Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Seunghee Ye Ma 8: Week 2 Oct 6

Week 2 Summary

This week, we will learn about sequences and real numbers. We first define what we mean by a
sequence and discuss several properties of sequences. Then, we will talk about what it means for a
sequence to converge to a limit.
In the second section, we will define real numbers. But first, we take a look at what it is about the
reals that make them so special. The formal definition of real numbers will also give us a plethora of
sequences to work with.

Topics

Page

1 Sequences 1
1.1 First Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Convergence and Limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Arithmetic With Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Real numbers 5

1 Sequences
1.1 First Definitions
Let’s begin by defining what we mean by a sequence of real numbers.

Definition (Sequence)
A sequence of real numbers is an infinite ordered set of real numbers {a1 , a2 , . . . } indexed by the natural
numbers.

A sequence of real numbers can be anything. For example, we could have the sequence an = n for all n; or
also an = n1 for all n. Of course, the sequence could be constant as well e.g. an = 1 for all n. However, it’s
important to note that a sequence does not have to have a “nice” pattern as the ones above. For example,
we could consider the sequence {an }∞ 1 where an is the n-th decimal digit of π.
When we talk about sequences, there are a few properties that

Definition A sequence {an } is said to be:

• bounded from below if there exists L ∈ R such that for all n, we have |an | ≥ L. Then, L is called
a lower bound for {an }

• bounded from above if there exists U ∈ R such that for all n, we have |an | ≤ U .Then, U is called
an upper bound for {an }

• bounded if it is bounded from above and from below

• monotonically increasing or nondecreasing if an ≤ an+1 for all n ∈ N

• monotonically decreasing or nonincreasing if an ≥ an+1 for all n ∈ N

Example 1.1 (Fibonacci sequence). Some sequences are defined recursively. One of the most famous such
sequences is the Fibonacci sequence which is defined as follows.

a1 = a2 = 1, an = an−1 + an−2 ∀n ≥ 3

Page 1 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

Is the Fibonacci sequence bounded from below/above or monotonically increasing/decreasing? Let’s show
that the Fibonacci sequence is bounded from below but not bounded from above; and it is monotonically
increasing.

Proof. Note that a1 = a2 = 1 > 0. Now, we show by induction that for all n ∈ N, an > 0. Base case is
true by the definition of the Fibonacci sequence. The induction step is equally as straightforward. Suppose
ak > 0 for all k ≤ n. Now, an+1 = an + an−1 . By induction hypothesis, we know that an , an−1 > 0 and
thus, their sum must also be nonnegative. Therefore, an+1 > 0. We conclude that an > 0 for all n ∈ N and
thus, the sequence {an } is bounded from below by 0.
The second claim is that {an } is not bounded from above. Let’s proceed by contradiction. Suppose
that there exists an upper bound U . We saw above that an > 0 for all n. In fact, by exactly the same
argument, we can show that 1 is a lower bound for {an }. The first few elements of the sequence are
a1 = a2 = 1, a3 = 2, a4 = 3, a4 = 5. In particular, an+1 = an + an−1 ≥ an + 1 for all n ≥ 5. Then,
we see that an+1 ≥ a2 + (n − 1) = n for all n. Since U ∈ R we can find a natural number N such that
rN > max(U, 5). However, since N ≥ 5, we have aN ≥ N > U and thus U is not an upper bound for
{an }. Hence, {aN } is not bounded from above. (Note that we also proved that the Fibonacci sequence is
monotonically increasing!)

Upper and lower bounds for sequences are very important since they give us a great deal of information
about how the sequence behaves. It will also be important in deciding whether a sequence converges, as we
will see later. Some values of upper/lower bounds are special:

Definition (Least Upper Bound and Greatest Lower Bound)


An upper bound U for the sequence {an } is called the least upper bound if for any upper bound U 0 ,
we have the inequality U ≤ U 0 .
A lower bound L for the sequence {an } is called the greatest lower bound if for any lower bound L0 ,
we have the inequality L ≥ L0 .

Example 1.2. Find the least upper bound and the greatest lower bound for the Fibonacci sequence.

Proof. First, least upper bound is easy. {an } is not bounded from above so there are no upper bounds.
Hence, a least upper bound does not exist.
The great lower bound is also not very hard to find. We saw above that 0 and 1 are lower bounds for
{an }. Since 1 > 0, 1 will be our candidate for the greatest lower bound. Now suppose L is a lower bound.
Then, we must have L ≤ an for all n ∈ N . In particular, L ≤ 1 = a1 . But this precisely means that 1, which
is a lower bound itself, satisfies the definition of the greatest lower bound. Hence, the greatest lower bound
for {an } is 1.

Why are least upper bounds and greatest lower bounds so special? Are they even that special? To
answer the second question, let’s take a step back and think about why we want the natural numbers, N:
we want to count things. Then, why do we need integers? We want to be able to subtract natural numbers.
Subtraction is a very “natural” thing to do once we have the natural numbers so we expand our scope of
numbers to the integers. Now that we can add and subtract, why talk about the rational numbers? Because
we want to be able to divide integers.
With our scope broadened to the rational numbers, we can add, subtract, multiply and divide. √ What
more do we want? Why do we need the real numbers? Is it because some numbers such as 2, π, and e
are irrational? That still doesn’t justify all the other irrational numbers that we are adding to get the real
numbers. The answer to this is the least upper bound property.

Proposition 1.1. Let A ⊂ R be a subset of real numbers which is bounded from above. Then, A has a least
upper bound

Page 2 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

It turns out that the least upper bound property is very useful. So much so that it justifies talking about
R instead of just Q. The following is an example of where the least upper bound property comes in handy.

Proposition 1.2 (Archimedean property). Let x, y ∈ R and suppose x > 0. Then, we can find a natural
number n ∈ N such that nx > y

While this proposition might be familiar and even sound trivial to you, proof of it is not entirely trivial.

Proof. Suppose we cannot find n ∈ N such that nx > y. Then, it means that for all n ∈ N, we nx ≤ y. In
other words, the set {nx | n ∈ N} is bounded from above by y. By the least upper bound property of R, we
can find a least upper bound for this set, say z.
Now, consider z − x. Since z − x < z and z is the least upper bound, we know that z − x is not an upper
bound for the set {nx | n ∈ N}. In particular, we can find N ∈ N such that N x > z − x. However, this
implies that
N x > z − x ⇒ N x + x > z ⇒ (N + 1)x > z
Since (N + 1)x ∈ {nx | n ∈ N }, the last inequality implies that z is not an upper bound for our set, which
is a contradiction.
Therefore, we conclude that we can find n ∈ N such that nx > y.

1.2 Convergence and Limit


Now, let’s talk about convergence of sequences. We all have an intuitive idea of what it means for a sequence
to converge. It should roughly be the sequence approaching a particular value as we go off to infinity. The
following definition is the mathematical way of saying exactly what we think a convergent sequence should
be.

Definition (Limit and Convergence)


Let {an }∞
1 be a sequence of real numbers. We say that the sequence converges if there exists a real
number L ∈ R such that for all ε > 0, we can find N ∈ N such that for all n ≥ N , we have

|an − L| < ε

If such L exists, L is called the limit of the sequence {an }.

If this is your first time seeing a definition with ε, spend some time to take in what the definition is realy
saying and convince yourself that it is exactly what we described in the paragraph above. We will introduce
more definitions and theorems that will involve statement such as “for all ε > 0” in the future so it will be
worth your effort understanding this definition.

Example 1.3. Show that the sequence {an }∞ n1


1 defined by an = (−1) n converges to 0

Proof. Let ε > 0 be a positive real number. Then, we want to find N ∈ N such that |an − 0| = |an | < ε for
all n ≥ N . Since ε > 0, 1ε > 0. Hence, we can find N ∈ N such that 1ε < N . Then, we have that N1 < ε.
Now, for all n ≥ N we have:
1 1
|an − 0| = |an | = ≤ <ε
n N
Therefore, the sequence {an } converges to 0 as n approaches infinity.
√ √
Example 1.4. Let an = n + 1 − n. Find wheter an converges, and if it does, find its limit.

Page 3 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

Solution. Let ε > 0. By plugging in a few large valuese of n, we “guess” that the limit of the sequence is 0.
And note that
√ √ √ √
| n + 1 − n| = n+1− n (1)
√ √
√ √ n+1+ n
= ( n + 1 − n) √ √ (2)
n+1+ n
n+1−n
= √ √ (3)
n+1+ n
1
= √ √ (4)
n+1− n
1
< √ (5)
n

Hence, we want to find N such that for all n ≥ N , √1n is small (more precisely, less than ε). Since ε > 0 we
know ε2 > 0. By the Archimedean property, we can find N ∈ N such that N > ε2 . In particular, we have
√1 < ε.
N
Now, let n ≥ N . Then, we have
√ √
|an − 0| = | n + 1 − n − 0| (6)
√ √
= | n + 1 − n| (7)
1 1
< √ ≤ √ <ε (8)
n N

Hence, we conclude that the sequence an converges to 0 as n approaches infinity.

1.3 Arithmetic With Limits


There are a few properties of limits that make working with them particularly nice. We will talk about these
properties later this course but we preview a couple of them right here.

Proposition 1.3. Let {an } and {bn } be sequences which converge to L and M , respectively, as n approaches
∞. Define a new sequence {cn } by letting cn = an + bn for all n. Then, {cn } approaches L + M as n
approaches infinity.

Proof. Let ε > 0. Since ε > 0, we know that 2ε > 0. Applying the definition of convergence to {an } and
{bn } with their respective limit, we see that we can find:

1. N1 such that for all n ≥ N1 , we have |an − L| < 2ε ; and


ε
2. N2 such that for all n ≥ N2 , we have |bn − M | < 2

Now, let N = max(N1 , N2 ). The claim is that this particular N satisfies the condition in the definition of
convergence. Let n ≥ N . Then, we have to show that

|cn − (L + M )| < ε

Since cn = an + bn , by algebraic manipulation and triangle inequality we see that

|cn − (L + M )| = |(an + bn ) − (L + M )| = |(an − L) + (bn − M )| ≤ |an − L| + |bn − M |

However, since N was chosen to be the maximum of N1 and N2 , we have that n ≥ N ≥ N1 and N2 .

Page 4 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

ε
Therefore, |an − L| < 2 and |bn − M | < 2ε . Putting these together we see that
ε ε
|cn − (L + M )| ≤ |an − L| + |bn − M | < + =ε
2 2
Therefore, for any ε > 0, we can find N such that for all n ≥ N , |cn − (L + M )| < ε. Thus, we conclude
that {cn } approaches L + M as n approaches infinity.

Proposition 1.4. Let {an }, {bn } be sequences that converge to limits L and M respectively. Define a new
sequence {cn } by cn = an · bn . Then, the sequence {cn } converges to LM as n approaches infinity.
The proof of Proposition 1.4 is similar to the proof of Proposition 1.3, only slightly more complicated.
We will omit the proof of Proposition 1.4 but instead look at the following example.
Example 1.5. Let {an } be a sequence defined as follows:
p
a1 = 1, and an+1 = 1 + a2n for all n ≥ 2

Show that {an } does not converge to a limit.


2
Solution. Suppose {ap
n } does converge to a limit L. Then, by Proposition 1.4, we know that an must converge
2 2
to L . Since an+1 = 1 + an , by squaring both sides we obtain

a2n+1 = 1 + a2n

By taking the limit of both sides as n goes to infinity we get

lim a2n+1 = lim (1 + a2n )


n→∞ n→∞

The left hand side is simply L2 as we saw above. The right hand side, however, is

lim (1 + a2n ) = 1 + lim a2n = 1 + L2


n→∞ n→∞

Therefore, we get L2 = 1 + L2 which is a contradiction. Hence, the original sequence {an } does not converge
to a limit.

2 Real numbers
There are lots of interesting sequences that arise naturally in mathematics but we now dive into studying
how we can use sequences to define the real numbers.
Definition (Real Numbers)
A real number is an expression of the form x = ±a1 a2 · · · am .b1 b2 · · · where {bn }∞
1 is a sequence of
integers between 0 and 9, inclusive. Alternatively, we think of such an expression as x = a + 0.b1 b2 · · · where
a ∈ Z and {bn ∈ Z |, 0 ≤ bn ≤ 9n ∈ N} (in other words, x is made up of its integral part before the decimal
point and the part after the decimal part).
Given a real number x as above, its truncation to n decimal places is the rational number
m
X ∞
X
tn (x) = a1 · · · am .b1 · · · bn := a1 10m−1 + · · · + am 100 + b1 10−1 + b2 10−2 + · · · = ai 10m−i + bj 10−j
i=1 j=1

Note that given a real number x, we can associate to it a sequence rational numbers {tn (x)}∞
1 .
Suppose we have two real numbers x and y following above definition. We will say that x and y are equal
as real numbers, and write x = y, if the sequence {tn (x) − tn (y)}∞
1 converges to 0.

Page 5 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

Proposition 2.1. Let x and y be real numbers. Suppose x = a1 · · · ar .b1 b2 · · · and y = a01 · · · a0s .b01 b02 · · · .
Then, x = y if 1. ai = a0i for all i = 1, . . . , r = s; and 2. bi = b0i for all i; or there exists N such that
bN = b0N − 1, and bn = 0, b0n = 9 for all n > N (or with roles of bi and b0i reversed).
In other words, real numbers x and y are equal if x and y are identically the same or if we can write

x = a1 · · · ar−1 ar .00 · · · , y = a1 · · · ar−1 [ar − 1].99 · · ·

or
x = a1 · · · ar .b1 · · · bN 000 · · · , , y = a1 · · · ar .b1 · · · [bN − 1]999 · · ·

Proof. Suppose x and y are real numbers with above properties. If x and y are identically the same, clearly
we have tn (x) = tn (y) for all n. Hence, {tn (x) − tn (y)}∞ ∞
1 is just the constant sequence {0}1 which clearly
converges to 0.
Now suppose x and y are not identically same but satisfy 1 and 2 above. Note that since ai = a0i for all
i = 1, . . . , r = s, we have that t0 (x) = t0 (y). Moreover, since bi = b0i for i = 1, . . . , N − 1, we know that
ti (x) = ti (y) for all i = 1, . . . , N − 1
bN b0N
Since tN (x) = tN −1 (x) + 10 N and tN (y) = tN −1 (y) + 10N we get

b0
   
bN
tN (x) − tN (y) = tN −1 (x) + − tN −1 (y) + NN (9)
10N 10
0
 
bN − bN
= (tN −1 (x) − tN −1 (y)) + (10)
10N
1
= (11)
10N
More over, since bi = 0, b0i = 9 for all i > N , we have the following:

bN +1 − b0N +1
 
tN +1 (x) − tN +1 (y) = (tN (x) − tN (y)) + (12)
10N +1
9
= (tN (x) − tN (y)) − N +1 (13)
10
1 9 10 − 9 1
= − N +1 = N +1 = N +1 (14)
10N 10 10 10

By induction, we can show that for all i > N , we have ti (x) − ti (y) = 101 i . Hence, we want to show that this
sequence converges to 0.
Let ε > 0 be given. Then, we can find M large enough that 10M > ε. Then, we have that 1ε < 101M .
Hence, for all i ≥ M we have
1 1
|ti (x) − ti (y)| = i ≤ M < ε
10 10
Therefore, {tn (x) − tn (y)}∞
1 converges to 0 as n approaches ∞. Thus, x = y as real numbers.

Definition (Cauchy Sequences)


Let {an } be a sequence. We say that {an } is a Cauchy sequence if for all ε > 0, we can find N ∈ N
such that for all n, m ≥ N , we have
|an − am | < ε

The essense of Cauchy sequences is that a sequence that is Cauchy is one whose terms get very close to
each other as n goes off to infinity. This should strongly hint that the sequence should converge. Before
we investigate the relationship between Cauchy sequences and convergent sequences, let’s first look at a few
examples of Cauchy sequences.

Page 6 of 7
Seunghee Ye Ma 8: Week 2 Oct 6

Example 2.1. Let x be a real number represented as x = a1 · · · am .b1 b2 · · · . Show that {tn (x)} forms a
Cauchy sequence.

Solution. Let ε > 0. Then, we can find N1 ∈ N such that 101N1 < ε. Let N = N1 + 1. Now, take n, m ≥ N .
Then, |tn (x) − tN (x)| < 101N and |tm (x) − tN (x)| < 101N . Hence,

|tn (x) − tm (x)| = |tn (x) − tm (x) + tN (x) − tN (x)| (15)


= |(tn (x) − tN (x)) − (tm (x) − tN (x))| (16)
≤ |tn (x) − tN (x)| + |tm (x) − tN (x)| (17)
1 1 2 1
< + N = N < N1 < ε (18)
10N 10 10 10
Thereore, {tn (x)} is a Cauchy sequence.

Page 7 of 7

You might also like