2

1 Sequences

Exercise 1.1. Consider sequences {an }n∈N , {bn }n∈N in Rm with

lim an = L, lim bn = M.
n→∞ n→∞

Show the following

(a) lim (an + bn ) = L + M.
n→∞

(b) For m = 1, lim (an · bn ) = L · M.

n→∞

Hint: an bn − LM = (an − L)(bn − M ) + (an − L)M + L(bn − M ).

an L
(c) lim = , if bn 6= 0 for all n and M 6= 0.
n→∞ bn M

Solution:

(a) Let ϵ > 0 be arbitrary, but fixed. Then there are N1 , N2 such that
ϵ
kan − Lk < for n ≥ N1
2
ϵ
kbn − M k < for n ≥ N2
2
Then for n ≥ max{N1 , N2 } we have

k(an + bn ) − (L + M )k = k(an − L) + (bn − M )k

ϵ
≤ kan − Lk + kbn − M k ≤ 2 · =ϵ
2

(b) Let ϵ > 0 be arbitrary, but fixed. We may assume that ϵ < 1. Then there are N1 , N2
such that
ϵ ϵ
|an − L| < min{ , 3M } for n ≥ N1
3
|bn − M | < min{ 3ϵ , 3L
ϵ
} for n ≥ N2

© Maurice Koster, UvA, 2024

 3

so that

|an · bn − L · M | = |(an − L)(bn − M ) + (an − L)M + L(bn − M )|

≤ |(an − L)(bn − M )| + |(an − L)M | + |L(bn − M )|
ϵ ϵ ϵ ϵ
< · + ·M + · L < ϵ.
3 3 3M 3L

(c) Take the continuous function f (x) = x1 , {cn } = {f (bn )}. Then

1
lim cn = f ( lim bn ) = f (M ) = .
n→∞ n→∞ M
Now apply the result above:

an 1
lim = lim an · cn = L · .
n→∞ bn n→∞ M

Exercise 1.2.
1
(a) Calculate for each x ∈ R: lim .
n→∞ 1 + x2n

(b) Draw the graph of the function f : R → R that is defined by

1
f (x) = lim .
n→∞ 1 + x2n

Solution:

(a) We have 
 0 if |x| < 1
lim x2n = 1 if x ∈ {−1, 1}
n→∞ 
∞ if |x| > 1.
Then as a result 
1  1 if |x| < 1
lim = 1
if x ∈ {−1, 1}
n→∞ 1 + x2n  2
0 if |x| > 1.

For the graph of the function, see the figure below:

© Maurice Koster, UvA, 2024

 4

1.0
0.8
0.6
0.4
0.2

-2 -1 1 2

Notice that for each fixed value of n, the function gn = x 7→ 1

1+x2n
is continuous – the
limiting value is not. In the figure we plotted f and g2 .

Exercise 1.3.
(a) Use a proof by induction to show that for each n ∈ N it holds: n! ≥ 2n−1 .
(b) Show that for each n ∈ N we have

1 1 1
1+ + + ... + < 2.
2! 3! n!
Hint: First show that for all n
1 1 1
1+ + + ... + < 2 − 2−n+1 ,
2! 3! n!
and use part (a) here.
(c) Show that
 
1 1 1 1 1
1 + + + . . . = lim 1 + + + . . . +
2! 3! n→∞ 2! 3! n!

exists.

Solution:
• Proof by induction. Let P (n) : n! ≥ 2n−1 .
– n = 1: 1! = 1 = 20 , so P (1) is true.
– Now suppose P (n) is true. For n ≥ 1 we have

(n + 1)! = n! · (n + 1) ≥ 2n−1 (n + 1) ≥ 2n−1 · 2 = 2n ,

so P (n + 1) is true.

© Maurice Koster, UvA, 2024

 5

• We have for all n ∈ N that

Xn
1 Xn
1 1 − ( 12 )n
≤ = = 2 − 2−n+1 < 2.
k=1
k! k=1
2 k−1 1− 2 1

• The sequence of sums

1 1
sn = 1 + + ... +
2! n!
is clearly increasing. Moreover, by part (a), it is also bounded, and therefore con-
verging. Later we will see that its sum equals e − 1.

Exercise 1.4. *

√
n
(a) Calculate lim n
n→∞

(b) Show that for all a, b ∈ N = {1, 2, 3 . . .} : ab ≥ a + b − 1.

(c) Use this result to show that n! ≥ nn/2 for all n ∈ N.
Hint: (n!)2 = (1 · 2 · · · n)2 = (n · 1)((n − 1) · 2) · · · (1 · n).
Which of the following sequences converge?
√
(d) an = n n!. Hint: Use the above exercise here.
p
n n
√
(e) bn = n!
qp
n n √ n
(f ) cn = n!

Solution:
(a) Take log first: we have

√ ln n
lim ln( n n) = lim = 0.
n→∞ n→∞ n

Hence the limit is equal to 1.

(b) Rewrite the equation

ab ≥ a + b − 1 ⇔ (a − 1)b ≥ a − 1,

© Maurice Koster, UvA, 2024

 6

which is clearly true as a, b ≥ 1.

(c) We have

(n!)2 = (1 · 2 · · · n)2 = (n · 1)((n − 1) · 2) · · · (1 · n)

Yn Yn
= (n − k + 1)k ≥ ((n − k + 1) + k − 1) = nn .
k=1 k=1

Now take square roots on both sides to get the result.

√ √
(d) We have that n! ≥ nn/2 so that n n! ≥ n → ∞ for n → ∞. Therefore the limit
does not exist.
(e) We have √ √ √ √ √
1· 2··· n ≤ ( n n)n = n.
n n n n
n! =
Now q
n √ √
1≤ n! ≤
n n
n.
The Squeeze Theorem implies that {bn } converges with limit 1.
(f ) Replicate the reasoning from the above exercise:
rq
√ q
n n √ √
1≤ n! ≤ n≤
n n n n
n,

and use the Squeeze Theorem again to deduce that lim cn = 1.

n→∞

Exercise 1.5. Show the following

(a) The sequence {an }n∈N defined by an = n!
nn
converges.
nn
(b) Show that the sequence {bn }n∈N defined by bn = n!
diverges.

Solution:
(a) For n > 1 we have

n! 1 · 2... · n 1·2 2
0< = ≤ = .
n n n · n · ... · n n·n n

© Maurice Koster, UvA, 2024

 7

Now use the Squeeze theorem to conclude that

n!
lim = 0.
n→∞ nn

(b) We have for n > 1

nn n2
bn = ≥ → ∞ for n → ∞
n! 2

Exercise 1.6. Consider the sequence {an }n∈N , recursively defined by


a1 = a > −6
√
an+1 = an + 6 for n ≥ 1

Here a > −6 is a constant. Is the corresponding sequence converging? If so, what is its
limit?

Solution:
We will show that for all a the sequence {an }n∈N is convergent. Notice that
√
a2 ≥ a1 ⇔ a + 6 ≥ a ⇔ a ∈ [−6, 3]

Case 1: a ∈ [−6, 3].

• We first show that the sequence is monotonically increasing, i.e., for all n ∈ N we
have an+1 ≥ an .
Proof: We use a proof by mathematical induction here. Let P (n) : an+1 ≥ an .
Then P (1) is true. Now suppose P (n) is true for some n ∈ N. Then
p √
an+1 ≥ an ⇒ an+1 + 6 ≥ an + 6 ⇒ an+2 ≥ an+1 .

So P (n + 1) follows from P (n). This completes the proof. Similarly, one may
show that {an }n∈N is (strictly) monotonically decreasing if a > 3.
• The sequence is bounded above by 3. Again we will show this using a proof by
mathematical induction. Let P (n) : an ≤ 3. Clearly we have a1 ≤ 3. Now
suppose an ≤ 3 for some n. Then
√ √
an ≤ 3 ⇒ an + 6 ≤ 3 + 6 = 3.

© Maurice Koster, UvA, 2024

 8

This completes the proof.

• We conclude that the sequence is monotonically increasing and bounded from above,
so it must converge by virtue of the Monotonic Convergence theorem for sequences.
We calculate its limit as follows:

L = lim an = lim an+1

n→∞ n→∞
√ q √
= lim an + 6 = lim an + 6 = L + 6.
n→∞ n→∞

Solving for L gives L = 3.

Case 2: a > 3. This can be treated as Case 1. You will find that the sequence is
(strictly) monotonically decreasing, and bounded from below by 3. Therefore, also in
this case, the sequence converges. Then the limit is calculated in the same fashion,
L = 3.

Exercise 1.7. Define a sequence {an }∞

n=1 in R by
2

 n 
n n 1/n
an = , (e + n) .
1+n

Does the sequence converge? If so, what is the limit?

Solution:
The sequence {an }n∈N converges if and only if its two components converge in R. We
have
ln(n) − ln(n + 1)
lim ln(an,1 ) = lim
n→∞ n→∞ 1/n
1
− n+1 1
−n2
= lim n
= lim = −1.
n→∞ − n12 n→∞ n(n + 1)

Then
lim an,1 = lim eln(an,1 ) = e−1 .
n→∞ n→∞

So the first coordinate converges. Now the second coordinate. Again we take the log and

© Maurice Koster, UvA, 2024

 9

apply de l’Hôpital:

ln(en + n)
lim ln(an,2 ) = lim
n→∞ n→∞ n
en + 1 en
= lim n = lim n = 1,
n→∞ e + n n→∞ e + 1

so that
lim an,2 = e.
n→∞

We see that the sequence converges and

lim an = (e−1 , e).

n→∞

Exercise 1.8. Consider the sequence {an }n∈N in R2 that is given by an = (sin n, cos n).
Does this sequence have a converging subsequence?

Solution:
This sequence is bounded, as kan k = 1 for all n. Therefore, by the Bolzano-Weierstrass
theorem, it has a converging subsequence.

Exercise 1.9. Consider the sequence {an }n∈N in R2 that is given by

 
π(n+1)
an = en sin( πn
2
), sin( 2
) .

(a) Show that the sequence {an }n∈N is unbounded.

(b) Does it have a converging subsequence?

Solution:
(a) For n = 1 + 4k and k = 0, 1, 2, . . . we have sin( nπ2
) = sin( 21 π + 2kπ) = 1 and
sin( (n+1)π
2
) = sin(π + 2kπ) = 0. So kan k = k(en , 0)k = en → ∞ for k → ∞.

© Maurice Koster, UvA, 2024

 10

(b) Note that when n = 4k for k = 0, 1, . . ., we get

an = e4k sin( 12 π(4k)), sin( 12 π(4k + 1))


= e4k sin(2kπ), sin(2kπ + 21 π) = (0, 1).

So {a4n }n∈N forms a converging subsequence.

© Maurice Koster, UvA, 2024