Topic 1 - Real sequences

Calculus
 Real sequences

• Definition (Real sequence). A real valued function is called a real sequence

if its domain is the set of natural numbers.
• Shortly, functions of the form f : N → R are real sequences.
• For a real sequence we shall use a, b, c, . . . and, for the function a : N → R
we use the notation (an ) ⊆ R.
• For a given index n ∈ N, let an := a(n). This real number is called the nth
member of the sequence (an ).
 Real sequences

• Examples.
? Arithmetic progression and geometric progression.

If λ, d ∈ R is fixed, then let an := λ + (n − 1)d whenever n ∈ N.

If λ, q ∈ R, q 6= 0 is fixed, then let an := λq n−1 whenever n ∈ N.

? Harmonic sequence. The sequence of the reciprocals of the natural num-

bers forms a harmonic sequence.
1
an := if n ∈ N.
n
? Fibonacci’s sequence: a1 = 1, a2 = 1, a3 = 2, a4 = 3, a5 = 5, ...
 Boundedness

• Definition (Boundedness of real sequences). We say that a real se-

quence (an ) is bounded from below if there exists k ∈ R such that an ≥ k
holds for all n ∈ N. Similarly, we say that (an ) is bounded from above if there
exists K ∈ R for which an ≤ K for all n ∈ N.
If (an ) is bounded from below and from above simultaneously, then it will be
called bounded.
• Geometrical meaning. Note that if (an ) is bounded from below and from
above, then there exists a bounded interval which contains the set

{an | n ∈ N}

• Exercise. Show that the sequences defined by an := n12 and bn := (−1)n

are bounded. Is the sequence cn := (−2)n bounded or not?
 Monotonicity

• Definition (Monotonicity of real sequences). We say that (an ) is decreas-

ing or strictly decreasing if, for all n ∈ N, we have

an ≥ an+1 or an > an+1 ,

respectively. Similarly, (an ) is increasing or strictly increasing if, for all n ∈ N,

we have
an ≤ an+1 or an < an+1 ,
respectively.
• A sequence (an ) is increasing and decreasing if and only if it is constant,
that is, there exists c ∈ R such that an = c holds for all n ∈ N. Give an
example of a real sequence, which is neither increasing nor decreasing!
 Convexity

• Definition (Convexity of real sequences). We say that (an ) is convex if,

for all n ∈ N with n ≥ 2, we have
an−1 + an+1
an ≤ .
2
The sequence (an ) is called concave if the sequence (−an ) is convex.

• Show that a real sequence (an ) is convex and concave (i.e, is affine) if and
only if it is an arithmetic progression.
 Convergence
• Definition (Convergence of a real sequence). We say that (an ) is convergent if
there exists A ∈ R such that, for all ε > 0 on can find n ∈ N such that |ak − A| < ε holds
whenever k ∈ N with k ≥ n.
Then the real number A is called a limit of the sequence (an ).

• Theorem (Uniqueness of the limit)*. If (an ) is convergent then it has at most one
limit.
• Motivated by the above theorem, we can introduce the following notation:
lim an .
n→∞

If (an ) is convergent and its limit is A, then we shall say that (an ) converges to A as n
tends to infinity and use the notation
lim an = A or an → A as n → ∞.
n→∞
 Divergent sequences

• Definition (Divergence of a real sequence). We say that (an ) is divergent

if it is not convergent. Furthermore, we shall say that
1. (an ) diverges to −∞ if, for all λ ∈ R one can find n ∈ N such that ak ≤ λ
whenever k ≥ n, or
2. (an ) diverges to +∞ if, for all λ ∈ R one can find n ∈ N such that ak ≥ λ
whenever k ≥ n.
If (an ) enjoys exactly one of the properties 1. or 2. we say that (an ) is
convergent in the extended sense.
 Connections

• Theorem (Convergence and boundedness).* If (an ) is convergent, then it

is bounded.

• The converse of the above theorem is not true! More precisely, there exist
divergent real sequences, which are bounded.
Question: What to assume which together with boundedness implies conver-
gence?

• Theorem (Convergence of bounded monotone sequences).* If (an ) is

decreasing or increasing and bounded from below or from above, respectively,
then it is convergent and

lim an = inf{an | n ∈ N} or lim an = sup{an | n ∈ N},

n→∞ n→∞

respectively.
 Cauchy property

• Definition (Cauchy property). We say that (an ) enjoys the Cauchy property
or, shortly, (an ) is a Cauchy-sequence if, for all ε > 0 we have n ∈ N such that
|ak − am | < ε whenever k , m ≥ n.

• Theorem (Characterization of convergence). A real sequence (an ) is

convergent if and only if it is a Cauchy sequence.

? In the case of real sequences, convergence and Cauchy property are

equivalent notions.
? Advantage of the above definition: We can apply it to test the convergence
of the given sequence even if we do not have any information about the
exact value of the limit.
 Subsequences
• Definition (Subsequence). The sequence (bn ) is called a subsequence of
(an ) if one can find (ϕn ) being strictly increasing such that

bn = aϕn , n ∈ N.

• Theorem (Convergence of a real sequence). If (an ) is convergent then,

for all strictly increasing sequence (ϕn ), the sequence (aϕn ) is convergent and

lim aϕn = lim an .

n→∞ n→∞

? A real sequence is not necessarily convergent even if it has convergent

subsequences!
? Each sequence is a subsequence of itself. (Indeed, let (ϕn ) be the iden-
tity, that is, ϕn = n for all n ∈ N.)
? If (an ) is a sequence such that (a2n ) and(a2n−1 ) are convergent tending
to the same limit, then (an ) is convergent and tends to the common limit.
? If a sequence is bounded, then all of its subsequences are bounded.
Consequently, if a sequence has an unbounded subsequence, then the
sequence in question must be unbounded.
 Accumulation points, density points

• Definition (Accumulation points, density points). We say that an ex-

tended real number p ∈ R ∪ {−∞, +∞} is an accumulation point or a density
point of (an ) if either
? p ∈ R and there exists a subsequence of (an ) which converges to p or
? p ∈ {−∞, +∞} and there exists a subsequence of (an ), which diverges
to p as n tends to infinity.
For the set of accumulation points of a sequence (an ) we shall use the notation
acc(an ).

• Note that if (an ) is convergent, then

lim an ∈ acc(an ),
n→∞

furthermore in this case the set acc(an ) has no further elements.

Consequently, for convergent sequences, the set of accumulation points is
never empty. In what follows we are going to show that acc(an ) is never empty.
 Accumulation points

• Theorem (Monotone Subsequence Theorem). Each real sequence has a

monotone subsequence.
• Corollary (Bolzano–Weierstrass Theorem).* Each bounded sequence has
a convergent subsequence.

• Let (an ) be a real sequence.

? If (an ) is not bounded: Then there exists a subsequence which diverges
to −∞ or +∞. Indeed, by the fact that the sequence is not bounded it is
not bounded from below or from above. If it is not bounded from below
then let b1 := a1 and, for n ≥ 2, let bn := ak , where k := min{` ∈ N |
a` ≤ bn−1 }. Then (bn ) is a subsequence of (an ), which diverges to −∞,
thus −∞ ∈ acc(an ). The complementary case can be treated similarly.
? If (an ) is bounded: Then, in view of the Bolzano–Weierstrass Theorem,
the set acc(an ) has at least one element.
Consequently, for any (an ), the set acc(an ) is indeed nonempty.
 Lower limit and upper limit

• A fundamental property of R is the following:

If a nonempty subset of R is bounded from below (or bounded from above)
then its infimum (or its supremum) exists.
• Definition (Lower and upper limit). Let (an ) be a real sequence. Then the
extended real number
lim inf an := inf acc(an ), or lim sup an := sup acc(an )
n→∞ n→∞

is called the lower limit or the upper limit of (an ), respectively.

 Lower limit and upper limit

? For all (an ) we have

lim inf an ≤ lim sup an .
n→∞ n→∞

? A real sequence (an ) is convergent of and only if

lim sup an ≤ lim inf an .

n→∞ n→∞

Then the common value is nothing else but the usual limit of the se-
quence.
 Exercises

• Determine the set of accumulation points of the following real sequences and
calculate their lower limit and upper limit! Investigate the following sequences
for convergence too!
1. an := (−1)n
n
2. an := 1 + (−1)
n
√
3. an := n + 1
4. an := sin( nπ
2
)

• Do we have a real sequence (an ) for which acc(an ) = [0, 1] or acc(an ) = ]0, 1[
holds?*
• Do we have a real sequence (an ) which takes each natural number infinitely
many times, more precisely, for which for all m ∈ N we have that the set

{n ∈ N | an = m}

has infinitely many elements?

 Convergence and operations

• Theorem. Let (an ) and (bn ) be convergent real sequences tending to the
limits A ∈ R and B ∈ R, respectively, and let λ ∈ R. Then the following
statements hold.
(1) The sum (an + bn ) of (an ) and (bn ) is convergent and tends to A + B as
n → ∞.
(2) The scalar multiple (λan ) is convergent and tends to λA as n → ∞.
(3) The product (an bn ) of (an ) and (bn ) is convergent and tends to AB as
n → ∞.
(4) If, in addition, bn B 6= 0 for all n ∈ N, then the ratio ( abnn ) of (an ) and (bn ) is
convergent and tends to BA as n → ∞.

• Statements (1) and (2) mean that the set of all convergent real sequences
forms a real vector space.
 Null sequences

• Definition (Null sequence). If (an ) ⊆ R converges to zero, then it is called a null

sequence.
• Remark. If you multiply a real number by zero, the result is zero. The analog of this
statement is not necessarily true for real sequences, or more precisely: if you multiply
a null sequence by an arbitrary real sequence, you cannot predict what the product will
be in terms of convergence.
? Let an := n1 and bn := n if n ∈ N. Then an → 0 and bn → +∞ as n → ∞.
However: an bn = 1 → 1 as n → ∞.
1
? Let an := n2
and bn := n if n ∈ N. Then an → 0 and bn → +∞ as n → ∞.
1
However: an bn = n
→ 0 as n → ∞.
? Let an := n1 and bn := n2 if n ∈ N. Then an → 0 and bn → +∞ as n → ∞.
However: an bn = n → +∞ as n → ∞.
• Theorem. If (an ) is a null sequence and (bn ) is bounded, then (an bn ) is a null se-
quence. Consequently, the product of a null sequence and a convergent sequence is a
null sequence.
• However, there is something to be said for the unbounded case: Let k ∈ Q and q ∈ R
with |q| < 1 be arbitrarily fixed. Then (nk q n ) is a null sequence.
 Limits of notable convergent real sequences
• Let k ∈ Q be arbitrarily fixed. Then

+∞ if k > 0,

nk → 1 if k = 0,

0 if k < 0.
• Let λ ∈ R with λ > 0. Then

+∞ if λ > 1,

√
n
λn → 1 if λ = 1, and λ→1

0 if λ < 1.
√
as
√ n → ∞. Moreover, we have n n → 1 as n → ∞. On the other hand, the sequence
n
( n!) diverges to +∞.
• The sequences of the form
constant polynomial exponential factorial
, , , and
polynomial exponential factorial double exponential
are null sequences.
• Let k ∈ Q be arbitrary but fixed and (xn ) ⊆ R such that xn → +∞ as n → ∞. Then
 k xn
lim 1 + → ek .
n→∞ xn
 Convergence and ordering

• Theorem. Let (an ) and (bn ) be convergent real sequences tending to A ∈ R

and B ∈ R, respectively.
(1) If there exists n ∈ N such that ak < bk for all k ≥ n, then A ≤ B holds.
(2) If A < B holds, then there exists n ∈ N such that ak < bk for all k ≥ n.

• Theorem (Squeeze Theorem). Let (an ) and (bn ) be convergent real se-
quences tending to the same limit L ∈ R. If (cn ) is a real sequence such that
there exists n ∈ N with ak ≤ ck ≤ bk whenever k ≥ n, then (cn ) is convergent
and tends to L as n → ∞.
√
• Show that cn := n 2n + 3n tends to 3 as n → ∞.
√
n
√
n
√n
√
n
cn = 2n + 3n ≤ 3n + 3n = 2 · 3n = 3 · 2 =: bn ,
√
n
√
n
cn = 2n + 3n ≥ 3n = 3 =: an .

The sequences (an ) and (bn ) are convergent having the same limit L := 3.
Moreover, we have an ≤ cn ≤ bn for all n ∈ N. In view of the Squezze
Theorem, (cn ) is convergent and tends to 3 as n → ∞.