MATH1010 — Assignment 3

Austin Kim
February 14, 2025

2.2
2.
(a.)
2n+1 2
lim 5n+4 = 5
2 2n+1
We need that for all ε > 0, ∃n ∈ N → ∀m ≥ n, 5 − 5n+4 < ε.

2(5n + 4) 5(2n + 1) 10n + 8 − 10n − 5

− =
5(5n + 4) 5(5n + 4) 5(5n + 4)
3
=
25n + 20
1
<
25n + 20
1
Then, if n = 25ε − 45 , we have

1 1
1 = 1
− 45 + 20


25 25ε ε − 20 + 20
1
= 1
ε
= |ε|

2 2n+1 1
Thus, 5 − 5n+4 < |ε|. To make n ∈ N guaranteed, we can simply set n = ⌈ 25ε − 45 ⌉.

(b.)

2n2 2n2
0− = −
3 + n3 3 + n3
2n2
=
3 + n3
2n2
≤ 3
n
2
=
n
2
To have n < ε, we need n = 2ε . Then n = ⌈ 2ε ⌉ suffices.

1
(c.)

sin(n2 ) sin(n2 )
0− √ = √
3
n 3
n
1
≤ √3
n

1
Then n = ε3 → n = ⌈ ε13 ⌉ suffices.

4.
(a.)
This is possible. Setting (−1)n and ε = 1.9 disproves convergence.

(b.)
This is impossible. For l ̸=, there is no n ∈ N such that |an − l| < ε for n large, since no matter
how large n is, there has to be another infinity ones after it (even if n is large, there is still only a
finite number of 1s behind it, so there must be another infinity after it). Thus, ε = l−1 2 suffices
to disprove convergence to l.

(c.)
This is possible. Let there be a sequence with alternating 1s and 2s, i.e. 1, 2, 1, 1, 2, 2, 1, 1, 1, 2, 2, 2, . . ..
I’m really not sure how to put this into set-builder notation. Just harvest my fucking soul already,
man.

6.
Let l, m unique limits for sequence an . Then taking ε = m−l
2 suffices to show that the sequence
cannot be within ε of both m and l. If you really want more than geometric intuition, here it is:
m−l
|m − xn | <
2
m−l
m − xn <
2
m l
+ < xn
2 2
but
m−l
|l − xn | <
2
m−l
xn − l <
2
m l
xn < + .
2 2
While the computation is for when m > xn > l, this is true without loss of generality.

8.
(a.)
Yes. M = 2 suffices.

2
(b.)
Yes. If there is a nonzero term, then at some point after it, there must be another zero. After
that, the process repeats infinitely, cycling every M terms.
e.g. xn′ . . . 0 . . . xn′′ . . . 0 . . . etc.

(c.)
No. Let it be that we have one nonzero term, then a zero, then two nonzero terms, then a zero,
etc., so that we get to n nonzero terms, then n + 1, etc. Then there is no one M that suffices,
since next, there will be the M + 1-term long string of nonzero numbers.

(d.)
A sequence is not zero-heavy if there is no M ∈ N such that for any N ∈ N, there exists n ∈ N
such that N ≤ n ≤ N + M where xn = 0.

2.3
3.
Statement: If xn ≤ yn ≤ zn ∀ n ∈ N and lim xn = l = lim zn , then lim yn = l.
Proof. Since both xn and zn converge to l, we have that for any ε > 0, there is n ∈ N and n′ ∈ N
for xn and zn , respectively, such that |l − xn | < 2ε and |l − zn′ | < 2ε . Then at m = max{n, n′ },
[xm , zm ] ⊆ (l − ε, l + ε). Then ym ∈ (l − ε, l + ε).

4.
(a.)

 
1 + 2an 1 2an
lim = lim +
1 + 3an − 4a2n 1 + 3an − 4a2n 1 + 3an − 4a2n
1 2an
= lim +
(−4an − 1)(an − 1) (−4an − 1)(an − 1)
  
1 1 1 an − 1 1
= lim · +2 · +
−4an − 1 an − 1 −4an − 1 an − 1 an − 1
  
1 1 1 1
= lim · +2 · 1+
−4an − 1 an − 1 −4an − 1 an − 1
1 1 2 2 1
= lim · + + ·
−4an − 1 an − 1 −4an − 1 −4an − 1 an − 1
1 1 2
= lim 3 · · +
−4an − 1 an − 1 −4an − 1
1 1 1
= 3 lim · + 2 lim
−4an − 1 an − 1 −4an − 1
= 3 · (−1) · (−1) + 2 · (−1)
=3−2
=1

3
(b.)

(an + 2)2 − 4 (an + 2)2

lim = lim
an an − a4n
a2n + 4an + 4 4
= lim −
an an
a2 4an 4 4
= lim n + + −
an an an an
= lim an + 4
=4

(c.)

2
an +3 2 + 3an
lim 1 = lim
an +5 1 + 5an
2 3an
= lim +
1 + 5an 1 + 5an
2 3an
= lim + lim
1 + 5an 1 + 5an
=2+0
=2

7.
(a.)
xn = (−1)n , yn = (−1)n+1 → (−1)n + (−1)n+1 = 1 · (−1)n + (−1) · (−1)n = (−1)n (1 + (−1)) = 0.

(b.)
Fails. xn is ε-close to some l past the nth term, but there is no l′ which xn + yn gets ε-close to,
since yn cannot stay ε-close to |l′ − l| or, in fact, to any number.

(c.)
Fails. If an = 1, then lim b1n = lim abnn = lim an
lim bn = 1
lim bn .

(d.)
Fails. Past some n, trying to say an − bn is bounded results in the proposed bound N having
this horrible experience that will almost certainly earn a one-star rating on TripAdvisor:

N ≥ a n − bn
> an − (l + ε)
→ an < N + l + ε

But since an is unbounded, eventually an > N + l + ε.

(e.)
Past some N , we have am bm → (l + ε) · bm . bm doesn’t converge, so certainly no l + ε multiple
will.

4
9.
(a.)
Statement: Let an a bounded sequence, and bn a sequence such that bn → 0. Then lim(an bn ) =
0.
Proof. There exists N ∈ N such that all an < N . Then, past m ∈ N, since all |bm | < ε, we have

|an · bn | < |an · ε|

= ε|an |
ε
at most ε · N . Then any k ∈ N such that |bk | < N results in

|ak · bk | = |ak | · |bk |

ε
< |ak | ·
N
≤ε

The Fraudebraic Fraud Fraudrem fails here because it is a fraud. Or because the proof for that
theorem assumes that a exists for lim an = a, where it might not generally be true for these
hypotheses.

(b.)
It goes to

|an · bn | = |an · (l + ε)|

= |an | · |l + ε|
≤ |an |(|l| + |ε|)
= |an ||l| + |an ||ε|

not guaranteed to converge, since |an ||l| may do something like oscillation (e.g. if an = sin(n),
then that term is |l sin(n)| not always in | 2l |).

(c.)
Kind of just relabel the an sequence here to the bn sequence from part a. Then certainly we
fulfill the condition for lim an = 0, since an = 0. Next, in the Algebraic Limit Theorem, we
also assume that bn converges to b, so bn is bounded like an from part a is, so we satisfy the
hypotheses from part a to arrive that the conclusion that the product goes to 0 = 0 · b = a · b.

2.4
3.
(a.)
We see that the sequence is monotonically increasing, and also that the sequence is bounded
p √ √
above by 2, since 2 + 2 < 2 + 2 = 2. Inductively, this is true for every single additional
nested addition, so for the√nth term of the sequence, we simply run the same inequality where
we replace the innermost 2 with 2 instead, then √ repeatedly compute until we arrive at 2.
As for the limit, consider that each term xn =√ 2 + xn−1 . Then,
√ if there is α the
√ supremum of
the sequence, then we have α = lim xn = lim 2 + xn−1 = 2 + α. Then α = 2 + α ⇒ α2 =
2
2+α ⇒ √ α − α − 2 = 0 ⇒ α = 2, −1. But the limit cannot have decreased from its starting
point, 2 > −1, since the sequence is monotonically increasing, so α = 2.

5
(b.)
For the same reasoning as above, it converges: it is bounded above by 2, and is also monotonically
increasing.
For the limit, let√α be the √ supremum of the√ set of values in the sequence. Then we have
α = lim an = lim 2an−1 = 2α. Then α = 2α ⇒ α2 − 2α = 0 ⇒ α = 0, 2. But clearly the
limit cannot be 0, so the limit is 2.

5.
(a.)
O1 First, let n = 2. Then, we have
  2
1 2
x22 = x1 +
2 x1
 2
1
= (2 + 1)
2
 2
1
= ·3
2
 2
3
=
2
9
=
4
≥2

Thus, it is true for a base case of n = 2. Then, assume it is true for n = k, k ∈ N. Then, we
must prove that it is true for n = k + 1.
1 2 1
x2k+1 = x +1+ 2
4 k xk
(x2k )2 + 4x2k + 4
=
4x2k
(2 + δ)2 + 4(2 + δ) + 4
=
4(2 + δ)
4 + 4δ + δ 2 + 8 + 4δ + 4
=
8 + 4δ
16 + 8δ + δ 2
=
8 + 4δ
16 + 8δ
>
8 + 4δ
=2

6
O
Now, step 2.
2
 
1 2
xn−1 − xn = xn−1 − xn−1 +
2 xn−1
xn−1 1
= xn−1 + +
2 xn−1
1 1
= xn−1 +
2 xn−1
x2n−1 + 2
=
2xn−1
22 + 2
≥
2xn−1
6
=
2xn−1
3
=
xn−1
≥0

O3 First, let l = inf{(x ) }, l ̸= 0 (because there’s no way for the limit to be under √2, judging
by the value of every term as proven in O
n n
1 ). Then, we have
 
1 2
l = lim xn = lim xn−1 +
2 xn−1
 
1 2
= l+
2 l
1 1
l= l+
2 l
2 1 2
l = l +1
2
1 2
l =1
2
l2 = 2
√
l=± 2
√
l= 2

(b.)
√ 1

c

In order to make the limit c, we make xn+1 = 2 xn + xn . Checking this gives
 
1 c
l = lim xn+1 = lim xn +
2 xn
1 c
= l+
2 l
1 c
l= l+
2 l
c
2l = l +
l
2l2 = l2 + c
l2 = c
√
l=± c
√
⇒l= c