# Infinite Sequences and Series

A sequence is an ordered set of numbers { } { }  
n n
a a a a a a a
5 4 3 2 1
= ,
where a
n
denotes the nth term. The sequence converges if
n
n
a
· ÷
lim = L (and L is finite),
otherwise it diverges.
Note: if
n
n
a
· ÷
lim

Theorem 1 Sandwich Theorem: if a
n
< b
n
< c
n
, and
n
n
a
· ÷
lim =
n
n
c
· ÷
lim = L, then
n
n
b
· ÷
lim = L also.

Theorem 2 If | | lim
n
n
a
· ÷
= 0, then
n
n
a
· ÷
lim = 0.

Theorem 3 If f is a function defined for real numbers x, ) ( lim x f
x · ÷
= L and f(n) =
a
n,
then
n
n
a
· ÷
lim = L also. (This allows the use of l'Hopital's rule to find the limits of
sequences.)

Theorem 4 A geometric sequence {x
n
} has limit zero for |x| < 1, but
n
n
x
· ÷
lim

does not
exist for |x| > 1 and x = -1.
n
n
1 lim
· ÷
= 1 when x = 1

A sequence is increasing if for all values of n, a
n
< a
n+1
; a sequence is decreasing
if for all values of n, a
n
> a
n+1
. A monotonic sequence is either increasing or
decreasing. A sequence is bounded above if there exists a number M such that M > a
n
and bounded below if there exists a number m such that m < a
n.
A bounded
sequence is bounded both above and below.

Theorem 5 A bounded monotonic sequence converges.

Given a sequence {a
k
} , if we add the terms we get a series. An infinite series is
denoted by:
1
k
k
a
·
=
¯
Given a series
1
k
k
a
·
=
¯
, we can define the partial sums
1
n
n k
k
S a
=
=
¯
These form a new sequence {S
n
}. If {S
n
} converges to S (that is
lim
n
n
S
÷·
= S) then we say the series converges and S is called the sum of the series.
Otherwise the series diverges. Although it is sometimes possible to find the exact value
of a series, in general this is a very difficult problem. The first step with a series is to
determine whether it converges or diverges. If it does converge, the sum can usually be
estimated numerically.
Geometric Series:
0
k
k
x
·
=
¯
converges if |x| < 1 and diverges otherwise.
The sum of this series is 1/(1 - x).
Harmonic Series:
0
1
k
k
·
=
¯
The partial sums are usually denoted by
1
1
n
n
k
H
k
=
=
¯
Since
H
n
grows like ln(n), this series diverges.

Theorem 6 If
1
k
k
a
·
=
¯
converges, then
k
k
a
· ÷
lim = 0.

Theorem 7 If
1
k
k
a
·
=
¯
and
1
k
k
b
·
=
¯
both converge, then
1
k k
k
a b
·
=
+
¯
converges also and
1 1 1
k k k k
k k k
a b a b
· · ·
= = =
+ = +
¯ ¯ ¯
Also for any constant c,
1 1
k k
k k
ca c a
· ·
= =
=
¯ ¯

Tests for Convergence

Nth Term Divergence Test If lim a
k

1
k
k
a
·
=
¯
diverges.
The converse of this is false. Just because the
k
k
a
· ÷
lim

= 0 the series
1
k
k
a
·
=
¯

may not converge. Write an example of this here: ____________

Integral Test: Let f be a positive, decreasing and continuous function for x > b.
If a
k
= f(k) then
1
k
k
a
·
=
¯
converges whenever ( )
b
f x dx
·
í
converges and
1
k
k
a
·
=
¯

diverges whenever ( )
b
f x dx
·
í
diverges.
P - series Test:
1
1
p
k
k
·
=
¯
converges for p > 1 and diverges for p < 1.

Comparison Test: If 0 < a
n
< b
n
then: if
1
k
k
b
·
=
¯
1
k
k
a
·
=
¯
; if
1
k
k
a
·
=
¯
diverges so does
1
k
k
b
·
=
¯

Limit Comparison Test: If 0 < a
k
, 0 < b
k
,
k k
k
b a
· ÷
lim = c and c is positive and finite
then:
1
k
k
b
·
=
¯
1
k
k
a
·
=
¯
either both converge or they both diverge.
If
k k
k
b a
· ÷
lim = 0 and
1
k
k
b
·
=
¯
converges, so does
1
k
k
a
·
=
¯

If
k k
k
b a
· ÷
lim = · and
1
k
k
b
·
=
¯
diverges, so does
1
k
k
a
·
=
¯

Common series used for comparisons are geometric series and p-series.

If 0 < a
k
, then   + ÷ + + ÷ + ÷ + ÷ = ÷
+
·
=
+
¯
n
n
k
k
k
a a a a a a a a
1
6 5 4 3 2 1
1
1
) 1 ( ) 1 (
is called an alternating series.
Alternating Series Test: (Leibniz’ Test)
If {a
k
} is decreasing and
k
k
a
· ÷
lim = 0 then
1
1
( 1)
k
k
k
a
·
+
=
÷
¯
converges.

A series
1
k
k
a
·
=
¯
is called absolutely convergent if
1
k
k
a
·
=
¯
converges.
A series
1
k
k
a
·
=
¯
may converge while
1
k
k
a
·
=
¯
diverges. Example? _________

Theorem 8 Convergence of
1
k
k
a
·
=
¯
implies convergence of
1
k
k
a
·
=
¯
. This means that
all of the tests above can be applied to { |a
n
| } if {a
n
} contains both positive and
negative terms.

Ratio Test: Given a series
1
k
k
a
·
=
¯
, let
1
lim
n
n
n
a
r
a
+
÷·
= If r < 1 then
1
k
k
a
·
=
¯
converges
absolutely; if r > 1 then
1
k
k
a
·
=
¯
diverges and if r = 1 no conclusion can be drawn.

Root Test: Given a series
1
k
k
a
·
=
¯
, let lim | |
n
n
n
r a
÷·
= If r < 1 then
1
k
k
a
·
=
¯

converges absolutely; if r > 1 then
1
k
k
a
·
=
¯
diverges and r = 1 no conclusion can be
drawn.
Power Series

The function f(x) =
0
( )
k
k
k
a x c
·
=
÷
¯
is called a power series centered at c or about c.
Of course, this function only makes sense at those points x where the series converges.
Power series, if they converge at more than one point, converge absolutely on an open
interval centered at c. The distance from the center c to the endpoints of the interval is
called the radius of convergence R. If the interval is the whole real line then R = ·. If
R is finite, then the power series diverges for |x - c| > R. Convergence at the endpoints of
the interval x = c ± R must be checked separately.

Our goal is to find a power series representation for a given function f(x).
Suppose f(x) is the given function; we want to determine the {a
k
} so that:
f(x) =
0
( )
k
k
k
a x c
·
=
÷
¯
for |x - c| < R. By evaluating f and all it’s derivatives at c, it
can be shown that a
k
must satisfy:
( )
( )
!
k
k
f c
a
k
=
Thus the power series for f(x) can only be the Taylor series of f(x) about x = c.
f(x) =
( )
2
0
( )
( ) ( ) '( )( ) ''( )( ) / 2!
!
k
k
k
f c
x c f c f c x c f c x c
k
·
=
÷ = + ÷ + ÷ +
¯

When c = 0, we have a special case of Taylor series called the MacLaurin Series:
. . . x
f
x
f
x f ' ) f( f(x) +
' ' '
+
' '
+ + =
3 2
! 3
) 0 (
! 2
) 0 (
) 0 ( 0
All of this assumes that the given function f(x) has a power series representation.
How can we be sure that this is the case? Here is an example of a function which is not
equal to its Taylor series except at x = 0: f(x) =
2
1 x
e
÷
for x = 0, and f(0) = 0. It can
be shown that all of the derivatives of f at zero are zero. Thus the Taylor series for f is the
function T(x) = 0 for all x, while f(x) is never zero when x = 0.
Taylor's formula (with remainder) tells us when f(x) is equal to its Taylor series.
Suppose that f(x) has n+1 derivatives in a interval I containing the points x and c. Then
there exists a number z between x and c such that:
f(x) =
2 ( )
( ) ( ) '( )( ) ''( )( ) / 2! ( )( ) / ! ( )
n n
n
f x f c f c x c f c x c f c x c n R x = + ÷ + ÷ + + ÷ +
where
( 1) 1
( ) ( )( ) /( 1)!
n n
n
R x f z x c n
+ +
= ÷ + ( )
n
R x is called the remainder term. Note that
( )
n
R x looks just like the (n+1)st term in the Taylor series except that the (n+1)st
derivative is evaluated at a point other than c. Now f(x) = ( )
n
T x + ( )
n
R x and f(x) is
equal to its Taylor series expansion provided that lim ( ) 0
n
n
R x
÷·
= .