Lecture 4: Monotone sequence and subsequence

Aritra Dey

Octobar 2024

We end our last lecture asking the question

Boundedness + (??) ⇒ Convergence.

In this lecture, we are going to answer the question. But before moving to the actual
topic let us recall something from our previous lecture.

Defination 1 :

Let S be a set in R be bounded above, then m is called the least upper bound
(supremum) if m ≥ s, ∀s ∈ S and if m′ is some other upper bound, then m < m′

We can write the above definition in the following way, as m is the least among all
the upper bounds, then we can ensure m − ϵ is not an upper bound. Now some number
is not an upper bound, means what? There must exist at least one element from the set
that is larger than m − ϵ as is written in the following definition.

Definition 2:

Let S be a set in R be bounded above, then m is a supremum if for some arbitrary

ϵ > 0 ∃s ∈ S, m − ϵ < s

If you have not understand the definition then stop here and read the definition
carefully first.

Monotone sequence
Definition:

Let X = (xn ) be a sequence of real numbers. We say that X is increasing if it satisfies

the inequalities

1
lecture 4 2

x1 ≤ x2 ≤ · · · ≤ xn ≤ xn+1 ≤ · · ·

We say that X is decreasing if it satisfies the inequalities

x1 ≥ x2 ≥ · · · ≥ xn ≥ xn+1 ≥ · · ·

We say that X is monotone if it is either increasing or decreasing.

Monotone Convergence Theorem

A monotone sequence of real numbers is convergent if and only if it is bounded.

Further:
(a) If X = (xn ) is a bounded increasing sequence, then

lim (xn ) = sup {xn : n ∈ N}

(b) If Y = (yn ) is a bounded decreasing sequence, then

lim (yn ) = inf {yn : n ∈ N} .

Here I am presenting a proof for the increasing sequence only, you can do the other
proof of your own.

Proof:

Suppose supn xn = M . The M is finite because the sequence is bounded (Here I use the
boundedness property of the sequence)
Then for given ϵ > 0, there exists n0 such that M − ϵ ≤ xn0 . (Here I have used the
definition 2 of least upper bound)
Since (xn ) is increasing, we have xn0 ≤ xn for all n ≥ n0 . This implies that

M − ϵ ≤ xn0 ≤ xn ≤ M ≤ M + ϵ for all n ≥ n0 .

That is xn → M .

√ √
Example 1. Let x1 = 2 and xn = 2 + xn−1 for n > 1. Then use induction to see
that 0 ≤ xn ≤ 2 and (xn ) is increasing. Therefore, by previous result (xn ) converges.
√
Suppose xn → λ. Then λ = 2 + λ. This implies that λ = 2.

Octobar 2024
lecture 4 3

xn+1
Example 2. Let x1 = 8 and xn+1 = 12 xn + 2. Note that xn
< 1. Hence the sequence
is decreasing. Since xn > 0, the sequence is bounded below. Therefore (xn ) converges.
λ
Suppose xn → λ. Then λ = 2
+ 2. Therefore, λ = 2.

Algebra of limits
Theorem Suppose that (xn ) and (yn ) are convergent real sequences and c ∈ R. Then
the sequences (cxn ) , (xn + yn ), and (xn yn ) converge, and

lim cxn = c n→∞

n→∞
lim xn

lim (xn + yn ) = lim xn + lim yn

n→∞ n→∞ n→∞
  
lim (xn yn ) = lim xn lim yn
n→∞ n→∞ n→∞

Proof. We let

x = lim xn , y = lim yn
n→∞ n→∞

The first statement is immediate if c = 0. Otherwise, let ϵ > 0 be given, and choose
N ∈ N such that

ϵ
|xn − x| < for all n > N
|c|
Then

|cxn − cx| < ϵ for all n > N

which proves that (cxn ) converges to cx. For the second statement, let ϵ > 0 be given,
and choose P, Q ∈ N such that

ϵ ϵ
|xn − x| < for all n > P, |yn − y| < for all n > Q
2 2
Let N = max{P, Q}. Then for all n > N , we have

|(xn + yn ) − (x + y)| ≤ |xn − x| + |yn − y| < ϵ,

which proves that (xn + yn ) converges to x + y. For the third statement, note that
since (xn ) and (yn ) converge, they are bounded and there exists M > 0 such that

|xn | , |yn | ≤ M for all n ∈ N

ϵ
and |x|, |y| ≤ M . Given ϵ > 0, choose P, Q ∈ N such that |xn − x| < 2M
for all

Octobar 2024
lecture 4 4

ϵ
n > P, |yn − y| < 2M
for all n > Q, and let N = max{P, Q}. Then for all n > N ,

|xn yn − xy| = |(xn − x) yn + x (yn − y)|

≤ |xn − x| |yn | + |x| |yn − y|
≤ M (|xn − x| + |yn − y|)
< ϵ,

which proves that (xn yn ) converges to xy.

Remarks. Note that the convergence of (xn + yn ) does not imply the convergence of
(xn ) and (yn ) separately; for example, take xn = n and yn = −n. If, however, (xn )
converges then (yn ) converges if and only if (xn + yn ) converges.

Monotonicity property of limit

Theorem If (xn ) and (yn ) are convergent sequences and xn ≤ yn for all n ∈ N, then

lim xn ≤ lim yn
n→∞ n→∞

Proof. Suppose that xn → x and yn → y as n → ∞. Then for every ϵ > 0 there exists
P, Q ∈ N such that

ϵ
|x − xn | < 2
for all n > P,
ϵ
|y − yn | < 2
for all n > Q.

Choosing n > max{P, Q}, we have

ϵ ϵ
x = xn + x − xn < yn + = y + yn − y + < y + ϵ.
2 2
Since x < y + ϵ for every ϵ > 0, it follows that x ≤ y.

Remarks. (Caution)This result, of course, remains valid if the inequality xn ≤ yn holds

only for all sufficiently large n. Limits need not preserve strict inequalities. For example,
1/n > 0 for all n ∈ N but limn→∞ 1/n = 0.
It follows immediately that if (xn ) is a convergent sequence with m ≤ xn ≤ M for all
n ∈ N, then

m ≤ lim xn ≤ M
n→∞

The following “squeeze” or “sandwich” theorem is often useful in proving the conver-
gence of a sequence by bounding it between two simpler convergent sequences with equal

Octobar 2024
lecture 4 5

limits.

Theorem : (Sandwich/Squeeze Theorem)

Suppose that (xn ) , (yn ) and (zn ) are sequences such that xn ≤ yn ≤ zn for all n and that
xn → x0 and zn → x0 . Then yn → x0 .
Proof: Let ϵ > 0 be given. Since xn → x0 and zn → x0 , there exist N1 and N2 such
that

xn ∈ (x0 − ϵ, x0 + ϵ) for all n ≥ N1

and

zn ∈ (x0 − ϵ, x0 + ϵ) for all n ≥ N2

Choose N = max {N1 , N2 }. Then, since xn ≤ yn ≤ zn , we have

yn ∈ (x0 − ϵ, x0 + ϵ) for all n ≥ N

This proves that yn → x0 .

Examples
−1 sin n sin n
1. Since n
≤ n
≤ n1 , by sandwich theorem n
→ 0.

n2 n2 n2 n·n2 n·n2
2. 2. Let yn = n3 +n+1
+ n3 +n+2
+ ··· + n3 +2n
. Then n3 +2n
≤ yn ≤ n3 +n+1
and hence
yn → 1.

1
3. Let x ∈ R and 0 < x < 1. We show that xn → 0. Write x = 1+a
for some a > 0.
1 1 1
Then by Bernoulli’s inequality, 0 < xn = (1+a)n
≤ 1+na
< na
. By sandwich theorem
n
x → 0.
1 1
4. Let x ∈ R and x > 0. We show that x n → 1. Suppose x > 1 and x n = 1 + dn
for some dn > 0. By Bernoulli’s inequality, x = (1 + dn )n > 1 + ndn > ndn which
x
implies that 0 < dn < n
for all n ∈ N. By sandwich theorem dn → 0 and hence
1 1 1 1
x → 1. If 0 < x < 1, let y =
n
x
so that y n → 1 and hence x n → 1.
1 1
5. We show that n n → 1. Let n n = 1 + kn for some kn > 0 when n > 1. Hence
n = (1 + kn )n > 1 for n > 1. By Binomial theorem, if n > 1, n ≥ 1 + 12 n(n − 1)kn2 .
Therefore n − 1 ≥ 12 n(n − 1)kn2 and hence kn2 ≤ n2 . By sandwich theorem kn → 0
1
and therefore n n → 1.

Octobar 2024
lecture 4 6

So far whatever we have done, gives some way to find if a sequence is convergent
or not, For example, monotone convergent theorem, is a very strong theorem to say if a
theorem to conclude that a sequence is convergent. But unfortunately, the converse of that
theorem is not true. That is a sequence can be convergent but may not monotone, though
n
it must be bounded. for example the sequence { (−1)
n
}n≥1 converges but not monotone.So
we cannot conclude that if a sequence is not monotone then it is not convergent! So how
can we find if a sequence is divergent or not?
In the next section, we present an very important notion of subsequence that will help
us to conclude, if a sequence is bounded or not.
sin(n)
In example 1. we see that n
is convergent. But we will show that sinn is not
convergent.

Problem: Set an = sin n, for any n ≥ 1. Prove that the sequence (an )n≥1 is divergent.

Solution. Arguing by contradiction, we assume that the sequence (an )n≥1 is convergent.
Let a = limn→∞ sin n. Using the identity

sin(n + 1) = sin n cos 1 + cos n sin 1

we deduce that the sequence (cos n) converges and, moreover,

a = a cos 1 + b sin 1,

where b = limn→∞ cos n. Using now

cos(n + 1) = cos n cos 1 − sin n sin 1

we deduce that

b = b cos 1 − a sin 1.

These two relations imply a = b = 0, a contradiction, since a2 + b2 = 1.

Problem: Let A1 , A2 , . . . , Ak be nonnegative numbers. Compute

lim (An1 + An2 + · · · + Ank )1/n

n→∞

Solution. Without loss of generality, we may assume that A1 = max {A1 , . . . , Ak }.

Therefore An1 ≤ An1 + · · · + Ank ≤ kAn1 . It follows that

Octobar 2024
lecture 4 7

A1 = n→∞ lim (kAn1 )1/n = A1 ,

lim (An1 + · · · + Ank )1/n ≤ n→∞
lim (An1 )1/n ≤ n→∞

which shows that the limit sought is A1 := min {Aj ; 1 ≤ j ≤ k}.

Subsequence

Subsequences : Let (xn ) be a sequence and let (nk ) be any sequence of positive
integers such that n1 < n2 < n3 < . . .. The sequence (xnk ) is called a subsequence.
Note that here k varies from 1 to ∞.

Example A subsequence of the sequence (1/n),

1 1 1 1
1, , , , , . . .
2 3 4 5
is the sequence (1/k 2 )

1 1 1 1
1, , , , , . . .
4 9 16 25
Here, nk = k 2 . On the other hand, the sequences

1 1 1 1 1 1 1 1
1, 1, , , , , . . . , , 1, , , , . . .
2 3 4 5 2 3 4 5
aren’t subsequences of (1/n) since nk is not a strictly increasing function of k in either
case.
Before we move to the next theorem understand a very simple situation. suppose that
there is a train which is moving from Chennai to Kolkata. the subsequence is like some
coaches of the train.
Now if the entire train is moving from Chennai to Kolkata then any compartment is
also moving from Chennai to Kolkata.
The same works for sequence as well, the next theorem exactly state the result,

Theorem Every subsequence of a convergent sequence converges to the limit of the

sequence.

Proof. Suppose that (xn ) is a convergent sequence with limn→∞ xn = x and (xnk ) is a
subsequence. Let ϵ > 0. There exists N ∈ N such that |xn − x| < ϵ for all n > N . Since
nk → ∞ as k → ∞, there exists K ∈ N such that nk > N if k > K. Then k > K implies
that |xnk − x| < ϵ, so limk→∞ xnk = x.

Octobar 2024
lecture 4 8

A useful criterion for the divergence of a sequence follows immediately from this result
and the uniqueness of limits.

Divergence Criteria If a sequence X = (xn ) of real numbers has either of the following
properties, then X is divergent.
(i) X has two convergent subsequences X ′ = (xnk ) and X ′′ = (xrk ) whose limits are
not equal.
(ii) X is unbounded.

 
Example The sequence 1, 21 , 3, 14 , · · · is divergent.
This is the sequence Y = (yn ), where yn = n if n is odd, and yn = 1/n if n is even. It
can easily be seen that Y is not bounded. Hence, by the sequence is divergent.

Example The sequence ((−1)n+1 ),

1, −1, 1, −1, 1, . . . ,

has subsequences (1) and (−1) that converge to different limits, so it diverges.

Octobar 2024