 Abel’s Theorem
A second order LDE in the form:
y 00 + P (x)y 0 + Q(x)y = 0
...has two solutions families y1 , y2 . Say you found one solu-

tion family y1 through some method, then the other solution
family is:
Z − R P (x)dx
e
y2 = y1 dx
y12
The previous theorem is also called reduction of order.
Solving Second Order LDEs

If you want to solve a homogeneous LDE, then we need to begin with a simple
case, constant coefficients:
ay 00 + by 0 + cy = 0
Let’s take a guess and say y = emx is a solution to this DE. If we can solve for
m in terms of a, b, c then this really is a solution.
y = emx
y 0 = memx
y 00 = m2 emx
Substituting that into our LDE:
ay 00 + by 0 + cy = 0
a(m2 emx ) + b(memx ) + c(emx ) = 0
emx (am2 + bm + c) = 0
This tells us that:
emx = 0
am2 + bm + c = 0
The second to last equation has no solutions, the last equation has two solutions
of: √
−b ± b2 − 4ac
m1,2 =
2a
And so not only was our guess correct, but it gave us ideally BOTH of our
solutions:
ay 00 + by 0 + cy = 0 =⇒ y1 = em1 x , y2 = em2 x
There are three cases for how this can be written:
13
1. m1 6= m2 , and m1 , m2 ∈ R
In this case our general solution is just a linear combination of the two
solutions since they are linearly independent:
y = c1 em1 x + c2 em2 x
−b
2. m1 = m2 = m = 2a
In this case, the two solutions are not linearly independent and so you
have only found one of the two solution families, namely: y1 = emx . You
can find the other through reduction of order : Applied to the equivalent
DE:
b c
y 00 + y 0 + y = 0
a a
Z −R bax
e d
y2 = y1 2 dx
y1
Using the fact that:
−b −b
m= =⇒ 2m =
2a a
Z 2mx
e
y2 = emx dx
e2mx
Z
mx
y2 = e dx
y2 = xemx
Now we have two linearly independent families of solutions and so the
general solution is:
y = c1 emx + c2 xemx
3. m1 , m2 ∈ C
In this case:
m1,2 = α ± βi
These are distinct and so you could write the general solution as:
y = c1 e(α+βi)x + c2 e(α−βi)x
However this can be written in a simpler way without complex numbers

as:
y = eαx (c1 cos(βx) + c2 sin(βx))
This method extends as you’d expect to higher order LDEs.
14
Solving Non-Homogeneous LDEs with Constant Coefficients
Now that we know how to find the solution to the homogeneous DE, we know
that we can find the general solution to the non-homogeneous DE by:
y = yc + yp
As in the general solution to the non-homogeneous DE is the general solution tot

he homogeneous complementary DE, plus some particular solution to the non-
homogeneous DE. We can find a particular solution based on g(x) (the function
after the equal sign). In the following table all capital letters are unknowns to
be solved for, lower case letters are to identify form.
g(x) Guess for yp (x)
a A
ax + b Ax + B
ax2 + bx + c Ax2 + Bx + C
.. ..
. .
an xn + · · · + a1 x1 + a0 An x n + · · · + A1 x 1 + A0
sin(αx) A cos(αx) + B sin(αx)
cos(αx) A cos(αx) + B sin(αx)
eαx Aeαx
(ax + b)eαx (Ax + B)eαx
(ax+ bx + c)eαx (Ax2 + Bx + C)eαx
If g(x) can be factored into some of the forms above, then your guess would be
the product of each corresponding guess for each factor.
Make the substitution y = ”your guess” and then solve for the constants.
The only complication comes if the general solution to the homogeneous
DE has some multiple of the particular solution in it. In this case you need
to multiply the particular solution by x as many times as needed to make it
linearly independent from yc .
Additionally, if your function g(x) cannot be written in the forms mentioned
above, then you must use the following method,
Variation of Parameters
This section defines a new method to solve higher order non-homogenous LDE.
This method does not require constant coefficients, however since it does require
us to find the homogenous solution for our examples at the time we will still
have constant coefficients.
This method is particularly convenient when To find the general solution to:
y 00 + Ay 0 + By = f (x)
Write it in the form:

y = yc + yp
15
Where yc is the general solution to the related homogenous DE written as:
yc = c1 y1 + c2 y2
... then by the method of variation of parameters the particular solution
would be:
yp = u1 y1 + u2 y2
... where:
0 y2
f (x) y20
u01 =
W (y1 , y2 )
y1 0
0
y 1 f (x)
u02 =
W (y1 , y2 )
This is Cramer’s rule applied to solve the following system:
y1 u01 + y2 u02 = 0
y10 u01 + y20 u02 = 0
Chapter 8: Laplace Transforms

8.1 The Laplace Transform
This section begins with a definition of a transform:
Definition 8.1.1 Transform
A transform is a function which takes in functions as inputs,

and returns functions as output.
The derivative (also called the derivative operator) is a transform because

you can apply it to a function and you get a function back out as a result:
d
: x2 7→ 2x
dx
The Laplace Transform is another example of a transform, which is defined
by the following:
Definition 8.1.1 The Laplace Transform
The Laplace Transform of some function f (t) is:

Z ∞
L {f (t)} = F (s) = e−st f (t)dt
0
16
As you can see this fits our definition of a transform as you can put f (t) into
the transform and you get another function out of it. The Laplace Transform is
useful in solving IVPs, simplifying the process to solving an algebraic equation.
Often when you compute the Laplace Transform of a function, you need
to put a restriction on s to make the limit converge, since this is an improper
integral. Also note that the output function is a function of s since the function
gets evaluated at specific values of t.
One of the most important properties of the Laplace transform is that it is
linear, meaning:
Theorem 8.1.2 Linearity of Laplace
L {αf (t) + βg(t)} = αL {f (t)} + βL {g(t)}

... for α, β ∈ R
The following is a table of a few common Laplace Transforms which can be

derived using the definition:

1
L {1} =
s

n!
L {tn } = for n ∈ N
sn+1

1
L {eαt } =
s−α

k
L {sin(kt)} =
s2 + k2

s
L {cos(kt)} =
s2 + k 2
Next we want to understand when the Laplace transform of a function exists or
doesn’t. For that we need to the following two definitions:
17
Definition 8.1.2 Piecewise Continuous
A function is piecewise continuous on [0, ∞) if for any in-

terval you choose in [0, ∞) contains a finite amount of finite
discontinuities, and the function is continuous otherwise.
Definition 8.1.3 Exponential Order
A function f is said to be of exponential order if there

exists constants c, M > 0, T > 0 such that:
|f (t)| ≤ M ect
... for all t > T .
Essentially if your function is increasing but grows slower than an exponential

function, it is of exponential order.
Theorem 8.1.2 Existance of L {f (t)}
If f (t) is peicewise continuous on the interval [0, ∞) and of

exponential order, then L {f (t)} exists for s > c.
Notice that the previous theorem does not state that these are the only condi-
tions for the existence of the transform, only that they are sufficient to guarantee
it. If a function does not contain these properties it’s transform may or may
not exist, but this theorem tells us that if a function does have these properties,
then its transform definitely exists.
18
8.2 The Inverse Laplace Transform
Definition 8.2.1 The Inverse Laplace Transform
If:
L {f (t)} = F (s)
... then:
f (t) = L −1 {F (s)}
... where L −1 {F (s)} is the inverse Laplace transform of
F (s).
To find the inverse Laplace of a function you want to match it to the forms
of the regular Laplace transform on page 17. This can be done by algebraic
techniques like partial fraction decomposition. It is also useful to know:
Theorem 8.2.1 Linearity of the Inverse Laplace Transform
The Inverse Laplace Transform is a linear transform, mean-

ing:
L −1 {αF (s) + βG(s)} = αL −1 {F (s)} + βL −1 {G(s)}
... for α, β ∈ R
Laplace of Derivatives
The Laplace transform is useful in solving initial value problems, in this section
we will see exactly how and why.
Let’s say we want to evaluate the Laplace of the first derivative of f instead
of the function itself: Z ∞
L {f 0 (t)} = e−st f 0 (t)dt
0
... by integration by parts:
Z ∞
= [e−st f (t)]t=∞
t=0 + s e−st f (t)dt
0
= −f (0) + sL {f (t)}
Meaning:
L {f 0 (t)} = sL {f (t)} − f (0)
19
As you can see, we have related the Laplace of the derivative of a function to
the function itself, and its Laplace. Similarly we can see:
L {f 00 (t)} = s2 L {f (t)} − sf (0) − f 0 (0)
L {f 000 (t)} = s3 L {f (t)} − s2 f (0) − sf 0 (0) − f 00 (0)

Or generally:
Theorem 8.2.2 Laplace of the Derivative
If f, f 0 , . . . , f (n−1) are all continuous on [0, ∞) and are of

exponential order, and if f (n) is piecewise continuous on the
same domain, then:
L {f (n) (t)}sn F (s) − sn−1 f (0) − sn−2 f 0 (0) − · · · − f n−1 (0)
This is useful because we can turn all the derivatives of a function into the
Laplace transform minus some initial values. This, along with the linearity of
the Laplace transform make it a very effective technique to solve IVPs. The
following is a fully worked example demonstrating this method:
Example 8.2.1 Laplace to Solve an IVP
Solve the following IVP using the Laplace Transform:
y 00 − 3y 0 + 2y = e−4t , y(0) = 1, y 0 (0) = 5
Begin with your equation:

y 00 − 3y 0 + 2y = e−4t
Take the Laplace of both sides:
L {y 00 − 3y 0 + 2y} = L {e−4t }
Use the linearity of the Laplace Transform:
L {y 00 } − 3L {y 0 } + 2L {y} = L {e−4t }
Apply theorem 8.2.2 which is how the Laplace transform relates to derivatives:
s2 Y (s) − sy(0) − y 0 (0) − 3(sY (s) − y(0)) + 2Y (s) = L {e−4t }
Apply the Laplace transform to the exponential:
1
s2 Y (s) − sy(0) − y 0 (0) − 3(sY (s) − y(0)) + 2Y (s) =
s+4
20
Substitute the values we know:
1
s2 Y (s) − s − 5 − 3(sY (s) − 1) + 2Y (s) =
s+4
Solve for Y (s) and apply PFD:
s+2 −16/5 25/6 1/30

Y (s) = = + +
s2 − 3s + 2 s−1 s−2 s+4
Apply the inverse Laplace transform to both sides:
−16/5 25/6 1/30
L −1 {Y (s)} = L −1 { + + }
s−1 s−2 s+4
Left side becomes the solution to the IVP:
16 −1 1 25 1 1 1
y=− L { } + L −1 { } + L −1 { }
15 s−1 6 s−2 30 s+4
Match the forms to find the inverse:
16 t 25 2t 1
y=− e + e + e−4t
15 6 30
Which is our solution to the IVP.
The general method can be described by this graphic:
8.4 Translation Theorems

In this section we investigate the effect of multiplying the input to the transform
by two specific functions. Another way to think about this is we are investigat-
ing:
L −1 {F (s − a)}
L −1 {e−as F (s)}
These theorems can help save time instead of having to apply things like IBP
multiple times.
21
Theorem 8.4.1 First Translation Theorem
If L {f (t)} = F (s) and a ∈ R, then:
L {eat f (t)} = F (s − a)
L −1 {F (s − a)} = eat f (t)
In a sense, multiplying the input of the transform by eat horizontally shifts the
Laplace transform right by a units. The proof of this is by definition:
Z ∞ Z ∞
L {e f (t)} =
at −st at
e e f (t)dt = e−(s−a)t f (t)dt = F (s − a)
0 0
Example 8.4.1 First Translation Theorem
Compute:
L {e5t t3 }
3! 6
L {e5t t3 } = L {t3 } |s→s−5 = 4
|s→s−5 =
s (s − 5)4
Often functions are either on or off. To turn a function on you can multiply
it by 1, and to turn it off you can multiply it by 0. If we want to turn the
function f (t) on at time t = 4, then we want a function U (t) that implements
the concept: ”Output 0 until t = 4, and then output 1”... which we could
multiply by f (t), as in f (t)U (t). The function which implements this is called
the unit step function or the heaviside function, and would be written in this
example as U (t − 4).
22
Definition 8.4.1 The Unit Step Function
The unit step function U (t − a) is defined as:

(
0 0≤t<a
U (t − a) =
1 t≥a
Note that at t = a, U = 1.
The unit step function is also useful in turning piecewise functions into direct
equation form. Using the following theorems relations:
1. Two case piecewise:
(
g(t) 0 ≤ t < a
f2 (t) =
h(t) t ≥ a
f2 (t) = g(t) − g(t)U (t − a) + h(t)U (t − a)
2. Three case piecewise, first and last 0:


0

 0≤t<a
f3 (t) = g(t) a ≤ t < b

0 t≥a

f3 (t) = g(t)[U (t − a) − U (t − b)]
This brings us to our second transition theorem. While the first one was
concerned with horizontally shifting the output of the transform a units to the
right, this theorem is concerned with moving the input function to the right
by a units with the additional condition that the function should be turned off
until t = a, meaning the input would be transformed to:
f (t − a)U (t − a)
Visually that is:
23
Theorem 8.4.2 Second Translation Theorem
If L {f (t)} = F (s) and a ∈ R+ , then:
L {f (t − a)U (t − a)} = e−as F (s)
L −1 {e−as F (s)} = f (t − a)U (t − a)
Additionally these is a way to write this theorem in the a more useful form for
when f is not in the correct shifted form:
L {g(t)U (t − a)} = e−as L {g(t + a)}
The Derivative of the Laplace Transform

In this section we’d like to evaluate the derivative of the Laplace transform as
in:
d
L {f (t)}
ds
... we begin by definition:
Z ∞
d
= e−st f (t)dt
ds 0
Z ∞
∂ −st
= e f (t)dt
0 ∂s
Z ∞
=− e−st f (t)dt
0
= −L {tf (t)}
This can be generalized to:
24
Theorem 8.4.3 Derivative of Transforms
If F (s) = L {f (t)} and n = 1, 2, 3, . . . , then:
dn
L {tn f (t)} = (−1)n F (s)
dsn
This is useful in examples like the following:
Example 8.4.2 Derivative of the Laplace Transform
Compute:
L {t sin(kt)}

d d k 2ks
L {t sin(kt)} = − L {sin(kt)} = − =
ds ds s2 + k 2 (s2 + k 2 )2
8.6 Convolution
This section is motivated by a question like the following:
1 1
Evaluate: L −1 { · 2 } = L −1 {L {et }L {sin(t)}}
s−1 s +1
This is not achieved by taking the product of the arguments, but rather we need
a new operation.
In this section we define a new operation called convolution which can be
applied to two functions. Convolution is not multiplication however we do use
a ∗ to denote it.
Definition 8.5.1 Convolution
The convolution (∗) of two functions f and g is defined as:

Z t
f (t) ∗ g(t) = f (τ )g(t − τ )dτ
0
The reason this is of any use to us is because of the following theorem:
25
Theorem 8.5.1 Convolution Theorem
If f (t) and g(t) are piecwise continuous on [0, ∞) and of

exponential order, then:
L {f ∗ g} = L {f (t)}L {g(t)} = F (s)G(s)
This can be used in reverse:
f ∗ g = L −1 {F (s)G(s)}
... to answer a question like the one stated at the beginning of this section:
L −1 {L {et }L {sin(t)}} = et ∗ sin(t)
Chapter 11: Fourier Series

This chapter is all about the Fourier Series which is an analytical tool used to
help solve partial differential equations. This chapter is just the theory of the
Fourier Series and how to find the Fourier Series of a function, we will not go
into it’s applications.
The Fourier Series (FS) is a series just like the Taylor Series. It is a way to
write the function as an infinite sum. While the Taylor Series came from our
study of Power Series, the FS comes from Linear Algebra. First we recall some
of those topics.
Recall that a vector is an element of a special kind of set called a Vector
Space. One such vector spaces is Rn which is where we get our visual under-
standing of vectors in space.
Vector spaces can also have something called an inner product which is a
way to combine vectors with a certain list of special properties. The exact list
is not important at this moment, but all inner product have the same list. In
Rn the inner product is the dot product, but in other vector spaces the inner
product is different. The inner product of two vectors is denoted:
< v~1 , v~2 >
Recall that in R3 we define two vectors to be orthogonal is they have a dot

product of zero.
Now let us consider a more abstract vector space, a vector space of functions.
This vector space passes all the tests of a vector space like we learned at the end
of MTH 141. This makes functions vectors just like arrows in space. We would
like to also have an inner product (like a dot product) for this vector space:
26
Definition 11.1.1 Inner Product of Functions
The inner product of two function f1 and f2 on the interval

[a, b] is:
Z b
< f1 , f2 >= f1 (x)f2 (x)dx
a
The following list of properties is true regarding the inner product of functions
f, g, h and a constant k:
1.
< f, g >=< g, f >
2.
< kf, g >= k < f, g >
3.
< f, f >= 0
... only if f = 0, otherwise the result is positive.
4.
< f + g, w >=< f, w > + < g, w >
Definition 11.1.2 Orthogonal Functions
Two function are orthogonal on [a, b] if their inner product

is equal to 0 on [a, b], as in (for functions f , and g):
Z b
< f, g >= f1 (x)f2 (x)dx = 0
a
Definition 11.1.3 Orthogonal Sets
A set of function is said to be orthogonal on [a, b] if the inner

product between any two elements is 0 on [a, b].
A set is orthonormal if all the elements of the set also have a length of 1,
length (also called norm) is defined next:
27
Definition 11.1.4 Norm of a Vector
The norm of a vector ~v is:

p
|~v | = < ~v , ~v >
This above definition also applies to functions. A norm is like the length of a
vector, but we want to generalize this concept to things that done have a length
like functions.
More on Vectors
Remember given some vector ~v in Rn we can write ~v as a linear combination of
orthonormal basis vectors. Given the basis:
{e~1 , e~2 , · · · , e~n }
We can write any vector v as:
~v = a1 e~1 + a2 e~2 + · · · + an e~n
... all you have to do is figure out the sequence an .

Our goal in a FS is to do the exact same thing with functions. Given the
orthonormal set:
nπx nπx
{1, cos , sin }
p p
... we can show that it is both orthonormal and spans all periodic functions
with a period of p. Notice also it is an infinite set since n is some integer. Given
that this is true, this means that any such periodic function can be written as
a linear combination of the functions in the set, as in:

nπx nπx
f (x) = k(1) + an cos + bn sin
p p
... all we have to do is figure out k, an , and bn . The equation above can be
rewritten as:
∞ X ∞
X nπx nπx
f (x) = k(1) + an cos + bn sin
n=1
p n=1
p
The coefficients k, an , and bn turn out to be:
28
Definition 11.1.5 Fourier Series
The Fourier Series of a function f defined on the symmet-

rical interval (−p, p) is given by:
∞
a0 X nπ nπ
f (x) = + an cos( ) + bn sin( )
2 n=1
p p
... where:
1 p
Z
a0 = f (x)dx
p −p
1 p
Z
nπ
an = f (x) cos( x)dx
p −p p
Z p
1 nπ
bn = f (x) sin( x)dx
p −p p
Now the most important thing is that this series actually converges to the
function wherever possible. Our function f doesn’t have to be fully continu-
ous, just piecewise continuous. At the discontinuities the FS converges to the
midpoint.
Theorem 11.1.1 Convergence of the Fourier Series
If f is piecewise smooth on [−L, L] then the Fourier Series

of f :
∞
a0 X nπ nπ
F (x) = + an cos( ) + bn sin( )
2 n=1
p p
converges for all x on [−L, L], in fact it converges to exactly:


f (x)

 −L < x < L and f is continuous at x
f (x+ )+f (x− )
F (x) = 2 −L < x < L and f is discontinuous at x
 f (−L+ )+f (L− )

x = L or x = −L

2
If the function happens to be even or odd, then the formulas simplify. Recall
that a function can be:
Even: f (x) = f (−x)
Odd: f (−x) = −f (x)
Additionally there are some useful properties of even and odd functions:
29
1.
even · even = even
2.
odd · odd = even
3.
even · odd = odd
4.
even − even = even
5.
odd − odd = odd
6. If f is even: Z a Z a
f (x)dx = 2 f (x)dx
−a 0
7. If f is odd : Z a
f (x)dx = 0
−a
Definition 11.1.6 Cosine or Sine Fourier Series
The FS of an even function (f ) on (−p, p) is called the

cosine series:
∞
a0 X nπ
f (x) = + an cos x
2 n=1
p
... where:
2 p
Z
a0 = f (x)dx
p 0
2 p
Z
nπ
an = f (x) cos x dx
p 0 p
The FS of an odd function (f ) on (−p, p) is called the sine
series:
∞
X nπ
f (x) = bn sin x
n=1
p
... where: Z p
2 nπ
bn = f (x) sin x dx
p 0 p
30
Half-Range Extensions
The FS is defined on (−p, p) but in practice a lot of the time you only want
the FS of the function on (0, L). All you have to do in this case is adjust your
choice of p to use the regular FS:
Function Choice of p
Cosine Series on (0, L) p=L
Sine Series on (0, L) p=L
Fourier Series on (0, L) p = L2
There is a lot more to FS than what was discussed here, I recommend reading
12.1 − 12.3 from the textbook Advanced Engineering Mathematics and watching
Dr. Trefor Bazett on youtube to get more depth to this topic.
Chapter 1: Parametric Equations and Polar Co-

ordinates
1.1: Parametric Equations
A parametric equation is a way to write a relation between x and y as a function
of some other variable t (usually thought of as time), even if the graph of this
relation is not a function.
The idea is that every point on the graph (x, y) is assigned a particular t
value, and so we think of both coordinated as functions of t as in (x(t), y(t)).
Definition 1.1.1 Parametric Equations
If x and y are continuous functions of t on an interval I,

then the equations:
x = x(t)
y = y(t)
... are called parametric equations with t as a parameter.
The graph of parametric equations is called a parametric
curve, or plane curve C.
For example if you graph the following parametric equations:

x(t) = t2 − 3
y(t) = 2t + 1
−2 ≤ t ≤ 3
... you get the following graph:
31
Notice that we had to limit the range of t, this made it so that the graph has a
start and end point. Also notice that the graph is not a function even through
x and y are both functions of t. Finally also notice that the graph has an
orientation, as t increases the curve is drawn bottom to top.
Now we discuss converting between parametric and rectangular form.
Parametric to Rectangular
This process is called eliminating the parameter.
1. Isolate t in one or both equations.
2. Substitute one into the other.
3. Solve for y if possible.
Rectangular to Parametric
This process is called parametrization of the curve. There is an infinite number
of ways to parametrize a rectangular curve. In general the process is:
1. Choose some function f (t) and set x(t) = f (t).
(a) f (t) must have a range of f (t) ∈ R.

2. Substitute f (t) into the equation for x.
3. Solve for y(t).
32
Definition 1.1.2 Cycloids
A cycloid is a type of parametric equation in the form:
x(t) = a(t − sin(t))
y(t) = a(1 − cos(t))
Cycloids can be thought of as the curve traced out by a point on a wheel:
1.2 Calculus of Parametric Curves

We begin this section with a definition of the derivative in parametric form:
Definition 1.2.1 Parametric Derivative
Given a plane curve C defined by x = x(t) and y = y(t),

where x0 (t) and y 0 (t) both exist, the derivative is then given
by:
dy dy/dt y 0 (t)
= = 0
dx dx/dt x (t)
This will return a function of t, which will output the slope of the tangent at
any point (x(t), y(t)).
Higher order derivatives are a little harder to define, so we just need to
additionally know that the section derivative is defined as:
d2 y (d/dt)(dy/dx)
2
=
dx dx/dt
Next we deal with integration:
33
Definition 1.2.2 Parametric Integration
Given some non-self-intersecting plane curve defined by the

parametric equations
x = x(t), y = y(t), a ≤ t ≤ b
... and assuming x(t) is differentiable, the area under the

curve is given by:
Z b
A= y(t)x0 (t)dt
a
We can also apply integration to find the arc length of a curve over some
range of t:
Definition 1.2.3 Parametric Arc Length
Given some plane cirve defined by parametric equations:
x = x(t), y = y(t), t1 ≤ t ≤ t2
... and assuming that x(t) and y(t) are both differentiable,
the arc length is given by:
s
Z t2 2 2
dx dy
s= + dt
t1 dt dt
1.3 Polar Coordinates

In rectangular coordinates, we can assign an ordered pair (x, y) to every point in
the grid based on its distance along the x − axis and distance along the y − axis
respectively. Polar coordinates as an alternative way of assigning an ordered
pair to every point in the grid.
Using polar coordinates, we can convert a point (x, y) into a new ordered
pair (r, θ) using the following conversion:
34
x = r cos(θ)
y = r sin(θ)
r2 = x2 + y 2
y
tan(θ) =
x
It is good to imagine a ray coming from the origin of length r and making
an angle θ with the positive x − axis and drawing the point at it’s tip. This
should remind you of a vector, and for good reason.
To make our lives easier, we sometimes choose to label the grid as:
The origin is called the pole in this context, and the horizontal axis is the polar
axis. The innermost circle denotes a radial distance of 1. We can see a few
points plotted on the following graph:
35
We should note that the polar plane is exactly the same geometrically as the
Cartesian plane, but we are labelling things in the polar context.
Polar curves are created when we define a function in the form r = f (θ).
Meaning the length of a ray is a function of the angle it makes with the positive
x−axis. Graphing polar functions may seem confusing at first, because we have
little experience with these types of functions. As you practice you will begin
to see patterns. Imagine being asked to sketch f (x) = ex x2 without learning
what the graphs of ex or x2 are.
The following is a list of graphs and their plots.
36
37
In rectangular coordinates we considered the symmetry of a function by
calling it even and odd. We have similar definitions in polar coordinates:
38
Definition 1.3.1 Symmetry in Polar Curves
There are three types of symmetry in polar curves:

1. Symetry about the Polar axis: Either:
f (θ) = f (−θ)
Or:
f (θ) = −f (π − θ)
2. Symmetry about the Pole: Either:
f (θ) = f (θ + π)
Or:
f (θ) = −f (θ)
π
3. Symmetry Across the line θ = 2:
f (θ) = f (π − θ)
1.4 Area and Arc Length in Polar Coordinates

Given some polar plot, we may be interested in measuring the area enclosed by
some
R part of it. Just like in rectangular coordinates where we used anR integral
( dx) to solve for this area, in polar coordinates we use an integral ( dθ). The
bounds of the integral are now angles, and the represents what rays bound the
integral, as in the following diagram:
We integrate all the lengths of the rays which produce the polar plot into an
area using the following definition:
39
Definition 1.4.1 Polar Integration
Suppose f is continuous and non-negative on the interval

α ≤ θ ≤ β and that interval has a length between 0 and 2π
(inclusive of 2π). Then the area of the region bounded by
r = f (θ) between the radial lines θ = α and θ = β is:
Z β Z β
1 2 1
A= (f (θ)) dθ = r2 dθ
2 α 2 α
For the area between two curves r1 = f (θ) and r2 = g(θ) (where informally
r1 > r2 ) we use:
1 β 2
Z
A= (r − r22 )dθ
2 α 1
... and α, β represent the angles of their points of intersection.
Finally we define a formula for the arc length of a polar function:
Definition 1.4.2 Polar Arc Length
Let f be a function with a continuous derivative on an in-

terval α ≤ θ ≤ β. The arc length of the graph from θ = α
to θ = β is:
s 2
Z βp Z β
dr
L= (f (θ))2 + (f 0 (θ))2 dθ = r2 + dθ
α α dθ
Chapter 2: Vectors in Space

2.5: Equations of Lines & Planes in Space
The vector equation of a line is one way to describe a line in 3D space. For it
you need a point on the line, and a direction vector from that point, and all
other points on the line will be in the direction (or opposite) of that vector in
space.
40
In general this can be written as:
~x = P~0 + td~
... where:
1. ~x are the coordinates to any point on the line.
2. P~0 is your initial point.

3. d~ is your direction vector.
4. t is an arbitrary scaling factor.
You can then then separate the components of the vector to get the parametric
form of a line:
x = x0 + ta
y = y0 + tb
z = y0 + tc
... assuming d~ = [a, b, c] . You can then eliminate the parameter to get the
T
symmetric form of a line:

x − x0 y − y0 z − z0
= =
a b c
A plane can be written by using the general idea that all vectors in the plane
are parallel to the normal. Say you have some starting point P~0 , and a normal
~n, then the vector (P~Q) which connects the starting point to any point on the
plane Q would be orthogonal to the normal, as in:
~n · P~Q = 0
[a, b, c]T · [x − x0 , y − y0 , z − z0 ]T = 0
41
a(x − x0 ) + b(y − y0 ) + c(z − z0 ) = 0
This is called the scalar equation of the plane. Sometimes we collect constants
are rewrite it as:
ax + by + cz = ax0 + by0 + cz0
This just makes the equation a little easier to use since the coefficients on the
left are just the normal vector, and the values on the right can just be calculates
by being given a point on the plane.
If you are given three points on the plane and no normal, you can find the normal
by taking the cross product of any two of the vectors between two points, and
then using that as the normal for the above process (picking a random point
that was given to be P0 ).
2.7: Cylindrical & Spherical Coordinates

In cylindrical coordinates, a point in 3D space is given by an ordered triple
(r, θ, z) where:
1. r is the radial distance on the xy−plane.
2. θ is the angle that radial line makes with the positive x−axis.
42
3. z is the standard coordinate off the xy−plane.
It is called cylindrical coordinates because given some radial distance, you con-
struct a cylinder until you know what θ and z are:
Given some rectangular coordinate (x, y, z), and the cylindrical coordinates of
the same point (r, θ, z), then they are related in the following equations:
Cylindrical to Rectangular
x = r cos(θ)
y = r sin(θ)
z=z
Rectangular to Cylindrical
r 2 = x2 + y 2
y
tan(θ) =
x
z=z
In spherical coordinates, a point in 3D space is given by an ordered triple

(ρ, θ, φ).
43
Where:
ρ is the distance between the point and the origin.
θ is the angle the radial line makes with the positive x−axis.
φ is the angle formed by the positive z−axis and the line segment OP
¯ .
Given some point (ρ, θ, φ) it can be related to Rectangular and Cylindrical

coordinates using the following:
Spherical to Rectangular
x = ρ sin(φ) cos(θ)
y = ρ sin(φ) sin(θ)
z = ρ cos(φ)
Rectangular to Spherical
ρ2 = x2 + y 2 + z 2
y
tan(θ) =
x
!
z
φ = arccos p
x2 + y 2 + z 2
44
Spherical to Cylindrical
r = ρ sin(φ)
θ=θ
z = ρ cos(φ)
Cylindrical to Spherical
p
ρ= r2 + z 2
θ=θ

z
φ = arccos √
r2 + z 2
Chapter 3: Vector-Valued Functions

3.1 Vector-Valued Functions (VVFs)
Definition 3.1.1 Vector-Valued Functions
A vector-valued function is a function in the form:

 
f (t)
~r(t) =  g(t) 
h(t)
Where each f, g, h are parametric functions of t. Plotting

VVFs is very similar to plotting polar functions.
45
Definition 3.1.2 Vector-Valued Limits
~ as t
A vector valued function ~r(t) approaches the limit L
approaches a,
lim ~r(t) = L
t→a
... given that:

~ =0
lim ~r(t) − L
t→a
Theorem 3.1.1 Calculating the Limit of a VVF
Let f , g, and h be functions of t. Then:

 
limt→a f (t)
lim ~r(t) =  limt→a g(t) 
t→a
limt→a h(t)
Provided all the limits exist.
We can then define the continuity of a vector valued function:
46
Definition 3.1.3 Vector-Valued Continuity
A vector valued function ~r(t) is continuous at t = a if the

following three conditions hold:
1.
~r(a) exists.
2.
lim ~r(t) exists.
t→a
3.
lim ~r(t) = ~r(a)
t→a
3.2 Calculus of VVFs

We begin with the formal definition of the derivative of a VVF, and then the
method we actually use to compute the derivatives:
Definition 3.2.1 VVF Derivative
The derivative of ~r(t) is:
d~r ~r(t + ∆t) − ~r(t)

~r 0 (t) = = lim
dt ∆t→0 ∆t
We however never actually compute the derivative using this method, instead
we use the following theorem:
Theorem 3.2.1 Calculating the Derivative of a VVF
Let f , g, and h be differentiable functions of t. Then:

 0 
f (t)
d~r  0 
~r 0 (t) = = g (t)
dt
h0 (t)
In general, the derivative operator works the same as with regular non-vector
valued functions. The following is an explicit list of rules:
1.
d d
[c~r(t)] = c [~r(t)]
dt dt
47
2.
d d~r(t) d~u(t)
[~r(t) ± ~u(t)] = ±
dt dt dt
3.
d d~u(t)
[f (t)~u(t)] = f 0 (t)~u(t) + f (t)
dt dt
4.
d d~r(t) d~u(t)
[~r(t) · ~u(t)] = · ~u(t) + ~r(t) ·
dt dt dt
5.
d d~r(t) d~u(t)
[~r(t) × ~u(t)] = × ~u(t) + ~r(t) ×
dt dt dt
6.
d d~r(f (t)) 0
[~r(f (t))] = · f (t)
dt dt
Finally a very important rule about VVF derivatives is that the derivative
VVF is orthogonal to the original VVF at every point (if the VVF is of constant
magnitude), since:
2
~r(t) · ~r(t) = ~r(t)
d d 2
[~r(t) · ~r(t)] = [ ~r(t) ]
dt dt
d~r(t) d~r(t)
· ~r(t) + ~r(t) · =0
dt dt
d~r(t)
2~r(t) · =0
dt
d~r(t)
~r(t) · =0
dt
Remember that if two vector have a dot product of 0 they are orthogonal.
Tangent Vectors
Given that we know the vector ~r(t0 ) and ~r 0 (t0 ) are perpandicular, the derivative
vector is tangent to the curve traced out by ~r(t) at t = t0 . We define the
normalized version of this derivative vector as the tangent vector (T~ (t)):
Definition 3.2.2 Tangent Vectors
The principal unit tangent vector of ~r(t) is given by:
~r 0 (t)
T~ (t) =
~r 0 (t)
48
Finally, we finish this section with integration of VVFs, which works how
you would expect:
Definition 3.2.3 VVF Integral
Let f , g, and h be differentiable functions of t. Then:

R 
R f (t)dt
Z
~r(t)dt = R g(t)dt 
h(t)dt
3.3 Arc Length and Curvature

Recall that informally, arc length is the length of the curve between two points.
We can compute the arc length using the following formula:
Definition 3.3.1 Arc Length
Let f , g, and h be differentiable functions of t. Then arc

length (s) of a VV curve C between t values of a and b is:
Z b p Z b
s= (f 0 (t))2 + (g 0 (t))2 + (h0 (t))2 dt = ~r 0 (t) dt
a a
We can then consider what happens if we make the upper limit of this value
a function of t: Z t
s(t) = ~r 0 (u) du
a
This would result in s being a function of t. This is a useful function to have,
as it returns the arc length of the curve from some starting value t = a to the t
value given as input to the function.
We can take this a step further and rearrange the function s(t) for t = t(s),
and then replacing t with t(s) in our function definition:
   
f (t) f (s)
~r(t) =  g(t)  =⇒ ~r(s) =  g(s) 
h(t) h(s)
This is called the arc-length parametrization of the curve C. This is

useful because now the parameter has more meaning, ~r(5) now returns the
point on the curve which is 5 units along the curve away from t = a.
49
Curvature
If you look at the following graph, you can see the tangent vectors change as
you move along the curve. The tangent vectors change a lot when the curvature
is high, and the tangent vectors don’t change much when the curvature is low.
This is how we define curvature, as how fast the tangent vectors change:
Definition 3.3.2 Curvature
Let ~r(s) be the arc length parameterization of the curve C.

The curvature κ at s is:
dT~
κ=
ds
There are some alternative formulas for curvature:
T~ 0 (t)
κ=
~r 0 (t)
For a 3D curve:
~r 0 (t) × ~r 00 (t)
κ= 3
~r 0 (t)
For a function y = f (x)
y 00
κ= 3
(1 + (y 0 )2 ) 2
We now define the normal and binormal vectors of the VVF, similar to how
we defined the tangent vector:
50
Definition 3.3.3 Normal and Binormal Vectors
Let C be a three-dimensional smooth curve represented

by ~r(t) over an open interval I. The principal unit normal
vector is defined to be:
~ 0
~ (t) = T (t)
N
T~ 0 (t)
The binormal vector at t is defined as:

~
B(t) = T~ (t) × N
~ (t)
At every point ~r(t0 ) on the curve C, the tangent, normal, and binormal
vectors form a right handed system centred at ~r(t0 ).
Another good technique to know is given some implicitly defined graph

(G(x, y) = c) , you can determine the tangent and normal to the curve by
using the following technique:

1
T~ = 0
y (x0 )
0
~ −y (x0 )
N=
1
Chapter 4: Differentiation of Functions of Several

Variables
4.4 Tangent Planes and Linear Approximations
The tangent plane to a curve at a point P0 is the plane which contains all the
tangent lines to the curve at that point. We can obtain the equation of the
tangent plane by the following formula:
51
Definition 4.4.1 Tangent Planes
Let S be the surface defined by a differentiable function

z = f (x, y), and let P0 = (x0 , y0 ) be a point in the domain
of f . Then the equation of the tangent plane to S at P0 is
given by:
z = f (x0 , y0 ) + fx (x0 , y0 )(x − x0 ) + fy (x0 , y0 )(y − y0 )
4.6 Directional Derivatives and the Gradient

The gradient is a mathematical tool used to represent more complex ideas. The
gradient is an operator which you can apply to a function, and returns a vector
in which each component is the partial derivative of that function with respect
to that component, as in:
52
Definition 4.6.1 Gradient
Given a function f (x1 , x2 , . . . , xn ) such that fx1 , fx2 , . . . fxn

exist, the gradient of f , denoted:
∇f (x1 , x2 , . . . , xn ) = grad f
... is defined as:

 
fx1
 fx2 
 
∇f (x1 , x2 , . . . , xn ) = 
 .. 

 . 
fxn
The gradient vector at any point on the surface is always normal to the tangent
plane at that point.
Directional Derivatives
The partial derivative with respect to x can be interpreted as the rate of change
of the function after taking one step in the î direction. Similarly, the partial
derivative with respect to y can be interpreted as the rate of change of the
function after taking one step in the ĵ direction.
But what if we want to know the rate of change of the function after taking a
step in the direction of an arbitrary unit vector û? This is called the directional
derivative.
53
Definition 4.6.2 Directional Derivative
The directional derivative of z = f (x, y) in the directional

of a unit vector defined as:

cos(θ)
û =
sin(θ)
... is defined as:

f (x + h cos(θ), y + h sin(θ)) − f (x, y)
Dû f (x, y) = lim
h→0 h
Notice that this is really generalizing partial differentiation, as if you set θ = 0

∂z
you get the definition of ∂x , and if you set θ = π2 you get the definition of ∂y
∂z
.
There is however an easier way to compute the directional derivative:
Theorem 4.6.1 Computing the Directional Derivative
Let f (x, y, z) be a differentiable fnction of three variables,

and let û be a unit vector. Then the directioanl derivative
of f in the direction of û is:
Dû f (x, y, z) = ∇f (x, y, z) · û
Properties
The following is a list of some notable properties of the gradient and the direc-
tional derivative:
If ∇f (x0 , y0 ) = 0, then Dû f (x0 , y0 ) = 0 for any û.
If ∇f (x0 , y0 ) 6= 0, then Dû (x0 , y0 ) is maximized when û points in the same

direction as ∇f (x0 , y0 ), and the maximum value is ∇f (x0 , y0 ) .
If ∇f (x0 , y0 ) 6= 0, then Dû (x0 , y0 ) is minimized when û points in the op-
posite direction as ∇f (x0 , y0 ), and the minimum value is − ∇f (x0 , y0 ) .
54
Chapter 5: Multiple Integration
In this section we study what integration means for a multi variable function.
5.1 Double Integral over Rectangular Coordinates

In single-variable integration, we took the value of the function for every value
between a and b, added them all together and called it:
Z b
f (x)
a
You can think of this as integrating on the line y = 0. We will study how
to integrate on other lines, but first we will talk about integrating over areas
instead.
In single variable functions, we divided the interval a to b into n subregions,
evaluated the function at an arbitrary point in the subregion, and then added
all those values together. We do a similar thing with integration of 2 variable
functions but this time it is into regions Rk , as in:
55
The area of each of those squares is ∆x · ∆y = ∆A. The double Riemann
sum definition of the double integral is as follows:
Definition 5.1.1 Double Integral
The double integral of the function f (x, y) over the rect-

angular region R in the xy−plane is defined as:
ZZ X n
m X
f (x, y)dA = lim f (x∗ij , yij
∗
)∆A
R m,n→∞
i=1 j=1
Notice that we do not put the definite bounds on the integral even though
we are defining a definite integral, this is because we are not evaluating this
function over an interval, we are evaluating this function for every point in a
2D space called R on the input plane.
The following is a list of important properties of the double integral. Assume
that f (x, y) and g(x, y) are integrable over the rectangular region R, S and T
are subregions of R, and that m and M are real numbers:
The sum f (x, y) + g(x, y) is integrable as:
ZZ ZZ ZZ
f (x, y) + g(x, y)dA = f (x, y)dA + g(x, y)dA
R R R
If c is a constant, then cf (x, y) is integrable as:

ZZ ZZ
cf (x, y)dA = c f (x, y)dA
R R
56
If R = S ∪ T and S ∩ T = ∅ except an overlap on the boundaries, then:
ZZ ZZ ZZ
= f (x, y)dA = f (x, y)dA + f (x, y)dA
R S T
If f (x, y) ≥ g(x, y) for all (x, y) ∈ R, then:

ZZ ZZ
f (x, y)dA ≥ g(x, y)dA
R R
If m ≤ f (x, y) ≤ M for all (x, y) ∈ R, then:

ZZ
m · (Area of R) ≤ f (x, y)dA ≤ M · (Area of R)
R
If f (x, y) can be factored into a product g(x)h(y), then the definite double
integral over the rectangular region R = [a, b] × [c, d] is:
! Z !
ZZ Z b d
f (x, y)dA = g(x)dx h(y)dy
R a c
Evaluating Double Integrals over a Rectangular Region

Over a rectangular, it is pretty straightforward to evaluate using the following
theorem:
Theorem 5.1.1 Fubini’s Theorem
Suppose that f (x, y) is continuous over a rectangular region

R = [a, b] × [c, d]. Then then the double integral is equal to
either of the following:
ZZ ZZ Z b Z b
f (x, y)dA = f (x, y)dx dy = f (x, y)dx dy
R R c a
ZZ ZZ Z b Z d
f (x, y)dA = f (x, y)dy dx = f (x, y)dy dx
R R a c
This is illustrated by the following image:
57
So far we only know to evaluate double integrals over rectangular regions in the
input space. In the next section we will learn how to evaluate them over more
complex regions.
5.2 Double Integral over General Regions

Before we get to completely general regions of the input plane to integrate over,
we study two simpler cases called Type I and Type II regions:
Definition 5.1.3 Type I and Type II Regions
A region D in the plane is of Type I if it lies between two

vertical lines, and the graphs of two continuous functions
g1 (x) and g2 (x), that is:
D = {(x, y) | a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x)}
A region D in the plane is of Type II if it lies between two

horizontal lines, and the graphs of two continuous functions
h1 (x) and h2 (x), that is:
D = {(x, y) | h1 (y) ≤ x ≤ h2 (y), c ≤ y ≤ d}
Examples of each are as follows:
58
If we are integrating over either Type I or II regions then we have a direct
formula we can use:
Theorem 5.2.1 Fubini’s Theorem (Strong Form)
For a function f (x, y) that is continuous on a region D of

Type I, then:
ZZ ZZ Z b Z g2 (x)
f (x, y)dA = f (x, y)dy dx = f (x, y)dy dx
D D a g1 (x)
For a function f (x, y) that is continuous on a region D of

Type II, then:
ZZ ZZ Z d Z h2 (x)
f (x, y)dA = f (x, y)dx dy = f (x, y)dx dy
D D c h1 (x)
The idea of this theorem is to call the inner integral a slice of the volume
called A(x) and integrate that between the two constant bounds. The geometric
interpretation for a Type I region is:
59
... additionally, it is effective to imagine the first integral integrating the
function across a line in the x or y direction (depending on type I or type II)
and then integrating that line between the constant bounds, as in:
From here, if you have a particularly complex region to integrate over, you
just need to decompose it into regions you know how to integrate over and use
the following theorem to add them back together:
60
Theorem 5.2.2 Decomposing Regions
Suppose a region D can be expressed as D = D1 ∪ D2 where

D1 and D2 do not overlap except at boundaries, then:
ZZ ZZ ZZ
f (x, y)dA = f (x, y)dA + f (x, y)dA
D D1 D2
Finally, you can reverse the order of integration by changing the way you think
about the region. A Type I region is Type II if you switch the independent and
dependent variables around, for example:
Which shows us that:

√ √
Z 2 Z 2−x2 Z 2 Z 2−y
x2 2
xe dy dx = xex dx dy
0 0 0 0
5.3 Double Integral in Polar Coordinates

Often it is easier to evaluate a double integral in polar coordinated. We want
to be able to convert a rectangular double integral to a polar double integral.
First we must define polar double integration.
In rectangular double integrals, we subdivided the region of the domain into
subregions which were all rectangles, here we will instead need to use polar
rectangles which are produced by varying θ from α to β, and r from a to b, as
in:
R = {(r, θ) | a ≤ r ≤ b, α ≤ θ ≤ β}
We then subdivide, find the area of each smaller region, evaluate the function
at an arbitrary point in the region:
61
Definition 5.3.1 Polar Double Integral
The double integral of the function f (r, θ) over the polar

rectangular region R in the rθ−plane is defined as:
ZZ X n
m X
∗ ∗
f (r, θ)dA = lim f (rij , θij )∆A
R m,n→∞
i=1 j=1
Note also that:

∗
∆A = rij ∆r∆θ
We can evaluate the polar integral over polar rectangular regions similar
to how we do it for rectangular regions... given some polar rectangular region
R:
R = {(r, θ) | a ≤ r ≤ b, α ≤ θ ≤ β}
... then:
ZZ ZZ Z β Z b
f (r, θ)dA = f (r, θ)r dr dθ = f (r, θ)r dr dθ
R R α a
To convert a function to a polar double integral from rectangular, we use

the following formula:
ZZ ZZ
f (x, y)dA = f (r cos(θ), r sin(θ))r dr dθ
R R
This is achieved by making the following substitutions:
x = r cos(θ)
y = r sin(θ)
dA = r dr dθ
62
Other Types of Polar Regions
Now we want to be able to integrate over a more complex polar region. Similar
to Type I and Type II in rectangular coordinates, in polar we have one type of
region which is simple to evaluate over, in the form of:
Which can be written as:
D = {(r, θ) | h1 (θ) ≤ r ≤ h2 (θ), α ≤ θ ≤ β}
Then:
Theorem 5.3.1 Polar Double Integrals
If f (r, θ) is continuous on a general polar region D as de-

scribed above, then:
ZZ Z β Z h2 (θ)
f (r, θ)r dr dθ = f (r, θ)r dr dθ
D α h1 (θ)
5.4 Triple Integrals

In this section we define the triple integral, which is how you integrate functions
of three variables. Visually you can imagine only the function’s input space
being some volume in R3 . From there, we want to divide this region into sub-
boxes with dimensions ∆x × ∆y × ∆z.
63
Through the same process as before, of making sub-regions, taking an arbitrary
point in each subregion to evaluate the function at, adding all those function
values together and then taking the limit as the size of those subregions goes to
zero, we get the formal definition of the triple integral:
Definition 5.4.1 Triple Integral
The triple integral of the function f (x, y, z) over the rect-

angular box B in the xyz−space is defined as:
ZZZ l X
X m X
n
f (x, y, z)dV = lim f (x∗ij , yij
∗ ∗
, zij )∆x∆y∆z
B l,m,n→∞
i=1 j=1 k=1
All the properties of the double integral also apply to the triple integral, so
they can be assumed.
You can imagine a triple integral begin what you get when you evaluate a
function at every point inside a region of R3 .
64
We begin the study by learning how to evaluate triple integrals over a rect-
angular box B given by:
B = [a, b] × [c, d] × [e, f ] = {(x, y, z) ∈ R3 | a ≤ x ≤ b, c ≤ y ≤ d, e ≤ z ≤ f }
Theorem 5.4.1 Fibini’s Theorem for Triple Integrals
If f (x, y, z) is continuous on a rectangular box B as de-

scribed above, then:
ZZZ Z f Z d Z b
f (x, y, z)dV = f (x, y, z)dx dy dz
B e c a
... or any of the 5 other permutations of dx dy dz.
Now we want evaluate a triple integral over more complex volumes, we begin
by looking at a volume which is between two curves, over some region D on the
xy−plane, as in:
We can think of D as the projection of the volume onto the xy−plane. We then
use the following theorem to evaluate a triple integral over this kind of space.
65
Theorem 5.4.2 Triple Integrals
The triple integral of a continuous function f (x, y, z) over

a general three-dimensional region E whch is overtop some
region D on the xy−plane between two surfaces u1 (x, y) and
u2 (x, y), as in:
E = {(x, y, z) | (x, y) ∈ D, u1 (x, y) ≤ z ≤ u2 (x, y)}
... is:
ZZZ ZZ Z u2 (x,y)
f (x, y, z)dV = f (x, y, z)dxdA
E D u1 (x,y)
If that region D is of Type I as in:
E = {(x, y, z) | a ≤ x ≤ b, g1 (x) ≤ y ≤ g2 (x), u1 (x, y) ≤ z ≤ u2 (x, y)}

Then the integral becomes:
ZZZ Z b Z g2 (x) Z u2 (x,y)
f (x, y, z)dV = f (x, y, z)dz dy dx
E a g1 (x) u1 (x,y)
If that region D is of Type II as in:
66
E = {(x, y, z) | c ≤ y ≤ d, h1 (y) ≤ x ≤ h2 (y), u1 (x, y) ≤ z ≤ u2 (x, y)}
Then the integral becomes:
ZZZ Z d Z h2 (y) Z u2 (x,y)
f (x, y, z)dV = f (x, y, z)dz dx dy
E c h1 (y) u1 (x,y)
Recall that there is always two ways to write the double integral, and so we
can change the variables as well in a triple integral in the same way.
5.5 Triple Integrals in Cylindrical and Spherical Coordi-

nates
In this section we define how we compute triple integrals in cylindrical (r, θ, z)
and spherical (r, θ, φ) coordinates.
Cylindrical Triple Integrals

Integrating in cylindrical coordinates is usefor for solids that are symmetric
around an axis, such as cylinders or cones.
If we define some box B in cylindrical coordinates:
B = {(r, θ, z) | a ≤ r ≤ b, α ≤ θ ≤ β, c ≤ z ≤ d}
Then we want to divide this box into sub-boxes, evaluate the function in each
box, sum them all together and then take the limit as the size of the boxes goes
to zero.
67
Definition 5.5.1 Cylindrical Triple Integral
Considering a cylindrical box B as described above, if the

function f (r, θ, z) on B then we can define the triple integral
in cylindrical coordinates as the limit of a triple Riemann
sum: ZZZ
f (r, θ, z)dV =
B
l X
X m X
n
∗ ∗ ∗ ∗
lim f (rijk , θijk , zijk )rijk ∆r∆θ∆z
l,m,n→∞
i=1 j=1 k=1
It is worth noting:
ZZZ ZZZ
g(x, y, z)dV = g(r cos(θ), r sin(θ), z)r dr dθ dz
B B
The following theorem is as you’d expect for integration of cylindrical coor-

dinates:
68
Theorem 5.5.1 Cylindrical Triple Integrals
Suppose that g(x, y, z) is continuous on a rectangular box

B as described by:
B = {(r, θ, z) | a ≤ r ≤ b, α ≤ θ ≤ β, c ≤ z ≤ d}
... then:
ZZZ Z d Z β Z b
g(x, y, z)dV = g(r cos(θ), r sin(θ), z)r dr dθ dz
B c α a
Spherical Triple Integrals

Integrating in cylindrical coordinates is use for for solids that are symmetric a
point, such as spheres or cones.
If we define some box B in spherical coordinates:
B = {(ρ, θ, φ) | a ≤ ρ ≤ b, α ≤ θ ≤ β, γ ≤ φ ≤ λ}
Then we want to divide this box into sub-boxes, evaluate the function in each
box, sum them all together and then take the limit as the size of the boxes goes
to zero.
69
Definition 5.5.2 Spherical Triple Integral
Considering a spherical box B as described above, if the

function f (ρ, θ, φ) on B then we can define the triple integral
in spherical coordinates as the limit of a triple Riemann sum:
ZZZ
f (ρ, θ, φ)dV =
B
l X
X m X
n
lim f (ρ∗ijk , θijk
∗
, φ∗ijk )(ρ∗ijk )2 sin(φ)∆ρ∆θ∆φ
l,m,n→∞
i=1 j=1 k=1
It is worth noting that:
dV = ρ2 sin(φ)dρ dθ dφ
The following theorem is as you’d expect for integration of spherical coordi-

nates:
Theorem 5.5.2 Spherical Triple Integrals
Suppose that g(ρ, θ, φ) is continuous on a rectangular box

B as described by:
B = {(ρ, θ, φ) | a ≤ ρ ≤ b, α ≤ θ ≤ β, γ ≤ φ ≤ λ}
... then:
ZZZ Z λ Z β Z b
g(ρ, θ, φ)dV = g(ρ, θ, φ)ρ2 sin(φ)dρ dφ dθ
B γ α a
Chapter 6: Vector Calculus

6.1 Vector Fields
A vector field is a way to assign every point in space with a vector. The following
are some examples:
70
Definition 6.1.1 Vector Fields
A vector field F~ in R2 is an assignment of a 2−dimensional

vector F~ (x, y) to each point (x, y) of a subset D of R2 . You
can think of a vector field in R2 as:
F~ R2 → R2
... as in a function with two inputs and two outputs.

A vector field F~ in R3 is an assignment of a 3−dimensional
vector F~ (x, y, z) to each point (x, y, z) of a subset D of R3 .
You can think of a vector field in R3 as:
F~ R3 → R3
... as in a function with three inputs and three outputs.
71
Typically, we write them as vectors:
 
P (x, y, z)
F~ (x, y, z) = Q(x, y, z)
R(x, y, z)
... additionally given some vector field, we can normalize all its components to
get a unit vector field, as in:
   
P P
F~ = Q → U ~ = 1 Q
R F~ R
We now discuss gradient fields, which are a special type of vector field, also
called conservative fields. Recall the gradient of a function is a vector assigned
to every point in it’s domain, the gradient is a vector field. If a vector field is
the gradient field of some function f , then it is a gradient field.
Definition 6.1.2 Gradient Fields
A vector field F~ is a gradient field if there exists some scalar

function f such that ∇f = F~ . We call f a potential func-
tion for F~ .
This should remind you of integration, if you ”indefinitely integrate” the

vector field, you get a potential function whose gradient field is the vector field
in question. But if potential functions differ by a constant, they would have the
same gradient field, which brings us to the next theorem:
Theorem 6.1.1 Potential Functions
Let F~ be a conservative vector field on an open and con-

nected domain. Let f and g be functions such that:
∇f = F~
∇g = F~
... then there is a constant C such that:
f =g+C
Given some arbitrary vector field, it is more likely not conservative. This is
because a conservative vector field must pass the following test. It should be
noted that the following test can only be used to prove that a vector field is not
conservative, and not that it is.
72
Theorem 6.1.2 Cross-Partial Property
Let F~ be a vector field with components who have continu-

ous second-order
mixed-partial derivatives on its domain.
P (x, y)
If F~ = is conservative then:
Q(x, y)
∂P ∂Q
=
∂y ∂x
 
P (x, y, z)
If F~ = Q(x, y, z) is conservative then:
R(x, y, z)
∂P ∂Q ∂Q ∂R ∂R ∂P
= , = , and =
∂y ∂x ∂z ∂y ∂x ∂z
6.2 Line Integrals

We would now like to learn how to integrate in a vector field. There are multiple
different kinds of integrals of vector fields. In this section we begin discussing
the line integral. First we learn what it means to take the line integral of a
regular function (called a scalar line integral ) and then we learn how to take
the line integral of a vector field (called a vector line integral ).
Scalar Line Integrals

The notation: Z b
f (x)dx
a
... means we want to integrate the function along the x − axis from a to b.
If we have a function of 2 variables, you can imagine drawing any arbitrary
curve on the domain plane, call this curve C. We know how to integrate all the
values of a function on the 2D input plane, but what if we just want to integrate
the function along the curve C, this can be visualized as below:
73
... and we denote this: Z
f (x, y) ds
C
We construct the line integral by breaking up the curve into a bunch of
straight line segments which have constant ∆t.
Note that in the picture Pi = ~r(ti ) You can then approximate the line integral
by adding up rectangles with height f (~r(t0 )) and width ∆s.
Definition 6.2.1 Scalar Line Integral
Let f be a function whose domain contains a curve C

parametrized by the function ~r(t), a ≤ t ≤ b. The scalar
line integral of f along C is defined as:
Z
f ds = lim f (~r(t∗ ))∆si
C n→∞
74
We can simplify this process by changing the integration variable to t, by
recalling (for a 3-variable function):
ds p
= ~r 0 (t) = (x0 (t))2 + y 0 (t))2 + z 0 (t))2
dt
... meaning: p
ds = (x0 (t))2 + y 0 (t))2 + z 0 (t))2 dt
... which brings us to the next theorem:
Theorem 6.2.1 Evaluating a Scalar Line Integral
Let f be a continuous function with a domain that includes

C parameterized by given by (in 3D):
 
x(t)
~r(t) = y(t)
z(t)
... a ≤ t ≤ b. Then:
Z Z b p
f (x, y, z)ds = f (~r(t)) (x0 (t))2 + y 0 (t))2 + z 0 (t))2 dt
c a
Note that parameterization of the curve C has no effect on the line integral as
long as the curve is traversed once as a ≤ t ≤ b.
Additionally, we can apply this idea to define arc length of a curve C to be
just the integral of 1 as t varies, as in:
Z
Arc Length of C = ds
C
Vector Line Integrals

We now want to integrate the vector field along some curve C. The meaning of
this is we want a way to measure how much the vector field is aligned with the
curve, and if the vector field represents a force field, how much work the force
~
field is doing on an object along that path (Recall work = F~ · d).
Imagine taking a curve in a vector field. At every point on the curve there
would be a corresponding vector from the vector field, and also a tangent vector
of the curve, as in:
75
If we take the dot product of those two vectors, we get some kind of measure
of how aligned those two vectors are. If we do this for ever point along the
arclength of the curve, we get the following definition of the vector line integral:
Definition 6.2.2 Vector Line Integral
The vector line integral of a vector field F~ along a curve C

parameterized by ~r(t) is:
Z
F~ · T~ ds = lim F~ (~r(t∗ )) · T~ (~r(t∗ ))∆si
C n→∞
We want a way to evaluate these vector line integrals, and so we want to

express it in terms of t. Recall that:
~r 0 (t)
T~ =
~r 0 (t)
ds = ~r 0 (t) dt
We can combine these facts into:
~r 0 (t)
F~ · T~ ds = F~ · ~r 0 (t) dt = F~ (~r(t)) · ~r 0 (t)dt
~r 0 (t)
... and thus: Z Z b Z

F~ · T~ = F~ (~r(t)) · ~r 0 (t)dt = F~ · d~r
C a C
... note at the end we introduce a new notation which reflects the fact that to
do the line integral you only need information about F~ and the derivative of ~r:
 0 
x (t)
d~r = y 0 (t)
z 0 (t)
76
Another standard notation for line integrals, given that the vector field is in the
form:  
P (x, y, z)
F~ = Q(x, y, z)
R(x, y, z)
... is:
Z Z Z b
~ dx dy dz
F ·d~r = P dx+Qdy+Rdz = P (~r(t)) + Q(~r(t)) + R(~r(t)) dt
C C a dt dt dt
With scalar line integrals, neither the parameterization nor direction of curve
matters. With vector line integrals the orientation of the curve matters, as in
if you go from left to right on a curve, you would get a different vector line
integral compared to if you integrated the other way. This make sense, as the
work done by gravity by moving up a fill is exactly negative the work done by
gravity moving down a hill.
The opposite orientation of a curve C is denoted −C, and so:

Z Z
F~ · d~r = − F~ · d~r
−C C
The following is a list of properties the vector line integral has:
77
Theorem 6.2.2 Evaluating a Scalar Line Integral
Let F~ and G
~ be continuous vector fields with domains that
include the oriented smooth curve C. Then:
1. Z Z Z
~ ~
(F + G) · d~r = ~
F · d~r + ~ · d~r
G
C C C
2. Z Z
k F~ · d~r = k F~ · d~r, where k ∈ R.
C C
3. Z Z
F~ · d~r = − F~ · r~r
−C C
4. Suppose instead that C is a piece wise smooth curve

in the domain of F~ , where C = C1 + C2 + · · · + Cn ,
then:
Z Z Z Z
F~ · d~r = F~ · d~r + F~ · d~r + · · · + F~ · d~r
C C1 C2 Cn
Flux and Circulation

Flux is the measure of flow of the vector field across some curve C. In a sense, we
want to know for every vector on the line in the vector field, what component of
that vector is normal to the curve at that point, which leads us to the definition
of flux:
Definition 6.2.3 Flux
The flux of F~ across C is:

Z Z b
~ ~n(t)
F· ds = F~ (~r(t)) · ~n(t)dt
C ~n(t) a
78
Note that for a curve parameterized by:

x(t)
~r(t) =
y(t)
... the normal vector at any point is:
y 0 (t)

~n(t) =
−x0 (t)
Circulation of F~ along a curve C measures the tendency of a fluid to move

in the direction of the curve where C is a closed curve. The only difference is
that C is a closed curve and so it forms a loop.
We use special notation for a circulation integral:
I
F~ · T~ ds
C
... which can then be evaluated as a regular line integral.
6.3 Conservative Vector Fields

In this section we continue our theoretical study of vector fields. Just like
in single variable calculus, if a function is integrable on an interval, we have
the fundamental theorem of calculus. There is a similar concept which can be
applied to conservative vector fields, namely the Fundamental Theorem of Line
Integrals.
Before we continue, we must define some concepts relating to curves and
regions:
79
Definition 6.3.1 Curve Definitions
A curve C is a closed curve if there is a paramterization

~r(t), a ≤ t ≤ b of C such that the parameterization traverses
the curve exactly once and ~r(a) = ~r(b).
A curve C is a simple curve if it does not cross itself.
Meaning the parameterization is one-to-one on (a, b).
Definition 6.3.2 Region Definitions
A region D is a connected region if, for any two points,

there is a path between them entirely inside D.
A region D is a simply connected region if any closed
curve C entirely within D can be shrunken continuously to
a point while staying entirely inside D. In two dimensions
this means there are no holes.
80
If ∇f = F~ is a gradient field, and we think of ∇ as a derivative, then f can be
thought of as an anti-derivative of F~ , also called the potential function. So just
like in single variable integrals where we can evaluate the whole thing using only
the value of the anti-derivative at the end points, if the anti-derivative exists,
we can do the exact same thing with line integrals in conservative fields.
Theorem 6.3.1 Fundamental Theorem for Line Integrals
Let C be a piecewise smooth curve with parameterization

~r(t), a ≤ t ≤ b. Let f be a function with first order partial
derivative that exist and are continuous on C. Then,
Z
∇f · d~r = f (~r(b)) − f (~r(a))
C
Proof: We already know that:

Z Z b
∇f · d~r = ∇f (~r(t)) · ~r 0 (t)dt
C a
... and since (using the chain rule):

d
f (~r(t)) = ∇f (~r(t)) · ~r 0 (t)
dt
81
... we can determine that:
Z Z b
d
∇f · d~r = f (~r(t))dt
C a dt
... then by the fundamental theorem of calculus:
Z
∇f · d~r = f (~r(t)) |ba
C
Z
∇f · d~r = f (~r(b)) − f (~r(a))
C
This lets us solve line integrals just like we solve single variable, by deter-
mining an anti derivative and then evaluating at the end points. One important
note is that any continuous function has an antiderivative, not every contin-
uous vector field has a potential function, only (by definition) gradient fields
do. We will spend some time learning about how to test if a vector field is a
gradient field. If a vector field is not a gradient field, you can evaluate the line
integral how we did in the previous section.
There are two important consequences of the fundamental theorem for line
integrals:
1. Circulation in a conservative field is 0.
This is because in a circulation C is a closed curve (meaning ~r(a) = ~r(b))
and so f (~r(a)) = f (~r(b)), or in other words
I
∇f · d~r = f (~r(b)) − f (~r(a)) = 0
C
2. Line integrals of conservative fields are independent of path, meaning

they only depend on the endpoints of the curve and not the path to get
between them.
If F~ is conservative, then F~ is independent of path. The converse is sometimes
true, if F~ is independent of path, and the domain D of F~ is open and connected,
then F~ is conservative.
82
Finding a Potential Function
When using the fundamental theorem for line integrals, we want to know a
potential function for the vector field. The following is a general strategy:
Strategy 6.3.1 Finding Potential Functions in R2
Given some vector field in the form:

~ P (x, y)
F =
Q(x, y)
... you can find a potential function f such that ∇f = F~

by:
1. Integrate P (x, y) with respect to x. This will be a
function in the form g(x, y) + h(y).
2. Take the partial derivative of g(x, y) + h(y) with re-

spect to y, resulting in the function gy (x, y) + h0 (y).
3. Make the equation:
g y (x, y) + h0 (y) = Q(x, y)
... to determine h0 (y).

4. Integrate h0 (y) with respect to y to find h(y).
5. Any function in the form:
f (x, y) = g(x, y) + h(y) + C
... if a potential function for F~ .
83
Strategy 6.3.2 Finding Potential Functions in R3
Given some vector field in the form:

 
P (x, y, z)
F~ = Q(x, y, z)
R(x, y, z)
... you can find a potential function f such that ∇f = F~

by:
1. Integrate P (x, y, z) with respect to x. This will be a
function in the form f = g(x, y, z) + h(y, z).
2. Take the partial derivative of f = g(x, y, z) + h(y, z)
with respect to y, resulting in the function fy =
gy (x, y, z) + hy (y, z).
gy (x, y, z) + hy (y, z) = Q(x, y, z)
... to determine hy (y, z). If at this point it is impos-

sible for hy (y, z) to exist, then no potential function
exists.
4. Integrate hy (y, z) with respect to y to determine
h(y, z), resulting in some h(y, z) = q(y, z) + p(z).
gz (x, y, z) + qz (y, z) + p0 (z) = R(x, y, z)
... to determine p0 (z).

6. Integrate p0 (z) with respect to z to determine p(z).
7. Any function in the form:
f (x, y, z) = g(x, y, z) + q(y, z) + p(z) + C
... if a potential function for F~ . We now know

g(x, y, z) from step 1 and q(y, z) from step 4, and p(z)
from step 6.
Test for Conservative Fields

We need to know if a given vector field is conservative, we can do this by:
84
Theorem 6.3.2 Test for Conservative Vector Fields
Let F~ =< P, Q, R > be a vector field on an open, simply

connected region D. Then Py = Qx , Pz = Rx , and Qz = Ry
thoughout D if and only if F~ is conservative.
6.4 Green’s Theorem

We will finish this course by understanding three important theorems:
1. Green’s Theorem (6.4)
2. Stokes’ Theorem (6.7)
3. The Divergence Theorem (6.8)

The sections in between are added theory to work up to those final two theorems.
Green’s theorem connects the circulation of a vector field, with the double
integral of the area enclosed. This is a direct continuation of the idea that to
evaluate an integral, we can use information about the boundary of integration.
We will now formally state Green’s theorem:
Theorem 6.4.1 Green’s Theorem, Circulation Form
Let D be an open, simply connected region with a boundary

curve C that is a piecewise smooth, simple closed curve
oriented counterclockwise. Let F~ be a vector field in the
form:
P (x, y)
F~ =
Q(x, y)
... with continuous partial derivatives on D. Then:
I I ZZ
~
F · d~r = P dx + Qdy = (Qx − Py )dA
C C D
85
There are many notes to make about this theorem:
1. Since: Z Z
P dx + Qdy = F~ · T~ dt
C C
... this is also called tangential form of Green’s theorem.
2. Green’s theorem only applies to vector fields in R2 .
3. The first two expression are just different notations for the line integral,
but the magic is that it is equal to the double integral on the right.
4. The theorem can be quickly proven for conservative vector fields since the
circulation integral is equal to 0 because of the fundamental theorem of
line integrals, and if F~ is conservative then Qx = Py =⇒ Qx − Py = 0.
5. The theorem can be used to turn a tricky line integral into a simpler
double integral, or a tricky double integral into a simpler line integral.
6. To use this theorem to calculate a double integral, we can use a vector

field with the property:
Qx − Py = 1
... such as: y
−
F~ = x2
2
... and then use the following process:

ZZ
Area = dA
D
ZZ
= (Qx − Py )dA
D
86
Z
= F~ · dR
~
C
Z
1
= −ydx + +xdy
2 C
7. The previous theorem was labelled circulation form because there is an-
other form of Green’s theorem which will be discussed later. This other
form is a direct result of the main Green’s theorem (circulation form) and
so really if you want to think of the Green’s theorem, it would be this one.
We now discuss the flux form of Green’s theorem:
Theorem 6.4.2 Green’s Theorem, Flux Form
Let D be an open, simply connected region with a boundary

curve C that is a piecewise smooth, simple closed curve
oriented counterclockwise. Let F~ be a vector field in the
form:
P (x, y)
F~ =
Q(x, y)
... with continuous partial derivatives on D. Then:
I ZZ
~ ~
F · N ds = (Px + Qy )dA
C D
This extension of Green’s theorem relates the flux through a closed curve C to
the double integral of the bounded region.
This form of the theorem is useful in calculating the flux across a curve by
turning it into a double integral.
87
Source-Free Vector Fields
A source free-vector field F~ is one where the flux of F~ around any closed curve
is 0. Equivalently:
1. The flux F~ · N~ ds across any closed curve C is zero.
H
2. Flux is independent of path.

3. There is a stream function g(x, y) for F~ . A stream function is like a
potential function but for source-free fields. A stream function for F~ =<
P, Q > is a function g such that:
P = gy
Q = −gx
... geometrically this means that the vector field is always tangential to
the curve since:
F~ (a, b) · ∇g(a, b) = 0
... for any point (a, b) in the domain of g.
4.
Px + Qy = 0
6.5 Divergence and Curl

In this section we study two important operations which can be applied to a
vector field, namely:
Divergence: div F~ = ∇ · F~
Curl: curl F~ = ∇ × F~
Divergence
Divergence is an operation which can be applied to a vector field and returns a
scalar value. Divergence at a point measures the tendency of the vector field to
flow out of the point. A positive divergence means more flow is coming out of
the point than coming in. Effectively it is the flux per unit volume at the point.
The following is its formal definition:
88
Definition 6.5.1 Divergence
If:  
P
F~ = Q
R
... then the divergence of F~ is defined by:
div F~ = Px + Qy + Rz
This can be interpreted as the dot product between the gradient vector:
 
∂
 ∂x
∂ 
∇ =  ∂y 
∂
∂z
... and the vector field F~ :

div F~ = ∇ · F~
Physically, divergence can be interpreted by imagining fluid flow. If the diver-
gence at a point is positive, then the surroundings of the point act as a source
of water. If the divergence at a point is negative, the surroundings act as a
sink that sucks in water. If the divergence is equal to 0 then we say the fluid is
incompressible.
The following image has zero divergence, and then negative divergence:
89
Divergence and Green’s Theorem
Recall the flux form of Green’s theorem:
I ZZ
F~ · N
~ ds = Px + Qy dA
C D
Note that the divergence of a 2D vector field is:

div F~ = Px + Qy
... and so Green’s theorem becomes:
I ZZ
~ ~
F · N ds = div F~ dA
C D
If you think of the divergence as a type of derivative, then this continues the
trend of being able to evaluate an integral (in this case a double integral) by
using information about the anti-derivative at the boundary. Using divergence
we can see that Green’s theorem is directly a higher dimensional analogue to
the fundamental theorem of Calculus.
Curl
The second operation we can apply to vector fields is curl. The curl of F~ at
a point P is a vector that measures the tendency of particles near P to rotate
about the axis that points in the direction of this vector.
To visualize this, imagine putting a paddlewheel in the vector field, every
point in the vector field can be assigned a vector with magnitude related to
how much the fluid is swirling, and direction related to the axis around which
it tends to swirl.
90
Definition 6.5.2 Curl
If:  
P
F~ = Q
R
... is a vector field, then curl F~ is defined as:

 
Ry − Qz
curl F~ =  Pz − Rx 
Qx − Py
... if F~ is in R2 then:  
0
curl F~ =  0 
Qx − Py
In general, computation of curl can be interpreted as the cross product of the
gradient vector with the vector field:
curl F~ = ∇ × F~
The following is a useful theorem for conservative vector fields:
Theorem 6.5.1 Curl Test
Let F~ =< P, Q, R > be a vector field in space on a simply

connected domain. If curl F~ = 0 then F~ is conservative.
91
Curl and Green’s Theorem
Recall the circulation form of Green’s Theorem:
I ZZ
F~ · d~r = Qx − Py dA
C D
Note that for a vector field in R2 :
curl F~ = (Qx − Py )k̂
This implies:
curl F~ · k̂ = Qx − Py
... and so Green’s theorem can once again be rewritten as:
I ZZ
F~ · d~r = curl F~ · k̂dA
C D
If we interpret the curl of a vector field as a type of derivative, then once again
this shows that you can evaluate an integral by knowing information about its
anti-derivative around the boundary of integration.
6.6 Surface Integrals

A surface integral is the two dimensional analogue to the line integral. Let’s
remind ourselves about how we define our integrals:
A regular integral is what you get when you evaluate a function at every
x value along the x − axis and add them all together.
A scalar line integral is what you get when you evaluate a function at
every point on any curve (regular integral is a special case of this) and
add them all together.
A double integral is what you get when you evaluate a function at every
point inside a region of the xy − plane, and then add them all together.
A scalar surface integral is what you get when you evaluate a function
at every point on any surface (double integral is a special case of this) and
add them all together.
Note that just like there is a scalar, and vector line integral, there is also a scalar
and vector surface integral.
Before we define the scalar surface integral, we need to get better at describ-
ing surfaces in 3D space. Just like in line integrals, we want to parametrize a
curve, here we want to parametrize a surface.
92
Definition 6.6.1 Parametric Surfaces
Given some surface S, there is a two-variable parameteriza-

tion in the form:
 
x(u, v)
~r(u, v) = y(u, v)
z(u, v)
You can interpret this as taking the uv−plane and mapping each point into
some surface in xyz−space. The uv−plane is called the parameter space or
parameter domain and is the set of points which can be plugged into ~r.
Common parametrizations include:
Cylinder of Radius R:
 
R cos(u)
~r(u, v) =  R sin(u) 
v
Sphere of Radius R:
 
R cos(θ) sin(φ)
~r(φ, θ) =  R sin(θ) sin(φ) 
R cos(φ)
0 ≤ θ ≤ 2π, 0 ≤ φ ≤ π
Explicit Surface
y = f (x, y)
 
u
~r(u, v) =  v 
f (u, v)
Surface Integral of a Scalar-Valued Function

We are now ready to define the scalar surface integral, and a method to evaluate
it:
Definition 6.6.2 Scalar Surface Integral
The surface integral of a scalar-valued function of f

over a piecewise smooth surface S is:
ZZ m X
X n
f (x, y, z)dS = lim f (~r(u∗ , v ∗ ))∆Sij
S m,n→∞
i=i j=1
93
Just like in line integrals, where we used:
ds = ~r 0 (t) dt
... here we use an analogous equation:
dS = ~tu × ~tv dA
Where:
~tu =< ∂x , ∂y , ∂z >
∂u ∂u ∂u
~tv =< ∂x , ∂y , ∂z >
∂v ∂v ∂v
... to create the following functional equation for the scalar surface integral:
ZZ ZZ
f (x, y, z)ds = f (~r(u, v)) ~tu × ~tv dA
C D
... where D is the region of the parameter space which defines the surface.
Orientations of Surfaces
Just like curves can be oriented with line integrals, so can surfaces. At every
point on a surface you can define two normal vectors N ~ and −N ~ which point
in opposite directions. Choosing one gives an orientation to your curve. Given
some orientable surface, we can define it’s normal:
~ ~
~ = tu × tv
N
~tu × ~tv
A surface can be oriented if it has an ”inner” and an ”outer” side, or an ”upward”

and a ”downward” side.
Surface Integral of a Vector Field

Imagine taking a surface in a vector field, chopping it up into small pieces, and
calculating F~ · N
~ at any point in that piece, doing that for all the pieces, and
adding them all together. This idea is flux through a surface, and is what a
vector surface integral calculates.
Definition 6.6.3 Vector Surface Integral
Let F~ be a continuous vector field with a domain that con-

~ . Then the
tains the oriented surface S with unit normal N
~
surface integral of F over S is:
ZZ ZZ
~ ~
F · dS = F~ · N
~ dS
S S
94
This type of integral is also called a flux integral.
A more functional form of this integral is:
ZZ ZZ ZZ
~ ~
F · dS = ~ ~
F · N dS = F~ (~r(u, v)) · (~tu × ~tv )dA
S S D
6.7 Stokes’ Theorem

Stokes’ theorem is a higher dimensional analogue to Green’s theorem. Once
again it allows us to relate an integral to its boundary conditions.
Recall that circulation form of Green’s theorem can be stated as:
I ZZ
~
F · d~r = curl F~ · k̂dA
C D
This can be interpreted visually as:
As in the tendency of the field to curl around the point on the plane has
an axis of rotation normal to the surface itself, and the double integral of this
curling tendency across the region D is equal to the circulation integral around
the boundary.
We now generalize this to curves which are not just on the xy−plane. Here,
the normal is not always k̂, as in:
95
If we want this to extend as expected, then the same logic should apply like
the surface above. The tendency of the field to curl around a point within the
surface has an axis of rotation N ~ which is normal to the plane, and the double
integral of this curling tendency across the surface S is equal to the circulation
integral around the boundary.
It turns out, its true.
Theorem 6.7.1 Stokes’ Theorem
Let S be a piecewise smooth oriented surface with a bound-

ary that is a simple closed curve C with positive orientation.
If F~ is a vector field with continuous partial derivatives then
on an open region containing S, then:
I ZZ ZZ
~
F · d~r = ~ ~
curl F · dS = (curl F~ ) · N
~ dS
C S S
6.8 The Divergence Theorem

Continuing the trend we see in these theorems, that we can evaluate an integral
by knowing information about the antiderivative at the boundary, we can imag-
ine that the surface integral of a surface which encloses some volume is related
in some way to the triple integral of the volume being enclosed. Recall the flux
form of Green’s theorem:
I ZZ
F~ · N
~ ds = div F~ dA
C D
This tells us that given some region in the plane, we can determine the flux
through its boundary by knowing the integral of the divergence inside. Imagine
then taking the region and stretching it out into xyz−space. The same reasoning
96
holds in one higher dimension, we can know about the flux through the surface
which was created by the boundary, by knowing information about the integral
of the divergence in the volume.
Theorem 6.8.1 The Divergence Theorem
Let S be a piecewise, smooth, closed surface that encloses a

solid E. Assume that S is oriented outward, and let F~ be a
vector field with continuous partial derivatives on an open
region containing E. Then:
ZZZ ZZ
div F~ dV = F~ · dS
~
E S
Concluding Remarks
In summary, the following is a list of all the important theorems in Calculus we
have learned so far:
The Fundamental Theorem of Calculus:

Z b
f 0 (x)dx = f (b) − f (a)
a
97
The Fundamental Theorem for Line Integrals:
Z
∇f · d~r = f (P1 ) − f (P0 )
C
Green’s Theorem, circulation form:

ZZ Z
(Qx − Py )dA = F~ · d~r
D C
Green’s Theorem, flux form:

ZZ Z
(Px + Qy )dA = F~ · N
~ ds
D C
Stokes’ Theorem:
ZZ Z
curl F~ · dS
~= F~ · d~r
S C
The Divergence Theorem:

ZZZ ZZ
~
div F dV = F~ · dS
~
E S
All applied in their particular domains, they are extensions of the Fundamen-
tal Theorem of Calculus into higher dimensions. The derivative, the gradient,
the divergence, and the curl of their respective types of functions are all differ-
ential operators, as in they reason about the rate of change of the function in
different ways. The unifying principle is:
Integrating a differential operator acting on a field over a domain
is equivalent to integrating the field components along the boundary.
Conclusion
This concludes all the content in this course. I really hope these notes helped,
I really enjoyed making them.
Good luck on the exam!
- Adam Szava
98

MTH312 Notes

Uploaded by

Copyright:

Available Formats

MTH312 Notes

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MTH312 Notes

Uploaded by

Copyright:

Available Formats

Differential Equations and Vector Calculus

(MTH 312): Notes

Introductory Topics and Definitions

Definition 0.1.1 Differential Equations

An equation containing derivatives of one or more depen-

1. General Form for nth order ODEs:

... where F is a function of n + 2 variables. We assume in this course that

y (n) = f (x, y, y 0 , y 00 , . . . , y (n−1) )

3. Differential Form for first order ODEs:

M (x, y)dx + N (x, y)dy = 0

These two equations are the key to solving DEs:

Definition 0.1.2 Solutions to a DE

Any function φ, defined on some interval I possessing n

Definition 0.1.3 Implicit Solutions

A relation G(x, y) = 0 is an implicit solution to an ODE

Review of Differential Equation Content from MTH

Definition 1.1.1 Terminology

A Differential Equation (DE) is an equation in variables

A DE is separable if it can be rearranged into the form:

As in you can separate x and y.

In order to solve a separable differential equation, you follow this formula:

4.5 Linear Differential Equations

Definition 1.2.1 Linear D.E. (L.D.E.)

A linear differential equation (LDE) is a DE which can

Notably the coefficient on the y 0 term is 1. In order to find a solution to this

Definition 1.2.2 Integrating Factor

In general (not just for LDEs) an Integrating Factor is a

Where p(x) is the same p(x) as in the LDE.

This helps us solve the LDE because of the following:

Chapter 2: First Order Equations

2.2 Separable Differential Equations

Definition 2.1.1 Seperable DE

A first order DE in the form:

Is seperable if and only if f (x, y) can be written as:

So the DE can be written in the form:

In this course we often abuse the differential notation, treating it as a frac-

2.5 Exact Differential Equations

... we can choose to write it as the following:

M (x, y)dy = P (x, y)dx

P (x, y)dx + Q(x, y)dy = 0

Given a differential equation in this form, we can check if it is exact.

Definition 2.2.1 Exact DE

A first order DE in the form:

P (x, y)dx + Q(x, y)dy = 0

Is exact if and only if:

P (x, y)dx + Q(x, y)dy = 0

We can use the following theorem to find it:

Theorem 2.2.1 Exactness Solutions

If u(x, y) has continuous partial derivatives ux (x, y) and

ux (x, y)dx + uy (x, y)dy = 0

P (x, y)dx + Q(x, y)dy = 0

Method to Solve a Exact DE

P (x, y)dx + Q(x, y)dy = 0

1. Check for exactness:

4. Write the equation:

6. Solve for h(y) by integrating h0 (y)

Definition 5.1.1 nth Order LDE

A nth order linear differential equation (LDE) can be written

Characteristics which identify an LDE:

The dependent variable is always of first degree.

Coefficient functions (a0 , a1 , . . . , an ) are all functions of only the indepen-