Notes∗ to MATH 156

Calculus for Business and Economics II

Winter 2020

Christoph Frei
[email protected]

∗ Thesenotes are subject to change, so it is best not to print

them more than a couple of lectures ahead. Reading them on
an electronic device or printing multiple pages per sheet will
save on paper.
Overview of MATH 154 and MATH 156
Main topics of MATH 154 Z 1
R= f (x) dx
0

slope = f 0 (x)

Functions Differentiation Integration

Main topics of MATH 156

f
y (x0 ,y0 )

Z 1
R= f (x) dx
1

Integration techniques Differentiation

Partial derivatives Probability and calculus

Business and economics applications in MATH 154/6

Supply and demand Compound interest

curves Differentiation Optimization problems

Main source of figures in these notes: textbook

Goals and objectives of MATH 156

• Extend your calculus skills and knowledge in four

important areas:
· integration techniques
· partial derivatives
· multivariate optimization
· probability and calculus.

• We will emphasize topics that are important for

business and economics, and show applications in
these areas.

• The course serves as a basis for many other courses

in business, economics, and science, where these
concepts are applied.

• The focus is on understanding and learning to

apply these concepts.

• Note that proving theorems is not a main part of

this course as opposed to the course MATH 118
(Honors Calculus II).
Who should take MATH 156?

• Everyone interested in a second calculus course

with applications to business and economics.

• Bachelor of Commerce students can take MATH

156, but are not required to do so.

• Students majoring in Economics are required to

take MATH 156. MATH 156 is a prerequisite to
ECON 384, 385, 386, and 399.∗ Economics stu-
dents can take MATH 156 in their first or second
year. It is recommended to take MATH 156 in
the term immediately following MATH 154.

• Students in the specialization programs Mathe-

matics and Economics as well as Mathematics and
Finance should take MATH 154 and 156 during
their first year.

∗ For students who take or took ECON 299 in Winter 2020

or earlier, the MATH 156 prerequisite will be waived for
ECON 384, 385, 386, and 399 in Fall 2020/Winter 2021
only; see https://www.ualberta.ca/economics/undergraduate-
programs/ba-econ-major.
Contents

1 Integration Techniques 7

1.1 Brief review of integration . . . . . . . . . . . . . . . 8

1.2 Integration by parts . . . . . . . . . . . . . . . . . . . . 12

1.3 Improper integrals . . . . . . . . . . . . . . . . . . . . . 18

1.4 L’Hôpital’s rule . . . . . . . . . . . . . . . . . . . . . . . . 26

1.5 Taylor approximation . . . . . . . . . . . . . . . . . . . 28

2 Partial Derivatives 32

2.1 Functions of several variables . . . . . . . . . . . 32

 2.2 Definition and application in economics . 41

2.3 The chain rule, tangent plane and

2.3 gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3 Multivariate Optimization 61

3.1 Extrema and saddle points . . . . . . . . . . . . . 62

3.2 Constrained extrema . . . . . . . . . . . . . . . . . . . 74

3.3 Total differential . . . . . . . . . . . . . . . . . . . . . . . 82

3.4 Double integrals . . . . . . . . . . . . . . . . . . . . . . . 86

4 Probability and Calculus 93

4.1 Probability distributions . . . . . . . . . . . . . . . . 94

4.2 Expected value and standard deviation . . 108

4.3 Covariance and correlation . . . . . . . . . . . . . . 119

 7

Chapter 1

Integration Techniques

Goal: extend topics discussed in MATH 154 (or MATH

114), namely:
· integration methods → integration by parts
· integrals → improper integrals
· limits → L’Hôpital’s rule
· differentials → Taylor approximation
 8

1.1 Brief review of integration

We review the essentials of integration. For more de-
tails, please see my notes to Chapter 5 of MATH 154.

• Definitions: There are three important definitions

related to integrals of a function f :
Name Definition Type
Antideri- any function F satisfying
a function
vative F 0(x) = f (x) for all x
R
f (x) dx = F (x) + C
Indefinite a set of
integral the collection of all functions
antiderivatives of f
Rb
a f (x) dx is equal to the
Definite
integral signed area of the region a number
under/above graph of f

• Computation of Indefinite Integrals:

1. Try to simplify the integrand f .

Z
c1 xr1 +···+cn xrn
In particular, write kxs dx in the
Z  
c1 r1 −s cn r −s
form kx + ··· + k x n d x.
 9

2. Use linearity of integrals:

Z  
c1f1(x) + c2f2(x) dx
Z Z
= c1 f1(x) dx + c2 f2(x) dx

3. Apply known integration rules (memorize them):

Rule Function Integral
R
Const. f (x) = k R
k dx = kx + C
 xr dx = 1 xr+1 + C if r =6 −1
Power f (x) = xr R r+1
 x−1 dx = ln |x| + C if r = −1
R x x
Exp. f (x) = exp(x) e dx = e + C
R
Sine f (x) = sin(x) sin(x) dx = − cos(x) + C
R
Cosine f (x) = cos(x) cos(x) dx = sin(x) + C

4. If the integrand contains a composed function,

use integration by substitution:
Z Z
f (g (x)) g 0(x) dx = f (u) du

• Computation of Definite Integrals: Find an an-

tiderivative F (x) using the above rules for indef-
 10

inite integrals and then apply

Z b
f (x) dx = F (b) − F (a).
a
This follows from the Fundamental Theorem of
Calculus. Additionally, when doing integration by
substitution, adjust the limits of integration:
Z b Z g(b)
f (g (x))g 0(x) dx = f (u) du.
a g(a)

• Examples:
Z
x2 + 3 x + 2
1. Evaluate 2
d x.
x +x
We first simplify the integrand, using

x2 + 3x + 2 (x + 1)(x + 2) 2
2
= = 1 + ,
x +x x(x + 1) x
so that
Z Z  
x2 + 3 x + 2 2
2
d x = 1 + dx
x +x x
= x + 2 ln |x| + C.
 11
Z
2. Evaluate 5x2 cos(x3 + 1) dx.

We make the substitution u = x3 + 1 because

its differential du = 3x2 dx appears in the in-
tegral, apart from the constant factor 3. Thus,
we compute
Z
5x2 cos(x3 + 1) dx
Z Z
5 3 2 5
= cos(x + 1)(3x dx) = cos(u) du
3 3
5 5
= sin(u) + C = sin(x3 + 1) + C.
3 3

Z 2 q
3. Evaluate 5x3 x2 − 1 dx
1
p
Because x2 − 1 can be seen as a composite
function of square root and x2 − 1, we apply
integration by substitution with u = x2 − 1 so
that du = 2x dx. We can write
Z q Z q
3 5
5x x2 − 1 dx = x2 − 1 x2 (2x dx),
2 | {z | {z }
√ }
= u =du

but there is still an x2, which we can replace

by x2 = u + 1. Moreover, the integration
 12

boundaries 1 and 2 become 0 and 3 so that

Z 2 q Z q
3 5 2
2
5x x − 1 dx = x2 − 1 x2(2x dx)
1 2 Z1
5 3√
= u(u + 1) du
2 Z0
5 3 3/2
= (u + u1/2) du
2 0
 
5 2 5/2 2 3/2 3
= u + u
25 3  0
5 2 5/2 2 3/2
= 3 + 3
2 5 3 √
=3 5/2 +5·3 1/2 = 14 3.

1.2 Integration by parts

→ Section 7.1 in the textbook

• Motivation: In addition to substitution, integra-

tion by parts is an important integration method.
We use it when the integrand is a product of two
factors, where one factor becomes simpler when
differentiated and the other factor does not be-
come harder when integrated.
 13

• Integration by Parts: For indefinite integrals:

Z Z
f (x)g 0(x) dx = f (x)g (x) − f 0(x)g (x) dx,

or equivalently, written in compact form (with

u = f (x), v = g (x) =⇒ du = f 0(x) dx, dv =
g 0(x) dx)
Z Z
u dv = uv − v du.

For definite integrals:

Z b Z b
b
f (x)g 0(x) dx = f (x)g (x) − f 0(x)g (x) dx,
a a a
where
b
f (x)g (x) = f (b)g (b) − f (a)g (a).
a

• Reason: Like for integration by substitution (which

is based on the chain rule for derivatives), integra-
tion by parts has a corresponding differentiation
rule, namely, the product rule, which says
d 
f (x)g (x) = f 0(x)g (x) + f (x)g 0(x).
dx
 14

Integrating both sides of this equation yields

Z Z
d 
f (x)g (x) dx = f 0(x)g (x) dx
dx Z
+ f (x)g 0(x) dx,
which leads to the integration by parts formula,
using Z d  
f (x)g (x) dx = f (x)g (x) + C.
dx
• Examples: Z
1. Evaluate xe5x dx.
The only way to evaluate this integral is by
using integration by parts. We set f (x) = x
because f 0(x) = 1 becomes simpler while an
antiderivative of e5x is 51 e5x. We compute
Z Z
1 1
x e5x dx = x e5x − 1 e5x dx
↓ ↑ 5 5
1 1 e5x
5
Z
x 5x 1 5x
= e − e dx
5 5
x 5x 1 5x
= e − e +C
5 25
5x − 1 5x
= e +C
25
 15

Note that in this example, we need to choose

x as the factor for derivation. If we chose x as
the factor for integration, we would get 12 x2,
which would make the integral more compli-
cated, rather than easier.
Z
2. Evaluate x2e5x dx.
This is similar to the first example, but now
there is x2 rather than x in the integrand. We
proceed similarly
Z Z
1 1
x2 e5x dx = x2 e5x − 2x e5x dx
↓ ↑ 5 5
2x 1 e5x
5
Z
x2 5x 2
= e − xe5x dx
5 5
x2 5x 2 5x − 1 5x
= e − e +C
5 5 25
25x2 − 10x + 2 5x
= e +C
125
using that
Z
5x − 1 5x
xe5x dx = e +C
25
 16

by the first example. Because C stands for

an arbitrary constant, we can write +C rather
than − 52 C in the above computation.
Z 2
3. Evaluate 2x3 ln(4x) dx.
1
The difficulty of this integral is the term ln(4x).
So, we want to make it easier by taking its
derivative: f (x) = ln(4x) gives f 0(x) = 4x 14=
1 by the chain rule. Hence, we compute
x
Z 2 2 21 1 Z
3 1 4
2x ln(4x) dx = x ln(4x) − x4 d x
1 ↑ ↓ 2 1 1 2 x
1 x4 1
2 x
Z
1 4 1 1 2 3
= 2 ln(8) − ln(4) − x dx
2 2 2 1
3 1 2 1 42
= 8 ln(2 ) − ln(2 ) − x
2 8 1
1
= 8 · 3 ln(2) − ln(2) − 2 +
8
15
= 23 ln(2) − .
8
Z 3
4. Evaluate ln(2x) dx.
2
 17

Similarly to the previous example, we make

the integrand easier by taking the derivative of
ln(2x): f (x) = ln(2x) gives f 0(x) = 2x12= 1
x
by the chain rule. However, there is no other
factor present that we can integrate. Hence,
we write the integrand as ln(2x) = 1 · ln(2x)
so that we have a factor 1 to integrate and
obtain
Z 3 3 Z 3
1
1 · ln(2x) dx = x ln(2x) − x dx
2 ↑ ↓ 2 2 x
x 1
x
Z 3
= 3 ln(6) − 2 ln(4) − 1 dx
2
= 3 ln(6) − 2 ln(4) − x|32
= 3 ln(6) − 2 ln(4) − (3 − 2)
= 3 ln(6) − 2 ln(4) − 1.
 18

1.3 Improper integrals

→ Section 7.4 in the textbook

• Motivation:
Z
So far, we considered definite inte-
b
grals f (x) dx over a finite interval [a, b] of
a
integration. However, often in applications, we
encounter situations with unbounded intervals of
integration. For such cases, we define improper
integrals.

• Definition: For a continuous function f , improper

integrals of f are defined as follows:
Z ∞ Z b
• f (x) dx = lim f (x) dx if the limit
a b→∞ a
Z b
lim f (x) dx exists.
b→∞ a
Z b Z b
• f (x) dx = lim f (x) dx if the limit
−∞ a→−∞ a
Z b
lim f (x) dx exists.
a→−∞ a
Z ∞ Z 0 Z ∞
• f (x) dx = f (x) dx + f (x) dx if
−∞ Z −∞ Z ∞ 0
0
both f (x) dx and f (x) dx exist.
−∞ 0
 19
Z ∞
• Computation: To compute f (x) dx, we first
Rb a
compute a f (x) dx and then take the limit for
b → ∞ of the resulting expression.

• Interpretation: As an example, let us consider

the function f (x) = x12 . The definite integral
Z b
1 dx gives the area of the region under the
2
1 x
graph of f (x) = x12 on [1, b], as the left figure
below
Z shows. The value of the improper integral
∞
1 dx equals the area of the region under the
1 x2
graph of f to the right of the vertical line x = 1,
as shown in the right figure below. If this value
is finite, we say that the improper integral exists.
 20

• Examples:
Z ∞
1. Evaluate 1 d x.
1 x2
Rewriting the improper integral as a limit and
evaluating the integral, we find
Z ∞ Z b
1 −2 dx
d x = lim x
1 x2 b→∞ 1 
b
= lim − x−1
b→∞ 1
 
1
= lim − +1
b→∞ b
= 1.
Z ∞
2. Evaluate 1 d x.
1 x
Similarly, we compute
Z ∞ Z b
1 1
dx = lim dx
1 x b→∞ 1 x 
b
= lim ln(x)
b→∞  1

= lim ln(b) − ln(1) ,
b→∞
Z ∞
which diverges to ∞ so that 1 dx does not
1 x
exist. The reason behind the difference com-
 21

pared to the first example is that x12 converges

faster to zero than x1 does, as x goes to ∞.
For x12 , the convergence is fast enough so that
the area of the region below the graph is finite
while it is not for x1 .
Z 0
−x2
3. Evaluate xe d x.
−∞
Since the limit is here for −∞, we write
Z 0 Z 0
xe −x2 dx = lim
2
xe−x dx.
−∞ a→−∞ a

Using integration by substitution with u = x2

and du = 2x dx, we find
Z 0 Z
−x2 1 0 −x2
xe dx = e (2xdx)
a 2 Za
1 0 −u
= e du
2 a 2
1 −u 0
=− e
2 a2
1 1 2
= − + e−a .
2 2
 22

Therefore, we conclude
Z 0 Z 0
xe −x2 dx = lim
2
xe−x dx
−∞ a→−∞ a
 
1 1 −a2
= lim − + e
a→−∞ 2 2
1
=− .
2

• Application: Recall that future income has a lower

present value because the money cannot be used
immediately. When there is a continuous income
stream over a time horizon [0, T ], its present value
is
Z T
PV = R(t)e−rt dt,
0
where
R(t) = rate of income generated at any time,
r = interest rate compounded continuously,
T = term.
If the time horizon is infinite, this formula be-
comes
Z ∞
PV = R(t)e−rt dt.
0
 23

A special case occurs when R(t) is constant. If

there are m payments per year of the same amount
P made at regular time intervals, this is called an
annuity and its present value is given by
Z T Z T
PV = R(t)e−rt dt = mP e−rt dt
0 0
mP −rt T mP  −rT

=− e = 1−e ,
r 0 r
where

P = size of each payment in the annuity,

r = interest rate compounded continuously,
T = term of the annuity (in years),
m = number of payments per year.

If the payments of an annuity continue indefi-

nitely, it is called a perpetuity. The present value
of a perpetuity equals the improper integral
Z ∞ Z T
PV = mP e−rt dt = lim mP e−rt dt
0 T →∞ 0
mP  −rT
 mP
= lim 1−e = .
T →∞ r r
 24

• Example: A donor of the University of Alberta

wants to establish a scholarship that pays out
each year (beginning next year) to a student a cer-
tain amount A. The continuously compounded
interest rate is 5%.

(a) Assume that A = $1, 000. How much money

does the donor need to fund the endowment?

We can use the formula for the present value of

a perpetuity so that the required amount to fund
the endowment is
mP 1 · $1,000
= = $20,000.
r 0.05

(b) The amount A in each year should be such

that it has the same purchasing power as $1,000
today. Because of inflation, a dollar in one year
has less purchasing power than a dollar today. We
assume that the yearly inflation rate is 2%. This
means that in 10 years, you will need $1.0210 to
buy the same as today for $1. How much money
does the donor need to fund this endowment?
 25

This endowment has a payoff in the amount of

R(t) = $1,000 · 1.02t in year t. Its present value
with r = 0.05 interest rate is
Z ∞ Z ∞
PV = R(t)e−rt dt = 1000·1.02te−0.05t dt.
0 0
To determine this improper integral, we write
t)
1.02t = e ln(1.02 = et ln(1.02).
Therefore, we compute
Z ∞
PV = 1000 · 1.02te−0.05t dt
Z0∞
= 1000 et ln(1.02)e−0.05t dt
0
Z b
= lim 1000 e(ln(1.02)−0.05)t dt
b→∞ 0
1000 b
= lim e(ln(1.02)−0.05)t
b→∞ ln(1.02) − 0.05 0
1000  
= lim e (ln(1.02)−0.05)b −1
b→∞ ln(1.02) − 0.05
1000
= ≈ 33,115.46,
0.05 − ln(1.02)
using lim e(ln(1.02)−0.05)b = 0 since the expo-
b→∞
nent is negative: ln(1.02)−0.05 ≈ −0.03020 < 0.
 26

1.4 L’Hôpital’s rule

→ Appendix B in the textbook

• Motivation: In MATH 154 (or MATH 114), you

have computed limits of ratios that are of indeter-
minate form 00 when numerator and denominator
are polynomials, for example,
x2 − 4 (x − 2)(x + 2)
lim = lim
x→2 x − 2 x→2 x−2
= lim (x + 2) = 4.
x→2
We now introduce a technique that works more
generally.

• L’Hôpital’s rule:
Assume that f and g are differentiable functions
around a point a and g 0(x) 6= 0 around a with
either
· both lim f (x) = 0 and lim g (x) = 0, or
x→a x→a
· both lim f (x) = ±∞ and lim g (x) = ±∞.
x→a x→a
Then
f (x) f 0(x)
lim = lim 0 .
x→a g (x) x→a g (x)
 27
f 0 (x)
• Warning: Note that to compute limx→a g0(x) ,
we
take the derivative of each numerator and denom-
inator; we do NOT apply the question rule for
differentiating.

• Examples:
ln(x)
1. Evaluate lim x−1 .
x→1
Because

lim ln(x) = ln(1) = 0, lim (x − 1) = 0,

x→1 x→1
the limit is of indeterminate form 00 so that we
can apply L’Hôpital’s rule, which yields
ln(x) 1/x 1/1
lim = lim = = 1.
x→1 x − 1 x→1 1 1
x .
2. Evaluate lim x−1
x→1
Because lim x = 1, we cannot use L’Hôpital’s
x→1
rule. Instead, we note that the limit is of the
1 , hence lim x
form 0 does not exist.
x→1 x − 1
Note that if we applied L’Hôpital’s rule, we
would have obtained 11 = 1, which is the wrong
result.
 28
x 2
3. Evaluate lim ex .
x→∞
Because lim x2 = ∞ and lim ex = ∞, the
x→∞ x→∞
limit is of the indeterminate form ∞∞ so that

we can apply L’Hôpital’s rule, which yields

x2 2x
lim x = lim x .
x→∞ e x→∞ e
Because the limit is still of the indeterminate
form ∞∞ , we apply L’Hôpital’s rule a second

time, and conclude

x2 2x 2
lim = lim x = lim x = 0.
x→∞ ex x→∞ e x→∞ e

1.5 Taylor approximation

→ Section 11.1 in the textbook

• Motivation: In MATH 154 (or MATH 114), you

learned that the graph of a function can be ap-
proximated by the tangent line. The tangent line
is given by a linear function, which is a polynomial
 29

of degree 1; compare the left figure below. We

now study a more accurate approximation by us-
ing a polynomial of degree 2; compare the right
figure below.

• Definition: For a twice differentiable function f ,

the Taylor approximation (also known as Taylor
polynomial of degree 2) at a point a is given by

0 f 00(a)
P (x) = f (a) + f (a)(x − a) + (x − a)2.
2

• Comment: The first two terms in the Taylor ap-

proximation, f (a) + f 0(a)(x − a), correspond to
the tangent line at the point a. The third term,
f 00 (a) 2 , is quadratic in x and improves the
2 (x − a)
linear approximation.
 30

• Examples:

1. Use a Taylor polynomial of degree 2 to ap-

proximate e0.1.
We use the Taylor approximation for f (x) =
ex at a = 0 because a = 0 is close to x = 0.1
and e0 = 1 can be explicitly computed. The
Taylor approximation is
f 00 (0)
P (x) = f (0) + f 0(0)(x − 0) + (x − 0)2
2
0 0 e0 2
=e +e x+ x
2
1
= 1 + x + x2
2
so that
0.1 1 2
e ≈ 1 + 0.1 + 0.1 = 1.105,
2
which is close to e0.1 = 1.10517091 . . .

2. Use a Taylor
√ polynomial of degree 2 to ap-
proximate 25.5.
We use the Taylor approximation for f (x) =
√
x at a = 25 because 25 is close to 25.5 and
 31
√
25 = 5 can be explicitly computed. We have
1 −1/2
f (x) = x1/2, f 0(x)
= x ,
2
00 1 −3/2 1
f (x) = − x = − 3/2
4 4x
so that the Tayler approximation becomes

0 f 00(25)
f (25) + f (25)(x − 25) + (x − 25)2
2
1
= 251/2 + 25−1/2(x − 25)
2
1 2
− (x − 25)
4 · 253/2
1 1
= 5 + (x − 25) − (x − 25)2.
10 1000
At x = 25.5, we thus find the approximation
√ 25.5 − 25 (25.5 − 25)2
25.5 ≈ 5 + −
10 1000
0.5 0.25
=5+ −
10 1000
= 5.04975,
√
which is close to 25.5 = 5.049752469 . . .