Applied Mathematical Sciences, Vol. 5, 2011, no.

64, 3173 - 3181

On the Pell, Pell-Lucas and Modified Pell Numbers

By Matrix Method
Ahmet Dasdemir

Department of Mathematics
Faculty of Arts and Sciences, Sakarya University
54100 Serdivan, Sakarya, Turkey
ahmetdasdemir37@gmail.com

Abstract
In this paper, we consider the usual Pell, Pell-Lucas and Modified
Pell sequences, and we define some new matrices, which are based on
Modified Pell and Pell-Lucas numbers, with determinants equal to 2n
or 23n .This unique properties provide very important formulas like the
Simpson F ormula of Modified Pell and Pell-Lucas numbers. Also we
show that the Modified Pell and Pell-Lucas numbers have the generating
matrices.Further, we present some elementary identities between them
by these matrices.

Mathematics Subject Classification: 11B37; 15A15

Keywords: Pell Number, Modified Pell Number, Pell-Lucas Number,

Generating Matrix, Simpson Formula

1 Introduction
In the literature, especially in mathematics and physics, there are a lot of in-
teger sequences, which are used in almost every field modern sciences. Admit-
tedly, the Fibonacci sequence is one of the most famous and curious numerical
sequences in mathematics and have been widely studied from both algebraic
and combinatorial prospectives.Also, there is the Pell sequence, which is as
important as the Fibonacci sequence.The Pell sequence {Pn } are defined by
recurrence Pn = 2Pn−1 + Pn−2 , n ≥ 2 with P0 = 0 and P1 = 1 and the
Pell-Lucas sequence {Qn } and Modified Pell sequence {qn } by the same re-
currence but with initial conditions Q0 = Q1 = 2 and q0 = 1 and q1 = 1,
respectively.Infact,
Qn = 2qn . (1)
3174 A. Dasdemir

Consequently, the known properties of {Qn } can be easily written for {qn }.
Thereby, a study of {qn } involves inevitably familiarity with {Qn }.Further
details about Pell, Pell-Lucas and Modified Pell sequences can be found in
[3, 5, 9, 13].For instance, in [3], Horadam gave some results as follows:
qn = Pn+1 − Pn , (2)
qn+1 = Pn+1 + Pn (3)
and
qn+1 + qn
Pn+1 = . (4)
2
Explicit Binet forms for {Pn }, {Qn } and {qn } are
αn − β n
Pn = , (5)
α−β
Qn = αn + β n , (6)
and
αn + β n
qn = (7)
α+β
where α and β are the roots of the characteristic equation x2 −2x−1 = 0.[1, 11]
The some of elements of the Pell, Pell-Lucas and Modified Pell sequences
are given by the following table;
n 0 1 2 3 4 5 6 7 8 9 ···
Pn 0 1 2 5 12 29 70 169 408 985 ···
Qn 2 2 6 14 34 82 198 478 1154 2786 ···
qn 1 1 3 7 17 41 99 239 577 1393 ···
Table 1. The values of {Pn }, {Qn } and {qn }.
In the last decades, some mathematicians have studied to find miscella-
neous affinities between the matrices and linear recurrences.They want to ob-
tain always the terms of linear recurrences by matrices, which are said to be
the Generating Matrices.Comparable matrix generators have been considered
previously for the Fibonacci and Pell sequences[7, 8, 12]. In [12], the following
property of the nth power of the Q-matrix is found
 
Fn+1 Fn
Qn = , (8)
Fn Fn−1
 
1 1
where Q = .Similarly, in [8] author showed that the members of the
1 0
Pell sequence can be derived by a matrix representation:
 
n Pn+1 Pn
M = , (9)
Pn Pn−1
Pell, Pell-Lucas and modified Pell numbers 3175

 
2 1
where M = .
1 0
The matrix method has also played an important and effective role stem-
ming from Number Theory.As an example of the usage of the matrix approach,
we can exemplify to obtain the Simpson F ormula for the Pell numbers[1],
namely,
Pn+1 Pn−1 − Pn2 = (−1)n , (10)
which may, of course, be established by means of the Binet form (5).
The permanent of an n-square matrix A = (aij ) is defined by
n
 
perA = aiσ(i) , (11)
σ∈Sn i=1

where the summation extends over all permutations σ of the symmetric group
Sn .The most important applications of permanents are in the areas of physics
and chemistry.One can find more applications of permanents in [11].The per-
manent of a matrix is analogous to the determinant, where all of the signs used
in the Laplace expansion of minors are positive.
The purpose of this article is to urge a greater use of the properties be-
tween the matrices and linear recurrences composed of the Pell, Pell-Lucas and
Modified Pell sequences.Emphasis in this paper will be considered a few square
2 × 2 matrices.Then we investigate the identities of these matrices.Even, we
shall introduce different matrices to obtain new results.Also we will show that
the identities presented before can be produce by the aid of them.

2 The Pell and Modified Pell Numbers

In this section, we introduce a new matrix whose rows are [3,1] and [1,1].
Clearly,
 
3 1
N= . (12)
1 1
Considering the matrix N and the equations (2-3), we can easily see
       
qn+2 Pn+1 Pn+2 1 qn+1
=N or = N (13)
qn+1 Pn Pn+1 2 qn

and ⎧  
⎪
0
⎪
⎪ 1
⎪
⎨ Pn = n 1 0 Nn , if n even,
22

1 

 (14)
⎪
⎪ 0
⎪
⎪ qn = 1
1 0 Nn , if n odd.
⎩ n−1
1
2 2
3176 A. Dasdemir

We also define a matrix E by

 
1 1
E= . (15)
1 −1
Moreover from the equations (2-4), we can feasible write
       
qn+1 Pn+1 Pn+1 1 qn+1
=E and = E . (16)
qn Pn Pn 2 qn
Note that the matrix E acts as the transformer (or reflector) mission between
the Pell and Modified Pell numbers.Furher there is a very interesting property
of the matrix E. Its powers are consecutively equal to multiplication of itself
n−1 n
by 2 2 and multiplication of the identity matrix by 2 2 . Clearly,
⎧  
⎪
⎪ n−1 1 1
⎪
⎪ 2 2 if n odd,
⎪
⎪ 1 −1
⎨
En =   (17)
⎪
⎪
⎪
⎪ 1 0
⎪
⎪ 22
n
if n even.
⎩
0 1
First we start with the following theorem in connection with the matrix N.
Theorem 2.1 Let the N be a matrix as in (12). For all positive integers n
and m, the following determinantal equalities are held:
⎧  
⎪
⎪ n Pn+1 Pn
⎪
⎪ 2 2 if n even,
⎪
⎪ Pn Pn−1
⎨
i. N n =  
⎪
⎪
⎪
⎪ qn+1 qn
⎪
⎪ 2
n−1
if n odd,
⎩ 2
qn qn−1
ii. det (N n ) = 2n ,
iii. per (N n ) = 2n−1 (q2n + 1) ,
iv. N n+m = N n N m ,
v. N n−m = N n N −m ,
Proof.
(i)(Induction on n) We will use the induction method on odd n and even n,
separately.Let n be an odd integer.The case for n = 1 is clear.Now we suppose
that it is correct for odd n integer.Then we show that it is correct for n + 2.By
the definition of Modified Pell sequence, we can write
 
n+2 n 2 n+1 qn+3 qn+2
N =N N =2 2 . (18)
qn+2 qn+1
Pell, Pell-Lucas and modified Pell numbers 3177

Thus, this part of the assertion for all odd n integers is to be true.Now let n be
an even integer.Then the case for n = 2 is clear.We assume that it is correct
for even n.Now we show that it is to be true for n + 2.Then, by the definition
of Pell sequence, we compute
 
n+2 n 2 n+2 Pn+3 Pn+2
N =N N =2 2 (19)
Pn+2 Pn+1

(ii) (Induction on n) When n = 1, (ii) is seen to be true.We assume that

it is to be true for n.Then we show that the equation holds for n + 1.

det N n+1 = (det N n ) det (N) = 2n 2 = 2n+1 , (20)

and the proof of (ii) is completed.

(iii) Using the Binet formulæ of the Pell and Modified Pell sequences and
from F, the proof of (ii) can be easily seen.
(iv − v) If we use a property of matrix multiplication, the proof of (iii) and
(iv) are readily seen.
We now zoom in the properties of the matrix N. Its characteristic equation
is
x2 − 4x + 2 = 0 (21)
and the eigenvalues of the nth power of N are
⎧ n
 √ √
⎪
⎪
⎨ 2 2 q n + P n 2, qn − P n 2 , if n even,
(λ, μ) = (22)
⎪
⎪
⎩ 2 n−1
 √ √
2 qn + Pn 2, qn − Pn 2 , if n odd.

To put it more simply, we write

⎧ n
⎪
⎨ 2 2 (αn , β n ) , if n even,
(λ, μ) = (23)
⎪
⎩ n−1
n n
2 2 (α , β ) , if n odd,

where α and β are the roots of the characteristic equation x2 − 2x − 1 = 0

(so-called Pell equation).
Since det N = 0, it is a invertible matrix. Thus, we can easily obtained
⎧  
⎪
⎪
⎪ 1 Pn−1 −Pn
⎪
⎪ n if n even,
⎪
⎨ 22 −Pn Pn+1
N −n =   (24)
⎪
⎪
⎪
⎪ qn−1 −qn
⎪
⎪ 1
if n odd.
⎩ n+1
2 2 −qn qn+1
As an example of the usage of elementary identities of matrices, we can
write the following theorem.
3178 A. Dasdemir

Theorem 2.2 For all positive integers n and m, the following determinan-
tal equalities are held:

i. Pn+1 Pn−1 − Pn2 = (−1)n ,

ii. Pn+1 Pn−1 + Pn2 = 12 (q2n + (−1)n ) ,

iii. qn+1 qn−1 − qn2 = 2(−1)n+1 ,

iv. qn+1 qn−1 + qn2 = q2n + (−1)n+1 ,

v. Pm+n = Pm Pn+1 + Pm−1 Pn ,

vi. 2Pm+n = qm qn+1 + qm−1 qn ,

vii. qm+n = Pm qn+1 + Pm−1 qn ,

viii. (−1)m+1 Pn−m = Pn qm−1 − qn−1 Pm ,

ix. (−1)m+1 qn−m = Pn qm+1 − Pn+1 qm ,

x. 2(−1)m+1 Pn−m = qn qm+1 − qn+1 qm ,

xi. P2n = 2Pn qn .

Proof. The proof of theorem is easily shown from Theorem 2.1.iv-v.

Note that the matrix N gives unlimited opportunities for us because it
allows to get an infinite numbers of fundamental identities involved the Pell
and Modified Pell numbers between themselves. We can also conclude that
Theorem 2.2.i-iii are the Simpson F ormula for the Pell and Modified Pell
numbers, respectively. But, they are found by Horadam beforehand, in [1, 3].
Here we only show that they can be obtained with the more different way from
method used in the papers of Horadam.

Theorem 2.3 Let E be a matrix as in (15). Then

   
n qn+1 qn qn+1 qn
EM = and E = 2M n (25)
qn qn−1 qn qn−1

where the matrix M is defined as before.

Proof. By the equations (2-3), we can easily see that these relationships
between the matrices E and M.Thus, the proof is completed.
Now we define a new matrix as follows:
 
1 1
L= . (26)
0 −1
Pell, Pell-Lucas and modified Pell numbers 3179

In mathematics, an involutary matrix is a matrix that is its own inverse.That

is, matrix A is an involution if A2 = I. One of the tree classes of elementary
matrix, namely the row-interchange matrix, that which represents multiplica-
tion of a row or column by -1, is also involutory, it is in fact a trivial example
of a signature matrix, all of which are involutory. Involutary matrices are all
square roots of the identity matrix. This is simply a consequence of the fact
that any nonsingular matrix multiplied by its inverse is the identity.
Considering the above information, we can write that the matrix L is a
involutory matrix. Thus, we can write
L2 = I (27)
Let us define a 2 × 2 matrix Hn as follows:
 
Pn+1 qn
Hn = . (28)
Pn qn−1
Then we have the following theorem.

Theorem 2.4 Let L and Hn be respectively as in (26) and (28). Then

M n = Hn · L and Hn = M n · L (29)
where M is defined as before.

Proof. Considering the equation 2, the proof of the first equation is easily
seen. Also, from the fact that L is an involutory matrix, the proof of the
second equation is easily seen.Thus, the proof is completed.
We can in fact extent the above theorem, namely for n, m ∈ Z +
Hn+m = Mn · Hm . (30)
Its proof can easily be seen by the basic matrix operations.

3 The Pell and Pell-Lucas Numbers

In this section we will consider on the findings of Section 2.Then we will pre-
dispose them to the Pell-Lucas numbers. First recall that aforementioned
equation[3]:
Qn = 2qn . (31)
So the known properties of the Modified Pell numbers all are easily transferable
to Pell-Lucas numbers.Then we can rewrite the above everything.To do this,
we shall define two new matrices.Define the 2 × 2 matrices R and F as follows:
 
6 2
R= (32)
2 2
3180 A. Dasdemir

and  
2 2
F = . (33)
2 −2
Frankly speaking, these matrices can be obtained with multiplication of
matrices N and E by 2, respectively.
The proof of the next theorem is analogous to the proof of Theorem (2.1),
so it will be omitted.

Theorem 3.1 Let the R be a matrix as in (32). For all positive integers n
and m, the following determinantal equalities are held:
⎧  
⎪
⎪ 3
n Pn+1 Pn
⎪
⎪ 2 2 if n even,
⎪
⎪ Pn Pn−1
⎨
i. Rn =  
⎪
⎪
⎪
⎪ Qn+1 Qn
⎪
⎪ 2
3
(n−1)
if n odd,
⎩ 2
Qn Qn−1

ii. det (Rn ) = 23n ,

iii. per (Rn ) = 23n−2 (Q2n + 2) ,

iv. Rn+m = Rn Rm ,

v. Rn−m = Rn R−m .

Similarly, employing the formula (31) or using the matrices R and F , we

can write as follows:

Qn+1 Qn−1 − Q2n = 8(−1)n+1 , (34)

Qn+1 Qn−1 + Q2n = 2(Q2n + 2(−1)n+1 ), (35)

8Pm+n = Qm Qn+1 + Qm−1 Qn , (36)
Qm+n = Pm Qn+1 + Pm−1 Qn , (37)
2(−1)m+1 Pn−m = Pn Qm−1 − Qn−1 Pm , (38)
(−1)m+1 Qn−m = Pn Qm+1 − Pn+1 Qm , (39)
and
8(−1)m+1 Pn−m = Qn Qm+1 − Qn+1 Qm (40)
It is well-known that the equation (34) was found by Horadam in [2] before-
hand. However, we show solely that they can be obtained by using R-matrix.
Finally, we write relationships between matrices M n and F as follows:
 
n Qn+1 Qn
FM = (41)
Qn Qn−1
Pell, Pell-Lucas and modified Pell numbers 3181

and  
n Qn+1 Qn
8M = F . (42)
Qn Qn−1
ACKNOWLEDGEMENTS. The author would like to thank the refer-
ees for constructive comments, which improved the presentation of the paper.

References
[1] A. F. Horadam, Pell identities, Fibonacci Quart., 9(3) (1971), 245-263.
[2] A. F. Horadam and J. M. Mahon, Pell and Pell-Lucas Polynomials Fi-
bonacci Quart., 23(1) (1985), 7-20.
[3] A. F. Horadam, Applications of Modified Pell Numbers to Representations,
Ulam Quart., 3 (1994), 34-53.
[4] Bro. J. M. Mahon and A. F. Horadam. Matrix and Other Summation
Techniques for Pell Polynomials, Fibonacci Quart. 24(4) (1986), 290-308.
[5] E. Kılıç and D. Taşc.ı, The Linear Algebra of The Pell Matrix, Bol. Soc.
Mat. Mexicana, 3(11) (2005).
[6] R. Melham, Sums Involving Fibonacci and Pell Numbers, Portugaliae
Mathematica, 56(3) (1999), 309-317.
[7] D. Kalman, Generalized Fibonacci numbers by matrix methods, Fibonacci
Quart., 20(1) (1982), 73-76.
[8] J. Ercolano, Matrix generator of Pell sequence, Fibonacci Quart., 17(1)
(1979), 71-77.
[9] N. Bicknell, A primer on the Pell sequence and related sequence, Fibonacci
Quart., 13(4) (1975), 345-349.
[10] H. Kayabaş, The applications of k− Fibonacci sequences, Selçuk Univer-
sity Graduate School of Natural and Applied Science, Ms Thesis, 2006.
[11] H. Minc, Permanents, Encyclopedia of mathematics and its applications,
Addison-Wesley, New York, 1978.
[12] V. E. Hoggat, Fibonacci and Lucas numbers, Palo Alto, CA: Houghton-
Mifflin, 1969.
[13] T. Koshy, Fibonacci and Lucas Numbera with Applications, Wiley- In-
terscience Publications, 2001.
Received: June 30, 2011