Infinite sequences

and series

February 20, 2019

 Intro
• There are many complicated functions
that we can not calculate value exactly
• Many functions that are difficult to
find integral
• Polynomials are easy to compute
(derivative, integral ...)
• We want to write function as sum of
polynomial
 Sequence
A sequence {an} is a function whose
variable n takes integer values

n = 1, 2, . . .

{an} can be written as a list of number

a1, a2, . . . , an, . . .

 General formula

Write an as a formula of n

an = f (n), n = 1, 2, . . .

Ex: an = 1/n2:
a1 = 1, a2 = 1/4, a3 = 1/9, . . .
 General formula

Write an as a formula of n

an = f (n), n = 1, 2, . . .

Ex: an = 1/n2:
a1 = 1, a2 = 1/4, a3 = 1/9, . . .
 Recurrence formula
Write an as function of previous terms
an−1, an−2, . . .
Ex: Fibonacci Sequence

an = an−1 + an−2,

a1 = a2 = 1
then

{an} = {1, 1, 2, 3, 5, 8, 13, . . . }

 Limit of sequence

Limit of sequence = limit of function an as n

goes to infinity.

lim an = L
n→∞
 Limit of sequence

Limit of sequence = limit of function an as n

goes to infinity.

lim an = L
n→∞
 Convergence

• If lim an = L < ∞ we say that

n→∞
{an} converges,
• If lim an = ∞ or the limit does not
n→∞
exist, say {an} diverges.
 Example
Does the following sequence converges?
4 8 16
2, , , , . . .
3 9 27

• Find the general formula:

2n
an = n−1
3
• then
2n
lim =0
n→∞ 3n−1
 Example
Does the following sequence converges?
4 8 16
2, , , , . . .
3 9 27

• Find the general formula:

2n
an = n−1
3
• then
2n
lim =0
n→∞ 3n−1
 Bounded sequence

Say that {an} has upper bound M if for all n

an ≤ M

Say that {an} haslower bound m if for all n

an ≥ m
 Bounded sequence

Say that {an} has upper bound M if for all n

an ≤ M

Say that {an} haslower bound m if for all n

an ≥ m
 Monotone sequence
{an} is increasing if for all n > N for some
fixed N
an+1 > an
{an} is decreasing if for all n > N for some
fixed N
an+1 < an
say {an} is monotone if increasing or
decreasing.
 Monotone sequence
{an} is increasing if for all n > N for some
fixed N
an+1 > an
{an} is decreasing if for all n > N for some
fixed N
an+1 < an
say {an} is monotone if increasing or
decreasing.
 Monotone sequence
{an} is increasing if for all n > N for some
fixed N
an+1 > an
{an} is decreasing if for all n > N for some
fixed N
an+1 < an
say {an} is monotone if increasing or
decreasing.
 Monotone convergence

Theorem: if {an} is monotone and bounded

then it converges.
 Example

√
Given an = 6 + an−1, a1 = 1. Prove that
{an} converges and find its limit.
Prove that {an} is increasing by induction:
• a2 > a1
• Suppose an+1 > an. we prove that

an+2 > an+1

Prove that {an} is increasing by induction:
• a2 > a1
• Suppose an+1 > an. we prove that

an+2 > an+1

 p p
an+2 = 6 + an+1 > 6 + an = an+1

because an+1 > an

Prove that {an} is bounded from above by 3
• a1 = 1 < 3
√
• if an < 3 then an+1 < 6+3=3
Prove that {an} is bounded from above by 3
• a1 = 1 < 3
√
• if an < 3 then an+1 < 6+3=3
{an} increasing and bounded so it
converges.
Suppose limn→∞ an = L, then

lim an+1 = lim an = L

n→∞ n→∞
√
Take limit on both sides of an+1 = 6 + an
√
L= 6+L

then L = 3.
What happen when we sum up all terms of a
sequence?
 Partial sum
Given sequence {an}. Let

SN = a1 + a2 + · · · + aN

SN is called the N−th partial sum

N
X
SN = an
n=1
 Infinite series

• {S1, S2, . . . , SN } is a sequence.

• If {SN } converges to L then say the
infinite series of an converges to L,
∞
X
an = lim SN = L
N→∞
n=1
 Infinite series

• {S1, S2, . . . , SN } is a sequence.

• If {SN } converges to L then say the
infinite series of an converges to L,
∞
X
an = lim SN = L
N→∞
n=1
• Usually we want to know if an infinite
series converges or diverges
• We will look at some special types of
infinite series
 Geometric series

Let a, r are real numbers. The series

∞
X
arn
n=0

is called the geometric series with common

ratio r
 Partial sum

SN = a0 + a1 + · · · + aN
= a + ar + ar2 + · · · + arN
= a(1 + r + · · · + rN )
1 − rN+1
=a
1−r
 Convergence

If |r| < 1 then rN+1 → 0 when N → ∞

so geometric series converges
∞
X a
arn =
1−r
n=0

If |r| ≥ 1 then the series diverges.

 Telescoping series
Example:
∞
X 1
n(n + 1)
n=1
Use partial fraction
1 1 1
= −
n(n + 1) n n + 1
 Telescoping series
Example:
∞
X 1
n(n + 1)
n=1
Use partial fraction
1 1 1
= −
n(n + 1) n n + 1
 Partial sum

1 1 1 1 1
SN = 1 − + − + − +
2 2 3 3 4
1 1
+ ··· + −
N N+1
1
=1−
N+1
∞
X 1
So =1
n(n + 1)
n=1
 General form
• Suppose we can write
an = bn − bn+1

• Then
a1 + a2 + · · · + aN
= b1 − b2 + b2 − b3 + · · · − bN+1
= b1 − bN+1
P
• an is called telescoping series
 Mixed series

• We can calculate the sum of series

mixed by geometric and telescoping
series
• For most of other series we can not
calculate the exact sum
 Series laws
P P
Suppose an and bn are convergent
series then
X X
• can = c an
P P P
• (an + bn) = an + bn
P P P
• (an − bn) = an − bn
 Example

Find sum of the series

∞  
X 3 1
+
n(n + 1) 2n
n=1
 Solution
∞ ∞
X 1 X 1
= −1
2n 2n
n=1 n=0
∞
X 1 1
: geometric r = , a = 1 so
2n 2
n=0
∞
X 1 1
= −1=1
2n 1 − 21
n=1
 Solution
∞
X 3
is telescoping series
n(n + 1)
n=1

N
X 3 3 3 3 3
= − + − ··· −
n(n + 1) 1 2 2 N+1
n=1
3 3
= − →3
1 N+1
∞
X 3
so =3
n(n + 1)
n=1
 Solution

So
∞  
X 3 1
+ n =3+1=4
n(n + 1) 2
n=1
 Positive series

• If an > 0 for all n larger than some N

P
then we say an positive.
P
• If an positive then {SN } increasing.
 Convergence tests

• Most Series can not be computed

exactly
• Need to know if a series converges or
not
• There are simple tests for that purpose
 (1) Divergence test

If
lim an 6= 0
n→∞
P
then an diverges.
 Why?

• Suppose limn→∞ an = L > 0

• There is a fixed M so that for all
n ≥ M:
an ≈ L

• aM + · · · + aN ≈ (N − M)L → ∞ if
N→∞
 Example
∞
X n3
Check convergence of
n3 + n + 1
n=1
Solution:
n3
lim = 1 6= 0
n→∞ n3 + n + 1

so by Divergence test the series diverges

 Example
∞
X n3
Check convergence of
n3 + n + 1
n=1
Solution:
n3
lim = 1 6= 0
n→∞ n3 + n + 1

so by Divergence test the series diverges

 Be careful !!
• The reverse is NOT true
• If lim an = 0 we can NOT say that
P
an converges
• If lim an 6= 0 then Diverges
P
• If lim an = 0 then an can diverge or
converge
 Be careful !!
• The reverse is NOT true
• If lim an = 0 we can NOT say that
P
an converges
• If lim an 6= 0 then Diverges
P
• If lim an = 0 then an can diverge or
converge
Test for positive series
 (2) Integral Test
Let f (x) be a positive, non-increasing
function, and define an = f (n). Then
X∞ Z ∞
an ≈ f (x)dx
n=1 1

means that the series and the integral

converges or diverges together.
 Example
∞
X 1
n ln n
n=2
1
f (x) = is positive and non-increasing.
x ln x
The integral
Z ∞
1
dx = ln ln x|∞
2 = ∞
2 x ln x
so the series also diverges.
 Example
∞
X 1
n ln n
n=2
1
f (x) = is positive and non-increasing.
x ln x
The integral
Z ∞
1
dx = ln ln x|∞
2 = ∞
2 x ln x
so the series also diverges.
 p−series

Let p be a positive integer. The p−series is

∞
X 1
np
n=1

By integral test, p−series converges if

p > 1, diverges if p ≤ 1
 Example: Harmonic series
∞
X 1
• The series is called harmonic
n
n=1
series
• Comparison doesn’t work, need to use
integral
1
• Function is positive and decreasing
x
R∞ 1
• 0 x dx = ln x|∞
0 = ∞

• so Harmonic series converges

 (3) Comparison Test

P P
Given 2 positive series an and bn.
Suppose an < bn for n > N for some fixed
N Then X X
an < bn
 P P
If an diverges then bn diverges.
P P
If bn converges then an converges.
 Example

∞
X 1
2n + 1
n=1

We have 1/(2n + 1) < 1/2n, so we compare

with (1/2)n. This one converges because
P

it is geometric series with ratio 1/2. So the

series converges.
 Example
∞
X ln n
converges or diverges ?
n
n=1

ln n 1
• >
n n
∞
X 1
• diverges by p series
n
n=1
∞
X ln n
• diverges by comparison
n
n=1
 (4) Limit comparison
P P
Given positive series an and bn and
an
lim =L
n→∞ bn

P P
• if 0 < L < ∞ then an ≈ bn
P P
• if L = 0 then an < bn
P P
• if L = ∞ then an > bn
 Example
∞
X n2 + 3n − 1
Check convergence of √ √
n2 n + n − 2
n=1

• Only care about the highest power

n2 1
• Compare with bn = 2√ = √
n n n
P
• lim an/bn = 1 and bn diverges, so
P
an diverges.
 Example
∞
X 1
Check convergence of
2n − 1
n=1

1 1
• > comparison test not good
2n − 1 2n
1/2n − 1
• lim =1
n→∞ 1/2n
X 1 X 1
• converges b/c
2n − 1 2n
converges
 Non positive series
P
Consider a series n an .
P
• an > 0 for all n > N: an is positive
series
• Can use tests: comparison, limit
comparison, integral
• What if an < 0 for infinitely many
values of n?
 Non positive series
P
Consider a series n an .
P
• an > 0 for all n > N: an is positive
series
• Can use tests: comparison, limit
comparison, integral
• What if an < 0 for infinitely many
values of n?
 Non positive series
P
Consider a series n an .
P
• an > 0 for all n > N: an is positive
series
• Can use tests: comparison, limit
comparison, integral
• What if an < 0 for infinitely many
values of n?
 Example
∞
X (−1)n
n
n=1

• Has infinitely many values of n so that

an > 0
• Has infinitely many values of n so that
an < 0
 Alternating series
P
• Series an is called alternating
series if an change sign at every n

an · an+1 < 0 for all n > N for some N.

• There is a convergence test that only

apply for Alternating series
(5) Alternating Series Test
Suppose {an} is an alternating series that
satisfies
• the sequence {|an|} decreasing :
|an+1| < |an| for all n ≥ N
( for some N)
• lim |an| = 0
n→∞
P
Then an converges.
(5) Alternating Series Test
Suppose {an} is an alternating series that
satisfies
• the sequence {|an|} decreasing :
|an+1| < |an| for all n ≥ N
( for some N)
• lim |an| = 0
n→∞
P
Then an converges.
 Example
∞
X (−1)n
Alternating harmonic series
n
n=1

(−1)n+1 (−1)n
• Alternating: <0
n+1 n
1 1
• Decreasing: <
n+1 n
1
• lim = 0
n→∞ n

• Converges by Alternating series test

 Example
∞
X (−1)n
Alternating harmonic series
n
n=1

(−1)n+1 (−1)n
• Alternating: <0
n+1 n
1 1
• Decreasing: <
n+1 n
1
• lim = 0
n→∞ n

• Converges by Alternating series test

 Example
∞
X (−1)n
Alternating harmonic series
n
n=1

(−1)n+1 (−1)n
• Alternating: <0
n+1 n
1 1
• Decreasing: <
n+1 n
1
• lim = 0
n→∞ n

• Converges by Alternating series test

 Proof Idea
Assume N = 1 and a1 > 0.
• a2k < 0, a2k+1 > 0
• S2k+2 > S2k
• S2k+1 < S2k−1
• S2 < S4 < · · · < S2k < S2k−1 < · · · <
S1
 Proof Idea
Assume N = 1 and a1 > 0.
• a2k < 0, a2k+1 > 0
• S2k+2 > S2k
• S2k+1 < S2k−1
• S2 < S4 < · · · < S2k < S2k−1 < · · · <
S1
 Proof Idea
Assume N = 1 and a1 > 0.
• a2k < 0, a2k+1 > 0
• S2k+2 > S2k
• S2k+1 < S2k−1
• S2 < S4 < · · · < S2k < S2k−1 < · · · <
S1
 Proof Idea
Assume N = 1 and a1 > 0.
• a2k < 0, a2k+1 > 0
• S2k+2 > S2k
• S2k+1 < S2k−1
• S2 < S4 < · · · < S2k < S2k−1 < · · · <
S1
 Error

P
Suppose an is an alternating series that
converges to a value S, then

|S − Sn| < |Sn+1 − Sn| < |an+1|

for every n.
 Example
∞
X (−1)n
Suppose converges to number S.
1 + 2n
n=1
How large is N so that the error |S − SN | is
less than 10−3?
Solution: |S − SN | < |aN+1| so we need

1/(1 + 2N+1) < 10−3

which implies that N = 9

 Example
∞
X (−1)n
Suppose converges to number S.
1 + 2n
n=1
How large is N so that the error |S − SN | is
less than 10−3?
Solution: |S − SN | < |aN+1| so we need

1/(1 + 2N+1) < 10−3

which implies that N = 9

 Example
∞
X (−1)n
Suppose converges to number S.
1 + 2n
n=1
How large is N so that the error |S − SN | is
less than 10−3?
Solution: |S − SN | < |aN+1| so we need

1/(1 + 2N+1) < 10−3

which implies that N = 9

 Absolute convergence

P
• Given an not positive.
P
• Consider |an|: positive series
P
• If |an| converges then we say
P
an absolutely converges
 Theorem

P
If an absolutely converges then it is also
converges
 Proof
P
Suppose |an| converges. Let

bn = an + |an|

2|an| ≥ bn ≥ 0 ∀n
P P
So if |an| converges, bn converges
X X X
an = bn − |an|

also converges.
Conditional convergence

P P
If an converges but |an| diverges then
P
we say that an conditionally converges
 Alternating harmonic
series
∞
X X (−1)n
an =
n
n=1
∞
X X 1
|an| =
n
n=1
P P
|an| diverges, but an converges
 (6) Ratio Test
P
Given an and
 
 an+1 
ρ = lim  
n→∞ an 

P
• if ρ < 1 then an absolutely
converges.
P
• if ρ > 1 then an diverges.
• if ρ = 1 then no conclusion.
 Example
∞
X (−1)nn!
nn
n=1

(n + 1)! nn
ρ = lim
n→∞ (n + 1)n+1 n!
 n
n 1
= lim = <1
n→∞ n + 1 e
so the series absolutely converges.
 Example
∞
X (−1)nn!
nn
n=1

(n + 1)! nn
ρ = lim
n→∞ (n + 1)n+1 n!
 n
n 1
= lim = <1
n→∞ n + 1 e
so the series absolutely converges.
 (7) Root Test
P
Let an positive and
p
n
ρ = lim |an|
n→∞

P
• if ρ < 1 then an absolutely
converges.
P
• if ρ > 1 then an diverges.
• if ρ = 1 then no conclusion.
 Example

∞
X 2n+1
nn
n=1

2(n+1)/n
ρ = lim =0<1
n→∞ n
so the series converges
 Which test to use?
• 1st step: if lim an 6= 0 then diverges
If lim an = 0 then continue
• Integral: function with ln
• Comp./Limit comp.: Rational, trig
• Root: (.)n
• Ratio: n!
• Alternate sign: Alternating series test
 Homework 1

11.1: 15, 35, 43, 79, 80

11.2: 21, 32, 43
11.3: 23, 25
11.4: 7, 17, 23
11.5: 15, 17