The Fibonacci Sequence

Michael B. Williams
Abstract
This note addresses two questions relating to the Fibonacci sequence.
First, what is a closed formula for the n-th term in the sequence? Second,
what is the limit of the ratio of two consecutive terms in the sequence?
We provide several methods of answering each question.
1 Introduction
The Fibonacci sequence is one of the most familiar sequences; it is
0, 1, 1, 2, 3, 5, 8, 13, . . . .
That is, after the second term, each term is obtained by adding the previous
two terms. If we call this sequence {F
n
}, then we have the following recursive
denition
F
n
=
_

_
0 if n = 0
1 if n = 1
F
n1
+F
n2
if n 2
.
A natural question is whether there exists a closed-form expression describing
F
n
, that is, can we nd a formula that does not involve any other terms of
the sequence? A second question involves asymptotic behavior. Obviously the
sequence grows without bound, but let us consider ratios of consecutive terms:
F
2
F
1
=
1
1
= 1,
F
3
F
2
=
2
1
= 2,
F
4
F
3
=
3
2
= 1.5,
F
5
F
4
= 1.6,
F
6
F
5
=
8
5
= 1.6, etc.
Upon continued inspection, this new sequence would appear to have a (nite)
limit, but what is it? We shall see that the answers to both questions are
intimiately related to the quantity known as the golden ratio, which occurs
frequently in mathematics, and indeed in nature.
1
2 A Closed Formula for F
n
2.1 The Linear-Algebraic Method
The recursive denition of the Fibonacci sequence implies that pairs of consec-
utive terms are important:
..
. . . , F
n
, F
n+1
,
. .
F
n+2
, F
n+3
. .
, . . .
It is therefore natural to consider pairs as 2-vectors, e.g. (F
n+1
, F
n
)
T
. In fact,
the recursive relation above can be described in this manner:
_
1 1
_
_
F
n+1
F
n
_
= F
n+1
+F
n
= F
n+2
.
More useful is considering two consecutive vectors in a matrix:
_
F
n+2
F
n+1
F
n+1
F
n
_
,
so that when n = 0 we have
_
F
2
F
1
F
1
F
0
_
=
_
1 1
1 0
_
.
Call this matrix M. Powers of this matrix encapsulate the entire Fibonacci
sequence; in fact, we claim that for all integers n 1, we have
M
n
=
_
F
n+1
F
n
F
n
F
n1
_
.
The proof is by induction on n. The base case n = 1 was demonstrated
above. If we suppose that the statement holds for some n, then we need to
prove it is true for n + 1. This is true by a simple computation:
M
n+1
= MM
n
=
_
1 1
1 0
__
F
n+1
F
n
F
n
F
n1
_
=
_
F
n+1
+F
n
F
n
+F
n1
F
n+1
F
n
_
=
_
F
n+2
F
n+1
F
n+1
F
n
_
,
which is what we needed to show. The proof is complete.
From this we obtain a formula describing the n-th term in the sequence:
M
n
_
1
0
_
=
_
F
n+1
F
n
_
. (1)
2
That is, F
n
= (M
n
)
12
= (M
n
)
21
. As it stands, this formula is not immediately
useful; powers of nondiagonal matrices are dicult to compute. Therefore, in
order to use this equation to describe F
n
we must diagonalize M. First, we nd
the eigenvalues, which are the roots of the characteristic polynomial:
|I M| =

1 1
1

= ( 1) 1 =
2
1.
Setting
2
1 = 0 and using the quadratic formula, we see that
=
1

5
2
.
Let us call
=
1 +

5
2
, =
1

5
2
.
Note that is the golden ratio, and that these roots satisfy the following rela-
tions:

2
= + 1,
2
= + 1, =

5, + = 1,
1

= ,
1

= .
Now we nd the eigenvectors. Row reduction gives
I M =
_
1 1
1
_

_
( + 1)
1
_
=
_

2
1
_

_
1
1
_

_
1
0 0
_
.
This implies that an eigenvector corresponding to the eigenvalue is (, 1)
T
. A
similar computation shows that an eigenvector corresponding to the eigenvalue
is (, 1)
T
. We put these eigenvectors into a change-of-basis matrix:
T =
_

1 1
_
.
The inverse of this matrix is
T
1
=
1

_
1
1
_
=
1

5
_
1
1
_
.
Then
M = T

MT
1
,
where the diagonalization of M is

M =
_
0
0
_
.
3
Now we can easily compute powers of M:
M
n
= [T

MT
1
]
n
= [T

MT
1
][T

MT
1
] [T

MT
1
]
= T

M
n
T
1
=
1

5
_

1 1
__

n
0
0
n
__
1
1
_
.
Finally, the desired formula is obtained using this, and (1):
_
F
n+1
F
n
_
= M
n
_
1
0
_
=
1

5
_

1 1
__

n
0
0
n
__
1
1
__
1
0
_
=
1

5
_

1 1
__

n
0
0
n
__
1
1
_
=
1

5
_

1 1
__

n

n
_
=
1

5
_

n+1

n+1

n
_
.
We conclude that
F
n
=
1

5
_

n
_
=
1

5
__
1 +

5
2
_
n

_
1

5
2
_
n
_
.
2.2 The Generating Function Method
Recall that the generating function for a sequence {a
n
} is
G(x) =

n=0
a
n
x
n
.
A simple but important example is the generating function for the sequence
with a
n
= 1 for all n:
G(x) =

n=0
x
n
=
1
1 x
.
4
Now, we nd the generating function of the Fibonacci sequence:
G(x) =

n=0
F
n
x
n
= 0 +x +

n=2
F
n
x
n
= x +

n=2
(F
n1
+F
n2
)x
n
= x +x

n=2
F
n1
x
n1
+x
2

n=2
F
n2
x
n2
= x +x

n=1
F
n
x
n
+x
2

n=0
F
n
x
n
= x +x(G(x) 0) +x
2
G(x).
This relation allows us to write
G(x) =
x
1 x x
2
=
x
(x +)(x +)
.
Using partial fractions, we nd that
G(x) =

5
1
x +
+

5
1
x +
=

5
1

1
1 +
x

5
1

1
1 +
x

=
1

5
_
1
1 x
+
1
1 x
_
=
1

5
_

n=0

n
x
n

n=0

n
x
n
_
=

n=0
1

5
(
n

n
)x
n
.
Equating coecients, we again conclude that
F
n
=
1

5
(
n

n
) =
1

5
__
1 +

5
2
_
n

_
1

5
2
_
n
_
.
5
2.3 The Inductive Method
This method assumes that we are somehow able to guess the correct forumla,
for we use induction on n to prove that it is indeed correct. Our two base cases
are trivial:
n = 0 :
1

5
(
0

0
) = 0 = F
0
,
n = 1 :
1

5
(
1

1
) = 1 = F
1
.
Assume that the formula holds for arbitrary n. Then for n + 1 we have
F
n+1
= F
n
+F
n1
=
1

5
(
n

n
) +
1

5
(
n1

n1
)
=
1

5
(
n
+
n1

n1
)
=
1

5
_

n1
( + 1)
n1
( + 1)
_
=
1

5
_

n1

n1

2
_
=
1

5
_

n+1

n+1
_
,
and so we are done.
2.4 An Intesting Connection
Here we remark that the golden ratio satises a relation similar to that of the
terms in the Fibonacci sequence, namely

n+2
=
n+1
+
n
.
Again, the proof is by induction on n. The base cases are veried as follows:
n = 0 :
2
= + 1 =
1
+
0
,
n = 1 :
3
=
2
= ( + 1) =
2
+
1
.
Suppose that the formula holds for arbitrary n. We show that it also holds for
n + 1:

n+1
=
n
= (
n1
+
n2
)
=
n
+
n1
),
so the proof is complete.
6
3 The Limit of F
n+1
/F
n
as n
3.1 The Elementary Method
This method is elementary in that it involves only basic calculus limits and
clever factoring; it does require the closed formula for F
n
, however:
F
n+1
F
n
=
(
n+1

n+1
)/

5
(
n

n
)/

5
=
( )(
n
+
n1
+ +
n1
+
n
)
( )(
n1
+
n2
+ +
n2
+
n1
)
=

n
i=0

ni

n1
i=0

n1i

i
=

n
i=0

ni

n1
i=0

ni

i
=
_
1 +

n

n1
i=0

ni

i
_
=
_
1 +
1

n1
i=0

ni

in
_
=
_
1 +
1

n1
i=0
(/)
ni
_
[1 + 0]
=
1 +

5
2
,
where we used the fact that (/)
ni
as n , for all i n.
3.2 The Recurrance Method
This method does not require a closed formula for F
n
. Setting a
n
= F
n+1
/F
n
,
we have
a
n
=
F
n+1
F
n
=
F
n
+F
n1
F
n
= 1 +
F
n1
F
n
= 1 +
1
a
n1
.
Say a
n
L. Then letting n in the above equation, we have
L = 1 +
1
L
.
This gives us the polynomial L
2
L 1 = 0, and we know the roots are and
. Since < 0, it cannot be the limit, because a
n
> 0 for all n. We conclude
that is the limit, so
F
n+1
F
n

1 +

5
2
.
7