Mathematics I

Part 2: Single Variable Calculus

Week 10

1. Real Sequences
Definition 1 (Real Sequence). A sequence of real numbers or a real sequence is a
function f : N → R defined by f (n) = xn for each n ∈ N .
We denote a sequence by (xn ) or {xn }, where xn is the nth term of the sequence.
There are different ways of expressing a sequence.
(1) Sequence defined by listing: (1, 4, 8, 11, 52, . . .).
(2) Sequence defined by rule: (xn ), where xn = 3n2 for all n ∈ N.
(3) Sequence defined recursively: (xn ), where x1 = 4 and xn+1 = 2xn − 5 for all
n ∈ N.
Note: A sequence (xn ) is always an infinite collection of elements. However, the set
of all terms of a sequence denoted by {xn : n ∈ N} may be finite or infinite.

The sequence (an ), where an = (−1)n , is −1, 1, −1, 1, −1, . . .. However {an : n ∈
N} = {1, −1} is finite. In a similar way, the sequence (bn ), where bn = n1 , is 1, 12 , 13 , . . .
and {bn : n ∈ N} = {1, 12 , 31 , . . .} is infinite.

Following are some examples of real sequences (an ):

(1) an = 2, constant sequence.
n
(2) an = 1 − (−1)
2n
.
n
(3) an = 1 + (−1) , oscillating sequence.
(4) an = an−1 + an−2 with a1 = 1 = a2 Fibonacci sequence.
Definition 2 (Convergent sequence). A sequence (xn ) is said to be convergent if there
exists ℓ ∈ R such that for every ε > 0, there exists m ∈ N satisfying
|xn − ℓ| < ε for all n ≥ m.
In this case, we say that ℓ is a limit of (xn ), and we write lim xn = ℓ or xn → ℓ.
n→∞

Note: For any x ∈ R, there is a unique m ∈ Z such that x − 1 < m ≤ x. We denote

this m by [x], called the greatest integer which is not greater than x, or the integral
part of x. For example, [3.859] = 3, [−3.581] = −4, [4] = 4, and [−23] = −23.
1
Example: Let xn := n
for all n ∈ N. Let ε > 0 be given. We want to find a
m ∈ N such that
1
− 0 < ε for all n ≥ m.
n
Note that n1 < ε implies n > 1ε . By Archimedean property, for any ε > 0 there exists
1
m ∈ N such that m > . We choose m = [ 1ε ] + 1. Thus for all n ≥ m, we have n1 < ε,
ε
and hence xn = n1 → 0.
1
2

Theorem 1.1. The limit of a convergent sequence is unique.

Proof. Let (xn ) be a convergent sequence. Suppose xn → ℓ1 and xn → ℓ2 . For any
ε > 0, there exists m1 , m2 ∈ N such that
ε
|xn − ℓ1 | < for all n ≥ m1
2
and
ε
|xn − ℓ2 | < for all n ≥ m2 .
2
We choose m = max{m1 , m2 }. Then
|ℓ1 − ℓ2 | = |(xn − ℓ2 ) − (x2 − ℓ1 )| ≤ |xn − ℓ2 | + |x2 − ℓ1 | < ε
for all n ≥ m. Therefore, we must have ℓ1 = ℓ2 . □
Example: Let a ∈ R be fixed and xn := a for all n ∈ N. Then xn → a. For any ε > 0,
we can consider m = 1 satisfying |xn − a| = |a − a| = 0 < ε for all n ≥ 1 = m.
2
Example: Let an := n2 +1
for n ∈ N. Note that
2 2
− 0 ≤ 2 for all n ∈ N.
n2 +1 n
√ √
For a given ε > 0, we choose m ∈ N such that m > √2 , say m = [ √2ε ] + 1. Then
ε

2 2 2
−0 = 2 ≤ 2 < ε for all n ≥ m.
n2 +1 n +1 n
Thus an → 0.
Remark 1.2. If a sequence does not converge to any real number ℓ, we say that the
sequence diverges or it is divergent. A sequence (xn ) is not convergent to ℓ if there
exists a ε > 0 with |xn − ℓ| ≥ ε and n ≥ m for all m ∈ N.
Example: The sequence an = (−1)n for all n ∈ N is divergent, that is, it is not
convergent. If possible, let an → a, then for ε = 21 , there exists m ∈ N such that
|(−1)n − a| < ε for all n ≥ m.
Note that
1 1
2 = |a2n − a2n+1 | = |(a2n − a) + (a − a2n+1 )| ≤ |a2n − a| + |a − a2n+1 | < + =1
2 2
for all n ≥ m, which is a contradiction. Thus an is a divergent sequence.
Definition 3 (Bounded sequence). For a sequence (xn ), if the set {x1 , x2 , x3 , . . .} is
bounded, then the sequence (xn ) is said to be bounded. Equivalently, the sequence (xn )
is bounded if there is a positive number M such that |xn | ≤ M for all n ∈ N. If (xn ) is
not bounded then it is said to be unbounded.
1 n2
The sequences xn = (−1)n , yn = n
are bounded. However, an = n, bn = n+1
are
unbounded sequences.
Theorem 1.3. Every convergent sequence is bounded.
 3

Proof. Let (xn ) be a convergent sequence that converges to ℓ. For any ε > 0, there
exist m ∈ N such that
|xn − ℓ| < ε for all n ≥ m.
Then
ℓ − ε < xn < ℓ + ε for all n ≥ m.
We choose m = min{x1 , x2 , . . . , xm−1 , ℓ − ε} and M = max{x1 , x2 , . . . , xm−1 , ℓ + ε},
then
m ≤ xn ≤ M for all n ∈ N.
Thus (xn ) is a bounded sequence. □
Remark 1.4. The converse of the above theorem is not true. There are bounded
sequences that are not convergent. For example, an = (−1)n is a bounded sequence, but
not convergent.
Remark 1.5. From the above theorem, we can conclude that if a sequence is unbounded,
then it is not convergent. For instance, for the sequence an := (−1)n n, we have |an | =
n. Thus the sequence (an ) is unbounded, and hence it is not convergent.
Theorem 1.6. Let (an ) and (bn ) be two convergent sequences with an → a and bn → b.
Then
(1) (an + bn ) → (a + b).
(2) (an − bn ) → (a − b).
(an bn ) → (ab).
(3) 
(4) abnn → ab if b ̸= 0.

(5) an ≤ bn for all n ∈ N implies a ≤ b.

Proof. Since an → a and bn → b, for any ε > 0, there exist m1 , m2 ∈ N such that
ε
|an − a| < for all n ≥ m1
2
and
ε
|bn − b| < for all n ≥ m2 .
2
(1) Note that
|(an + bn ) − (a + b)| ≤ |an − a| + |bn − b| < ε
for all n ≥ m = max{m1 , m2 }. Therefore, (an + bn ) → (a + b).

(2) The proof is almost identical to (1).

(3) We have
|an bn − ab| = |an (bn − b) + (an − a)b| ≤ |an ||bn − b| + |an − a||b|.
Since (an ) is convergent, and hence it is bounded. There exists a number K such that
|an | ≤ K for all n ∈ N. Since b is a fixed real number, we must have a number M such
that |b| ≤ M . Thus
Kε M ε (M + K)
|an bn − ab| < + = ε = ε1
2 2 2
4

for all n ≥ m = max{m1 , m2 }. As a result, (an bn ) → (ab).

(4) The proof is almost identical to (3).

(5) We prove the statement by contradiction. If possible, let b < a. We have

ε ε ε ε
a − < an < a + and b − < bn < b +
2 2 2 2
for all n ≥ m = max{m1 , m2 }. We choose ε = a − b > 0. As a result,
ε a−b a−b ε
b+ =b+ =a− =a− .
2 2 2 2
In this case, we see that
ε ε ε ε
b− < bn < b + = a − < an < a +
2 2 2 2
for all n ≥ m. This implies bn < an for all n ≥ m, which is a contradiction. Thus we
must have a ≤ b. □
Theorem 1.7 (Sandwich Theorem). Let there exists m ∈ N such that an ≤ cn ≤ bn
for all n ≥ m. If limn→∞ an = c = limn→∞ bn , then limn→∞ cn = c.
Proof. Since an ≤ cn ≤ bn for all large n ∈ N, we must have
lim an ≤ lim cn ≤ lim bn
n→∞ n→∞ n→∞
or
c ≤ lim cn ≤ c.
n→∞
Thus we must have limn→∞ = c. □
Remark 1.8. Suppose there exists m ∈ N such that |bn | ≤ cn for all n ≥ m. If cn → 0,
then as a consequence of the Sandwich theorem, we have bn → 0.
Example: Let a ∈ R and an := an for all n ∈ N. Then (an ) is convergent if and only
if −1 < a ≤ 1.

Solution: If a = 0, then it is clear that 0n = 0 and it is convergent.

If a = 1, then an := 1n → 1.
If a = −1, then an := (−1)n is an oscillating sequence and it is not convergent.
1
Let 0 < |a| < 1, then we consider r = |a| . Thus r > 1, and hence we take r = 1 + h
for some h > 0. Using the Binomial Theorem, we have
rn = (1 + h)n = 1 + nh + · · · + hn .
As a result, rn > nh, and hence
1 1
0 ≤ |an | = |a|n =n
< → 0.
r nh
Thus the Sandwich Theorem implies an → 0.
If |a| > 1, we consider s := |a|. Then s > 1, and we take s = 1 + h for some
h > 0. Using the above argument, we see that |an | = sn > nh for all n ∈ N. Hence the
sequence (an ) is unbounded, and hence it is not convergent.
 5

n3 +3n2 +4
Example: Let an := n4 +8n2 +2
for all n ∈ N. Note that
n3 + 3n2 + 4 1 3 4
0≤ 4 2
≤ + 2 + 4 → 0.
n + 8n + 2 n n n
Therefore, an → 0.
On can also see that
1
n3 + 3n2 + 4 n
+ n32 + n44 0
an := 4 2
= 8 2 → = 0,
n + 8n + 2 1 + n2 + n4 1
1
due to the fact that n
→ 0.
1
Example: For the sequence an := n
sin( n1 ), we have
1
0 ≤ |an | ≤ → 0,
n
giving an → 0.
Definition 4. A sequence (an ) is said to be increasing if an ≤ an+1 for all n ∈ N, that
is, a1 ≤ a2 ≤ a3 · · · .
A sequence (an ) is said to be decreasing if an ≥ an+1 for all n ∈ N, that is, a1 ≥ a2 ≥
a3 ≥ · · · .
A sequence is monotonic if it is either increasing or decreasing.
The sequence an = n1 is a decreasing sequence and bn = n + n1 is an increasing se-
quence. Thus both are monotonic sequences. However xn = (−1)n is neither increasing
nor decreasing, and hence it is not monotonic.
Theorem 1.9. An increasing sequence (an ) that is bounded above is convergent, and
limn→∞ an = sup{an : n ∈ N}.
Proof. Let (an ) be an increasing sequence that is bounded above. Therefore, there
exists a number M such that
|an | ≤ M for all n ∈ N.
Thus well-ordering property of real numbers provides that sup{an : x ∈ N} exists in
R, say s = sup{an : x ∈ N}. As a result, s − ε is not an upper bound for (an ) for any
ε > 0, and hence there exists m ∈ N such that
am > s − ε.
Since (an ) is increasing, we must have
an ≥ am for all n ≥ m.
Therefore,
s − ε < am ≤ an ≤ s < s + ε for all n ≤ m.
Thus for any ε > 0, there exist m ∈∈ N such that
|an − s| < ε for all n ≥ m,
and hence an → s. □
6

Remark 1.10. A decreasing sequence (an ) that is bounded below is convergent, and
limn→∞ an = inf{an : n ∈ N}.
 
Example: Let a1 := 32 and an+1 := 21 an + a2n for n ∈ N. It is easy to see that
an > 0 for all n ∈ N, and hence (an ) is bounded below by 0. Note that
2 − a2n
 
1 2
an+1 − an = an + − an = for all n ∈ N.
2 an 2an
Thus (an ) is decreasing if and only if a2n ≥ 2 for all n ∈ N. It is clear that a21 = 94 > 2
2 2
and a2n+1 − 2 = (an4a−2)
2 ≥ 0 for all n ∈ N. As a result, the sequence (an ) is a decreasing
n
sequence that is bounded below, and hence must be convergent. Suppose an → ℓ, then
an+1 → ℓ. Therefore,
 
1 2 1 2
ℓ = lim an+1 = lim an + = (ℓ + )
n→∞ n→∞ 2 an 2 ℓ
giving
2ℓ2 = ℓ2 + 2,
√
and hence
√ ℓ = 2 as (an ) is a sequence of positive numbers. This concludes that
an → 2.
Definition 5. Let (xn ) be a real sequence. If (nk ) is a sequence of positive integers
such that n1 < n2 < n3 < · · · , then (xnk ) is called a subsequence of (xn ).
For example xn = 1 and yn = −1 are two subsequences of an = (−1)n . Similarly,
1 1
zn = 2n and wn = 2n+1 are subsequnces of cn = n1 .
Theorem 1.11. If a sequence (xn ) converges to ℓ, then every subsequence of (xn ) must
converge to ℓ.
Proof. Let (xnk ) be a subsequence of the sequence (xn ). Since (xn ) → ℓ, for any ε > 0,
there exists m ∈ N such that
|xn − ℓ| < ε for all n ≥ m.
In particular, we have
|xnk − ℓ| < ε for all nk ≥ m.
Thus xnk → ℓ. □
Remark 1.12. If a sequence has two subsequences that converge to two different limits,
then the sequence is divergent. For example, the sequence ((−1)n ) has two subsequences
((−1)2n ) → 1 and ((−1)2n+1 ) → −1. Thus ((−1)n ) is divergent.