3-Sequence and their convergence

Shivangi Chandel

Jindal School of Government and Public Policy

Course: Advanced Mathematical Tools

Shivangi Chandel (JSGP) Mathematics ADMT 3052 1 / 20

Sequence and their convergence

1 Sequences in R

Shivangi Chandel (JSGP) Mathematics ADMT 3052 2 / 20

What does it mean for n to ‘reach’ infinity?
If you were to walk to the end of the road which is 1 km long, you
would first have to walk half that distance i.e. 1/2 kms.. Even before
walking that distance, you would have to cover half of that distance,
which will be 1/4 kms.. By the same argument, you would have to
cover 1/8 kms. before you cover 1/4 kms., 1/16 kms. before 1/8
kms. and so on. Figure 1 explains your movement across the road.

1
1/2
1/4
1/8
..
.
Figure: Zeno’s Dichotomy Paradox

Zeno pointed out since an infinite number of distances must be covered

before one could reach the end of the road, the task of covering this finite
length would be impossible!
Shivangi Chandel (JSGP) Mathematics ADMT 3052 3 / 20
 Many mathematicians who came much after him had done the math
to show that the sum of infinitely many sub-distances
1/2 + 1/4 + 1/8 + 1/16 + ...... is equal to 1.
Adding infinitely many numbers (in a series) does not always become
undefined.
Many infinite series, like the one in Dichotomy Paradox, have a finite
sum. These series are called as convergent series.
Is there is rule by which we can calculate the sum of any convergent
series?

Shivangi Chandel (JSGP) Mathematics ADMT 3052 4 / 20

 The answer is yes. The method involves constructing a ‘sequence’ of
partial sums, like

S0 =1/2,
S1 =1/2 + 1/4,
....,
Sn =1/2 + 1/4 + ... + 1/2n+1 ,
....

and calculating the ‘limit’ of this sequence which will be nothing but
1/2 + 1/4 + · · · .
I have introduced two new terms in the above method, sequence and
limit. Let us take a closer look at what these terms imply.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 5 / 20

What is a sequence?

An infinite sequence of reals is an infinite list of real numbers that is

ordered as the natural numbers i.e corresponding to each natural
number n is an element of the sequence, say xn .
In other words, a sequence of real numbers is a map/function from N
to R.
Notation {xn }, {x0 , x1 , x2 , x3 , · · · , xn , · · · }, (xn )∞
n=0
Examples include
I {1, 1, 1, 1, · · · }
I {1, 2, 3, 4, · · · }
1 1
I {1, , , · · · }
2 3
I {1, −1, 1, −1, · · · }
−1 1 −1 1
I {1, , , , ···}
2 4 8 16

Shivangi Chandel (JSGP) Mathematics ADMT 3052 6 / 20

Limit of a sequence

1 1 1 1
Consider the sequence {1, , , , ..., n , ...} versus the sequence
2 4 8 2
{1, 2, 3, .., n, ...}.
What is the difference between the two?
The elements in the first sequence get closer and closer to 0 while the
elements in the second sequence keep on increasing indefinitely
(Check for yourself).
The first one is what we call a convergent sequence where as n
becomes larger and larger, the elements in the sequence get closer
and closer to a finite value which is called the limit of that sequence.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 7 / 20

Definition 1.1
Let {xn } be a sequence of real numbers. A real number x is called the
limit of the sequence {xn } if given any real number  > 0; there exists a
natural number N , such that |xn − x| < , ∀ n ≥ N .

If x is the limit of the sequence, we say that the sequence converges

to x and write

lim xn = x, or simply xn −→ x
n−→∞

.
The notation N is used to emphasize that the choice of N depends
on the value of . - it is often convenient to write N instead of N .
In most cases, a “small” value of  will usually require a “large” value
of N to guarantee that the distance between xn and x is less than 
for all n = N, N + 1, N + 2, · · · .

Shivangi Chandel (JSGP) Mathematics ADMT 3052 8 / 20

 Define, for a ∈ R and  > 0, the −neighbourhood of a as the set
V (a) = {b ∈ R : |b − a| < }.
Alternative definition of convergence of a sequence is:
Let {xn } be a sequence of real numbers. A real number x is called
the limit of the sequence {xn } if for every -neighbourhood V (x) of
x, there exists a natural number N such that for all n ≥ N,

xn ∈ V (x).

That is, only finitely many elements x1 , · · · , xN−1 lie outside the
−neighbourhood of the limit x, and this holds for every  > 0.
Notice that both definitions can only be used to verify that a proposed
value x is indeed the limit. They do not determine what this x will be.
1
Show that lim = 0.
n−→∞ n

Shivangi Chandel (JSGP) Mathematics ADMT 3052 9 / 20

 Proofs of most theorems on sequences and their limits require the
triangle inequality:
|x + y | ≤ |x| + |y |, x, y ∈ R,
or the subtraction variant of the triangle inequality:
|x − y | ≥ | |x| − |y | | , x, y ∈ R.

Theorem 1.1
A sequence in R can have at most one limit.

Proof.
Suppose not. Suppose that x 0 and x 00 are both limits of {xn }. Then, for each  > 0
there exists a N 0 ∈ N such that |xn − x 0 | < /2 for all n ≥ N 0 , and there exists a N 00 ∈ N
such that |xn − x 00 | < /2 for all n ≥ N 00 . Define N = max N 0 , N 00 . Then for all n ≥ N,
we apply the Triangle Inequality to get

|x 0 − x 00 | = |x 0 − xn + xn − x 00 | ≤ |x 0 − xn | + |xn − x 00 | < /2 + /2 = 

Since  > 0 is an arbitrary positive number, we conclude that x 0 − x 00 = 0 or

x 0 = x 00 .
Shivangi Chandel (JSGP) Mathematics ADMT 3052 10 / 20
Convergence =⇒ Boundededness
Definition 1.2
A sequence {xn } of real numbers is said to be bounded if there exists
areal number K > 0 such that |xn | ≤ K for all n ∈ N.

Theorem 1.2
A convergent sequence of real numbers is bounded.

Proof.
Suppose that xn −→ x and let  = 1. Then, by definition of convergence,
there exists a natural number N such that |xn − x| < 1 for all n ≥ N.
Then for any n ≥ N,

|xn | = |xn − x + x| ≤ |xn − x| + |x| < 1 + |x|

If we set K = max{|x1 |, |x2 |, · · · , |xN−1 |, 1 + |x|}, then it follows that

|xn | ≤ K for all n ∈ N.
Shivangi Chandel (JSGP) Mathematics ADMT 3052 11 / 20
Algebra on Limits

Theorem 1.3
1 Let {xn } and {yn } be sequences of real numbers that converge to x
and y , respectively, and let c ∈ R. Then the sequences {xn + yn },
{xn − yn }, {xn yn }, and {cxn }, converge to x + y , x − y , xy , cx
respectively.
2 If {xn } converges to x and {zn } is a sequence of non-zero real
numbers that converges to z and if z 6= 0, then the quotient sequence
{xn /zn } converges to x/z.

Proof.
Theorem 3.2.3 (pages 61-63) Robert G. Bartle, Donald R. Sherbert
(1999). Introduction to real analysis. Third edition. Wiley, New York.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 12 / 20

 Coming back to the Dichotomy paradox, the limit of the sequence,
{Sn } with Sn = 1/2 + 1/4 + ... + 1/2n+1 for n = 0, 1, 2, · · · if it
exists, is the sum of the infinite series 1/2 + 1/4 + · · · .

1 1 1
lim Sn = lim ( + + ... + n+1 )
n−→∞ 2 4  2
n−→∞

1
= lim 1 − n+1 (Why ?)
n−→∞ 2
1
= 1 − lim n+1 (Why ?)
n−→∞ 2
= 1.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 13 / 20

Monotone Sequences

Definition 1.3
Let {xn } be a sequence of real numbers. We say that {xn } is increasing if
xn ≤ xn+1 for all n ∈ N. We say that {xn } is decreasing if xn ≥ xn+1 for
all n ∈ N. We say that {xn } is monotone if it is either increasing or
decreasing.

Sequences like {1, 2, 3, · · · }, {1, 2, 2, 3, 3, 3, · · · }, {a, a2 , a3 , · · · } for

a > 1 are increasing.
Sequences like {1, 1/2, 1/3, · · · }, {1, 1/2, 1/2, 1/3, 1/3, 1/3, · · · },
{a, a2 , a3 , · · · } for a < 1 are decreasing.
Sequences like {1, −1, 1, · · · }, {1, −2, 3, −4, · · · } are not monotone.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 14 / 20

Monotonicity + Boundedness =⇒ Convergence

Theorem 1.4 (Monotone Convergence Theorem (MCT))

A monotone sequence of real numbers is convergent if and only if it is
bounded. Further,
(a) If {xn } is a bounded increasing sequence, then
lim xn = sup{xn : n ∈ N}.
(b) If {yn } is a bounded decreasing sequence, then
lim yn = inf{yn : n ∈ N}.
where sup A or supremum of a set A is the least upper bound of A and
inf A or infimum of a set A is the greatest lower bound of A.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 15 / 20

Proof idea.
( =⇒ ) We have shown earlier that every convergent sequence is bounded.
(⇐=) (a) Let {xn } be a bounded and increasing sequence. Then because, the sequence
is bounded, there must exist a least upper bound among all the bounds of the set
{xn : n ∈ N} ( also known as the Completeness property). Call that x. Then it can be
shown that lim xn = x.
For any given  > 0, x −  is not an upper bound of the set {xn : n ∈ N}. Therefore,
there exists an element of the sequence xN such that x −  ≤ xN . Because {xn } is an
increasing sequence, for all n ≥ N,

x −  < xN ≤ x < x + .

Therefore, we have |xn − x| <  ∀n ≥ N. Since  > 0 is arbitrary, xn −→ x.

(b) For a bounded decreasing sequence {yn }, {−yn } is a bounded increasing sequence.
Therefore part (a) can be used to get,
lim yn = −(lim −yn ) = − sup{−yn : n ∈ N} = inf{yn : n ∈ N}.
√
Checkpwhether the sequence {xn } is convergent where x0 = 2 and
xn = 2 + xn−1 for all n > 1. What is its limit, if it exists?
1
Show that lim( √ ) = 0. (Use MCT for both)
n
Shivangi Chandel (JSGP) Mathematics ADMT 3052 16 / 20
Subsequences

Definition 1.4
Let {xn } be a sequence of real numbers and let n1 < n2 < · · · < nk < · · ·
be a strictly increasing sequence of natural numbers. Then the sequence
{xnk } given by
{xn1 , xn2 , · · · , xnk , · · · }
is called a subsequence of {xn }.

Examples: (a) {1, 1/4, 1/8, · · · }, {1/4, 1/15, 1/33, · · · } are

subsequences of sequence {1/n}. (b) {1, 1, · · · } and {−1, −1, · · · }
are obvious subsequences of sequence {1, −1, 1, −1 · · · }

Shivangi Chandel (JSGP) Mathematics ADMT 3052 17 / 20

 As you could guess correctly, any or all subsequences of a sequence
must converge to the limit of the sequence itself. Can you argue why?
(Use contradiction.)
Interesting thing about subsequences is that we can use them to
define the Divergence Criterion for a sequence:
If a sequence of real numbers is unbounded then it is divergent.
It is also not convergent if it have atleast two convergent subse-
quences whose limits are not equal.
Sequence {(−1)n } is divergent. So is {1, 1/2, 3, 1/4, · · · }.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 18 / 20

 While not every sequence is a monotone sequence, every
sequence has a monotone subsequence (How?).
I In any sequence {xn }, look for ’peaks’ i.e. those elements of the
sequence that are never exceeded by any term that follows them in the
sequence. Thus, xm is a peak of the sequence {xn } if xm ≥ xn for all
n ≥ m.
I Then, there are two cases, either the sequence has infinitely many
peaks or finitely many peaks.
I In a sequence with infinitely many peaks, you can generate a
decreasing subsequence by listing these peaks in increasing subscripts.
I In a sequence with finitely many peaks, you can generate an increasing
subsequence from elements of the sequence that are beyond the last
peak.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 19 / 20

Bolzano-Weierstrass Theorem (BWT)

Theorem 1.5 (Bolzano-Weierstrass Theorem (BWT))

A bounded sequence of real numbers has a convergent subsequence.

Proof.
Let {xn } be a bounded sequence of real numbers. From previous result,
{xn } will have a monotone subsequence {xnk }. Further, {xnk } is bounded
because {xn } is bounded. From Monotone Convergence Theorem, it
follows that the subsequence {xnk } is convergent. Hence proved.

Application of BWT:
The Bolzano-Weierstrass theorem allows one to prove that if the set
of allocations is compact and non-empty, then the system has a
Pareto-efficient allocation.
We will revisit this again once we define compact sets.

Shivangi Chandel (JSGP) Mathematics ADMT 3052 20 / 20