Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Math 120—Homework #3

due Friday, October 21, 2022 by 9:50AM


There is a 15% late penalty for Homework submitted after the due date and time. (Please take
the submission time of 9:50AM seriously, and plan to finish well before then—preferably the night
before, really.) Solutions to the Homework will be available online 24 hours after the due date and
time, and no late work will be accepted once those 24 hours pass.
When writing mathematics, always try to make your writing look like the writing in your favourite
mathematics textbook, the one you find clearest and easiest to understand. In particular, write in
complete sentences, with words explaining the mathematical things you are doing and putting them
into context (more words than math symbols, more likely). You are not just trying to convey to
your reader the answer you obtained: you are also trying to persuade your reader that your answer
is correct in a way that is easy for them to understand.
For this assignment, you do not need to use formal induction proofs.

Problem 1. When n is a nonnegative integer, recall that n! (read “n factorial”) denotes the product
n! = n × (n − 1) × (n − 2) × · · · × 3 × 2 × 1. For example, 4! = 4 × 3 × 2 × 1 = 24. By convention,
0! = 1 and 1! = 1. In this problem, k always denotes an integer.
d
X
(a) Suppose that p(x) = an xn is a polynomial of degree d (so that ad ̸= 0). Prove that if
n=0
0 ≤ k ≤ d then the kth derivative of p(x) equals
d
(k)
X n!
p (x) = an xn−k ,
n=k
(n − k)!
(k)
while if k > d then p (x) = 0 for all x ∈ R.
dk 
(b) If b > 1 is any constant, find and prove a formula for log b x for all k ≥ 1.
dxk
d
X
(a) Given polynomial p(x) = an xn and to find a general solution for kth derivative of p(x)
n=0
when 0 ≤ k ≤ d

The zeroth derivative of the polynomial will return the polynomial itself. Let’s find the first
derivative of the polynomial (k = 1);
X d
=⇒ p(x) = an xn = a0 + a1 x + a2 x2 + .... + an−1 xn−1 + an xn
n=0
=⇒ p (x) = a1 + 2a2 x + .... + (n − 1)an−1 xn−2 + (n)an xn−1
(1)

The second derivative (k = 2);


=⇒ p(2) (x) = 2a2 + (3)(2)a3 x.... + (n − 1)(n − 2)an−1 xn−3 + (n)(n − 1)an xn−2

The third derivative (k = 3);


=⇒ p(3) (x) = (3)(2)a3 .... + (n − 1)(n − 2)(n − 3)an−1 xn−4 + (n)(n − 1)(n − 2)an xn−3
Looking at these, we start to notice a pattern;
(n − 1)! (n)!
=⇒ p(3) (x) = (3!)a3 .... + an−1 xn−4 + an xn−3
(n − 4)! (n − 3)!
Given that k = 3 we can re-write it as;
(n − 1)! (n)!
=⇒ p(k=3) (x) = (3!)a3 .... + an−1 x(n−1)−k + an xn−k
((n − 1) − k)! (n − k)!
d
X (n)!
=⇒ p(k=3) (x) = an xn−k
n=0
(n − k)!
P
Given our condition of 0 ≤ k ≤ d, we can replace the lower value of with k
d
X (n)!
=⇒ p(k) (x) = an xn−k
n=k
(n − k)!
When we put k = d we get;
d
(d)
X (n)!
=⇒ p (x) = an xn−d
n=d
(n − d)!
=⇒ p(d) (x) = (d!)ad
Which is a constant value. If we take a value of k above d we are differentiating a constant
value.
=⇒ if k > 1 then, p(k) (x) = 0

dk 
(b) Given b > 1 is some constant and to find and prove a general formula for k logb x for
dx
all k ≥ 1

Let’s take the first derivative of the function;


d  1 1
=⇒ logb x =
dx x ln(b)
Second derivative;
d2  −1 1
=⇒ 2
logb x = 2
dx x ln(b)
Third derivative;
d3  2 1
=⇒ 3
logb x = 3
dx x ln(b)
We can notice a pattern a here;
d3  −((3 − 1)!)3−1 1
=⇒ log b x =
dx3 x3 ln(b)
We can now generalise this for any value k ≥ 1;
dk  −((k − 1)!)k−1 1
=⇒ log b x =
dxk xk ln(b)
Problem 2. Be as explicit as possible about the specific limit facts and laws you use from class
and the Canvas module.

(a) Using the formal definition of infinite limits, directly prove the ∞ × ∞ Limit Law.
(b) Derive the ∞ × −∞ Limit Law from the ∞ × ∞ Limit Law.
(c) If p(x) is a polynomial of odd degree, prove that there exists a real number z such that
p(z) = sin z. (Hint: remember Problem 6 on Homework #1.)
(d) If g(x) is a function that is continuous on [− π2 , π2 ], prove that there exists a real number z
such that g(z) = tan z.

(a) Given to show that if lim f (x) = ∞ and lim g(x) = ∞ then lim f (x)g(x) = ∞
x→∗ x→∗ x→∗

From the formal definition of the limit we can say that;

For some given N, ∃ a δ1 > 0 s.t. for |x − t| < δ1 then f (x) > N

For some given N, ∃ a δ2 > 0 s.t. for |x − t| < δ2 then g(x) > N

We have to show to show that for some fixed value of t and for N ∃ δ > 0 s.t. for |x−t| > δ
then, f (x)g(x) > N

Let δ = min(δ1 ), δ2 , and hence,

For |x − t| < δ both f (x) > N and g(x) > N are true simultaneously.

=⇒ f (x)g(x) > N 2 > N for N > 1

Therefore,
=⇒ lim f (x)g(x) = ∞
x→∗

(b) Given to show that if lim f (x) = ∞ and lim g(x) = −∞ then lim f (x)g(x) = −∞
x→∗ x→∗ x→∗

We can prove this quite easily if consider a function g1 (x) = −g(x). From the Infinite limit
negation lemma, lim g1 (x) = ∞
x→∗

Using the same method as the previous question we arrive the conclusion,

=⇒ f (x)g1 (x) < N 2 < N for N ∈ (0, 1)

Therefore,
=⇒ lim f (x)g(x) = −∞
x→∗
(c) Let’s consider the polynomial of p(x),
n−1
X
We can write it as, ai xi s.t. n = 2k and n − 1 > 0
i=0
Consider the limit of the polynomial when x → ±∞
n−1
X
=⇒ lim p(x) = lim ai xi = lim an−1 xn−1 + an−2 xn−2 + .... + a1 x + a0
x→±∞ x→±∞ x→±∞
i=0
n−1 an−2 a1 a0
=⇒ lim p(x) = lim x (an−1 + + .... + n−2 + n−1 )
x→±∞ x→±∞ x x x
Using the multiplication limit law;
an−2 a1 a0
=⇒ lim p(x) = lim xn−1 lim (an−1 + + .... + n−2 + n−1 )
x→±∞ x→±∞ x→±∞ x x x
=⇒ lim p(x) = ±∞(an−1 )
x→±∞
=⇒ lim p(x) = ±∞
x→±∞

From this we can say that ∃ some N, M ∈ R s.t. for x > N , ∃ N ′ ∈ R s.t. f (x) > N ′
and for x < M , ∃ M ′ ∈ R s.t. f (x) < M ′ . Therefore there are no constraints on f (x) and
continuously extends till ±∞.

Now consider the equation p(x) − sin(x). Since sin(x) is restricted as −1 ≤ sin(x) ≤ 1
and the polynomial is −∞ ≤ p(x) ≤ ∞, the overall function is not restricted anywhere
and continuously extends till ±∞.

Also since, both p(x) and sin(x) are continuous their difference is also continuous. Thus,
we can say that p(x) − sin(x) is continuous all throughout the set of real numbers i.e.
(−∞, ∞)

Using Intermediate Value Theorem, we can say that since zero lies between p(x) − sin(x)
at x → ∞ and p(x) − sin(x) at x → −∞ and hence there exists some value z ∈ (−∞, ∞)
s.t. p(z) − sin(z) = 0.

=⇒ p(z) = sin(z)

(d) Let us consider the function g(x) − tan(x), we are given that g(x) is continuous on the
interval [ −π , π]
2 2

For tan(x) the function is continuous on interval ( −π , π ), and hence when we look at g(x)−
2 2
−π π
tan(x) we have to restrict the interval to ( 2 , 2 ) for the function g(x) − tan(x) to be
continuous.

Using Intermediate Value Theorem, we can say that since zero lies between g(x) − tan(x)
at x → π2 and g(x) − tan(x) at x → −π
2
and hence there exists some value z ∈ ( −π , π ) s.t.
2 2
g(z) − tan(z) = 0.

=⇒ g(z) = tan(z)
Problem 3. If f (x) is a function that is defined for all real numbers x, we say that f (x) has a global
maximum at x = a if f (a) ≥ f (x) for all x ∈ R, and we say that f (x) has a global minimum at
x = b if f (b) ≤ f (x) for all x ∈ R.
Be sure that your answers actually prove that the polynomials do or do not satisfy the definition of
a global maximum/minimum, rather than giving a graphical or intuitive argument.

(a) Prove that a polynomial of odd degree has neither a global minimum nor a global maximum.
(b) If p(x) is a nonconstant polynomial of even degree, prove that either p(x) has a global
maximum and no global minimum, or vice versa. (Make sure your proof doesn’t work for
polynomials of odd degree!)
n−1
X
(a) Consider a polynomial defined as p(x) = ai xi such that n = 2k, n ̸= 0 and a0 ̸= 0.
i=0

Let’s consider the limit of the function at x → ±∞

n−1
X
=⇒ lim p(x) = lim ai xi = lim an−1 xn−1 +an−2 xn−2 +....+a2 x2 +a1 x+a0
x→±∞ x→±∞ x→±∞
i=0

n−1
X an−2 a2 a1
=⇒ lim p(x) = lim ai xi = lim xn−1 (an−1 + + .... + n−3 + n−2 +
x→±∞ x→±∞
i=0
x→±∞ x x x
a0
)
xn−1

Now we can use the multiplication limit law;

n−1
X an−2 a2
=⇒ lim p(x) = lim ai xi = lim xn−1 lim (an−1 + + .... + n−3 +
x→±∞ x→±∞
i=0
x→±∞ x→±∞ x x
a1 a0
n−2
+ n−1 )
x x

=⇒ lim p(x) = ±∞(an−1 )


x→±∞

=⇒ lim p(x) = ±∞
x→±∞

From this we can say that ∃ some N, M ∈ R s.t. for x > N , ∃ N ′ ∈ R s.t. p(x) > N ′
and for x < M , ∃ M ′ ∈ R s.t. p(x) < M ′ . Therefore there are no constraints on p(x) and
continuously extends till ±∞.
Since N ′ and M ′ are arbitrary, we will never have a single point where f (xo ) ≥ f (x) or
f (xo ) ≤ f (x) as we can make N ′ and M ′ arbitrarily large or small. Hence, polynomials
with odd degrees do not have a global maximum or minimum.

n
X
(b) Consider a polynomial defined as p(x) = ai xi such that n = 2k, n ̸= 0 and a0 ̸= 0.
i=0

Let’s consider the limit of the function at x → ±∞

n
X
=⇒ lim p(x) = lim ai xi = lim an xn + an−1 xn−1 + an−2 xn−2 + .... + a2 x2 +
x→±∞ x→±∞ x→±∞
i=0
a1 x + a0

n
X an−1 an−2 a2
=⇒ lim p(x) = lim ai xi = lim an xn (1 + + 3
+ .... + +
x→±∞ x→±∞
i=0
x→±∞ an x an x an xn−2
a1 a0
n−1
+ )
an x an x n

Now we can use the multiplication limit law;

n
X an−1 an−2
=⇒ lim p(x) = lim ai x i = lim an xn lim (1 + + + .... +
x→±∞ x→±∞
i=0
x→±∞ x→±∞ an x an x 3
a2 a1 a0
n−2
+ n−1
+ )
an x an x an x n

=⇒ lim p(x) = ∞(1)


x→±∞

=⇒ lim p(x) = ∞
x→±∞

This gives that there exists some N > 0 such that |x| > N where =⇒ p(x) ≥ p(0). We
can considering the restriction on the function to the closed and bounded interval [−N, N ].
Since p(x) is continuous it will attain a minimum value at a some point xo ∈ [−N, N ]
=⇒ p(xo ) ≤ p(0) and since outside this range p(x) ≥ p(0), the absolute minimum of the
polynomial is obtained at x = xo

We can similarly show the property of absolute maximum of the polynomial by consid-
ering a function p′ (x) = −p(x) =⇒ lim p′ (x) = −∞ and considering the restriction of
x→±∞
the polynomial on the closed and bounded interval [−N, N ]. This gives us that the absolute
maximum of the polynomial is obtained at x = xo

Hence, for polynomials of even degrees, they either have an absolute maximum or abso-
lute minimum.

Problem 4. Let N (x) and D(x) be polynomials, let ℓN and ℓD be their leading coefficients, and
let n be the degree of N (x) and d be the degree of D(x).
(a) If d > n, prove that limx→∞ N (x)/D(x) = 0 and limx→−∞ N (x)/D(x) = 0. Make sure
you not only show the steps of your argument but also explain why those steps are valid.
(b) If d = n, prove that limx→∞ N (x)/D(x) = ℓN /ℓD and limx→−∞ N (x)/D(x) = ℓN /ℓD .
Make sure you not only show the steps of your argument but also explain why those steps
are valid.
(c) If d < n, what are limx→∞ N (x)/D(x) and limx→−∞ N (x)/D(x)? You don’t have to prove
your answers, just state them (they will depend on n, d, ℓN , and ℓD ).
n
X d
X
i
(a) Let: N (x) = ai x and D(x) = ai x i
i=0 i=0

Let’s consider the given limit,

n
X
( ai x i )
i=0
=⇒ lim N (x)/D(x) = lim d
x→∞ x→∞ X
( ai x i )
i=0

(ln xn + an−1 xn−1 + .... + a1 x + ao )


=⇒ lim N (x)/D(x) = lim
x→∞ x→∞ (ld xd + ad−1 xd−1 + .... + a1 x + ao )

Since we are given that d > n, we divide the numerator and denominator by xd−1 . Upon
doing this we get the terms in the numerator with powers of x in the denominator whereas
for the denominator we have an x term and a constant followed by terms with powers of x
in the denominator.

(ln /xd−n−1 + an−1 /xd−n + .... + a1 /xd−2 + ao /xd−1 )


=⇒ lim N (x)/D(x) = lim
x→∞ x→∞ (ld x + ad−1 + .... + a1 /xd−2 + ao /xd−1 )
Since we have another continuous function we can just replace the value of x into the
limit of the function. This causes all the terms with powers x in the denominator become
zero and we are left with;

ln
=⇒ lim N (x)/D(x) = lim = if d − n = 1
x→∞ x→∞ ld x + ad−1

else,

0
=⇒ lim N (x)/D(x) = lim = if d − n ̸= 1
x→∞ x→∞ ld x + ad−1

Both these limit values become zero. We get the exact same result if we take x → −∞
instead.

Therefore, lim N (x)/D(x) = 0


x→±∞

(b) Consider the limit, lim N (x)/D(x) and that n = d


x→∞

Taking the equation from the polynomial in the previous question we get;

n
X
( ai x i )
i=0
=⇒ lim N (x)/D(x) = lim d
x→∞ x→∞ X
( ai x i )
i=0

(ln xn + an−1 xn−1 + .... + a1 x + ao )


=⇒ lim N (x)/D(x) = lim
x→∞ x→∞ (ld xd + ad−1 xd−1 + .... + a1 x + ao )

Since we have n = d we can divide the numerator and denominator by either xn or xd .


In the numerator we are left with ln followed by terms with powers of x in the denominator,
and in the denominator we have ld followed by terms with powers of x in the denominator.

(ln + an−1 /x + .... + a1 /xd−1 + ao /xd )


=⇒ lim N (x)/D(x) = lim
x→∞ x→∞ (ld + ad−1 /x + .... + a1 /xd−1 + ao /xd )
Since we have another continuous function we can just replace the value of x into the
limit of the function. This causes all the terms with powers x in the denominator become
zero and we are left with;

(ln ) ln
=⇒ lim N (x)/D(x) = lim =
x→∞ x→∞ (ld ) ld

We get the exact same result if we take x → −∞ instead.

ln
Therefore, lim N (x)/D(x) =
x→±∞ ld

(c) If we apply a method similar to part(a) and divide the numerator and denominator by xn−1 ,
we have two possibilities:

CASE I - Either the denominator entirely becomes zero (when n − d ̸= 1) or,

CASE II - we are left with ld in the denominator (when n − d = 1)

CASE I has an indeterminable form and hence the limit does not exist. For CASE II
if ld = 0, the limit does not exist, while for ld ̸= 0 we either get the value as ∞ or −∞
depending on the values of ld , ln

Problem 5. For all parts of this problem, use the definition of the limit (and our facts about
nearness/neighbourhoods) to prove the statements directly.

(a) Prove the ∞ × Positive Constant Limit Law: if lim f (x) = ∞ and C > 0, then we have
 x→∗
lim Cf (x) = ∞.
x→∗
(b) Prove the ∞ + Bounded Below Limit Law: if lim f (x) = ∞ and g(x) is bounded below
 x→∗
near x = ∗, then lim f (x) + g(x) = ∞.
x→∗
(c) Prove the Composition Two-Sided Limit Lemma: if lim f (x) = b and f (x) ̸= b near x = ∗,
 x→∗
and if lim g(x) = c, then lim g f (x) = c.
x→b x→∗

(a) Given, lim f (x) = ∞ and C > 0


x→∗

Assume g(x) = C, hence, =⇒ lim g(x) = C


x→∗
For g(x), we can say that,for some ϵ > 0 ∃ δ1 > 0 s.t. when |x − ∗| < δ1 , |g(x) − C| < ϵ

Let: ϵ = 2c ;
C
=⇒ |g(x) − C| < 2

=⇒ − C2 < g(x) − C < C


2

C 3C
=⇒ 2
< g(x) < 2

For f (x) we can say that, for every N ∈ R ∃ some δ2 > 0 s.t. when |x − ∗| < δ,
=⇒ f (x) > 2N
C

2N C
Let δ = min(δ1 , δ2 ) =⇒ If |x − ∗| < δ then, f (x) > C
and g(x) > 2

=⇒ f (x)g(x) > ( 2N
C
)( C2 )
=⇒ Cf (x) > N

And hence, we can say that, lim Cf (x) = ∞
x→∗

(b) Given, lim f (x) = ∞ and g(x) is bounded below near x = ∗


x→∗

For g(x) we can say that,∃ some A ∈ R s.t. g(x) ≥ A near x = ∗

For f (x), we can say that, for every N ∈ R ∃ some δ > 0 s.t. when |x − a| < δ,
=⇒ f (x) > N − A

=⇒ f (x) + g(x) > N − A + A

=⇒ f (x) + g(x) > N



And hence, we can say that, lim f (x) + g(x) = ∞
x→∗

(c) Given, lim f (x) = b and f (x) ̸= b near x = ∗, and if lim g(x) = c
x→∗ x→b

For g(x), we can say that for some arbitrary ϵ > 0 ∃ δ > 0 s.t. for |y − b| < δ,
|g(y) − c| < ϵ

For f (x), we can say that for δ > 0, ∃ ϵ1 > 0 s.t. when |x − t| < δ, |f (x) − b| < ϵ1

Let: ϵ1 = δ;
=⇒ |f (x) − b| < δ and, since |y − b| < δ is also true near x = ∗ they are true simulta-
neously.

Set y = f (x);

=⇒ |g(f (x)) − c| < ϵ



Hence, we can say that lim g f (x) = c
x→∗

Problem 6. Remember that you can use statements from earlier pages in the “Limit facts and
rules” Module, including the Boundedness page, to prove statements from later pages.
(a) Prove the Nonnegativity Limit Lemma: if f (x) ≥ 0 near x = ∗ and lim f (x) exists, then
x→∗
lim f (x) ≥ 0.
x→∗
(b) Prove that if lim f (x) = ∞, then lim f (x) does not exist.
x→a x→a

(a) Assume that lim f (x) = −F s.t. F > 0


x→∗

We can say that, for some ϵ > 0 ∃ δ > 0 s.t. when |x − ∗| < δ, |f (x) + F | < ϵ
F
Let, ϵ = 2

F
=⇒ 0 < |f (x) + F | < 2
near x = ∗

=⇒ − F2 < f (x) + F < F


2
near x = ∗

=⇒ − 3F
2
< f (x) < − F2 near x = ∗

But this suggests that, near x = ∗ the function goes to some positive values and some
negative values at the same time, but due to the definition of a function, this not possible.
Hence, if f (x) ≥ 0 near x = ∗ and lim f (x) exists, then lim f (x) ≥ 0
x→∗ x→∗

(b) Assume that lim f (x) exists and that lim f (x) = F
x→∗ x→∗

We can say that, for some ϵ > 0 ∃ δ > 0 s.t. when |x − ∗| < δ, |f (x) − F | < ϵ

Let ϵ = 1

=⇒ 0 < |f (x) − F | < 1 near x = ∗

=⇒ −1 < f (x) − F < 1 near x = ∗


=⇒ F − 1 < f (x) < F + 1 near x = ∗

But this tells us that, the function is bounded between two distinct points and from our
given data (lim f (x) = ∞) the function is only bounded below away from zero near x = ∗,
x→a
but these two statements cannot be true simultaneously. Hence, if lim f (x) = ∞, then
x→a
lim f (x) does not exist
x→a

Problem 7.
1
(a) Using the definition of the limit, prove that lim = ∞ if and only if both lim h(x) = 0
x→∗ h(x) x→∗
and h(x) > 0 near x = ∗.  
1
(b) Using the definition of the limit, prove that lim k = c if and only if lim+ k(y) = c.
x→∞ x  y→0
(c) If lim f (x) = ∞ and lim g(x) = c, prove that lim g f (x) = c. (Hint: although we
x→∗ x→∞ x→∗
can’t directly apply the Composition Limit Laws and Lemmas, try to use parts (a) and (b)
to change to a situation where we can.)
1
(a) To prove, lim = ∞ if and only if both lim h(x) = 0 and h(x) > 0 near x = ∗
x→∗ h(x) x→∗

We can say that, ∀ ϵ > 0 ∃δ > 0 s.t. |h(x) − 0| < ϵ as |x − ∗| < δ, if lim h(x) = 0
x→∗

1
Let ϵ > 0, ∃N ∈ R, s.t. N > ϵ

1
Also, ∃δ s.t. h(x) < N
as |x − ∗| < δ

=⇒ h(x) < ϵ
1 1 1
Also, since h(x)
>N > ϵ
> 0, =⇒ h(x)
> 0, =⇒ h(x) > 0

 
1
(b) To prove that, lim k = c if and only if lim+ k(y) = c
x→∞ x y→0

We can say that, ∀ ϵ > 0 ∃δ > 0 s.t. |k(y) − c| < ϵ as y − 0 < δ, if lim+ k(y) = c
y→0

Let: ϵ > 0, ∃ M ∈ R s.t. |k( x1 ) − c| < ϵ as x > M


1
Let: δ = M

1
=⇒ x

1
Both x
< δ and y < δ are true simultaneously near x = ∗
1
Set x
=y
=⇒ |k(y) − c| < ϵ


(c) To prove that, if lim f (x) = ∞ and lim g(x) = c, then lim g f (x) = c
x→∗ x→∞ x→∗

For f (x),we can say that, for each N ∈ R ∃ δ > 0 s.t. f (x) > N as |x − ∗| < δ

For g(y), we can say that, for ϵ > 0 ∃ M ∈ R s.t. |g(y) − c| < ϵ as y > M

Let: M = N ;

=⇒ f (x) > M, y > M and since both the statements are true near x = ∗ they are true
simultaneously as well.

Set y = f (x);

=⇒ |g(f (x)) − c| < ϵ

Problem 8 is a bonus problem! You can get full marks for this assignment even if you skip
Problem 8 completely. Any marks you receive for Problem 8 will be added to your score.
Problem 8. For this problem, we should never write down any expression of the form bz where z
is irrational; the context of this problem is to extend the definition of br for rational numbers r to
all real numbers using the proposed extension Eb (x).
Let b > 1. For any real number z, define the set Sb (z) = {br : r ∈ Q, r < z}, and define Eb (z) to
be the least upper bound for Sb (z) in the real numbers. Prove that Eb (x + y) ≥ Eb (x)Eb (y) for
all real numbers x and y. (Hint: if w is an upper bound for Sb (x + y) and r < y is any rational
number, first prove that w/br ≥ Eb (x), then prove that w ≥ Eb (x)Eb (y).)

your answer goes here

You might also like