JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS

ARTICLE NO.

222, 79]91 1998.

AY975903

Solutions of Linear Difference Equations with

Variable Coefficients
Ranjan K. Mallik*
Department of Electronics and Communication Engineering, Indian Institute of
Technology, Institution of Engineers Building, Panbazar, Guwahati 781001, India
Submitted by Richard A. Duke
Received May 5, 1997

The explicit solution of a linear difference equation of unbounded order with

variable coefficients is presented. As special cases, the solutions of nonhomogeneous and homogeneous linear difference equations of order N with variable
coefficients are obtained. From these solutions, we also get expressions for the
product of companion matrices, and the power of a companion matrix. Q 1998
Academic Press

Key Words: Linear difference equation; variable coefficients; explicit solution;

companion matrices.

1. INTRODUCTION
Linear difference equations or linear recurrences play a significant role
in different areas of science and engineering. Pioneering work on the
asymptotics of linear difference equations was done by Poincare
w1x.
Asymptotics of solutions of linear recurrences with coefficients having
series representations have also been studied by Adams w2x, Birkhoff w3x,
Trjitzinsky w4]6x, Culmer, Harris w7x, Sibuya w8x, Immink w9x, Wimp, and
Zeilberger w10x. Further work on convergence properties of linear recurrence sequences has been presented by Kooman and Tijdeman w11x. A
survey of the literature on explicit solutions of linear recurrences reveals
that in the case of linear recurrences with constant coefficients, the explicit
solutions in terms of coefficients are well known. That is no longer the
case when the coefficients vary with the index w12x. A method of solving
linear matrix difference equations with constant coefficients by using
*E-mail: [email protected].
79
0022-247Xr98 $25.00
Copyright Q 1998 by Academic Press
All rights of reproduction in any form reserved.

80

RANJAN K. MALLIK

operator identities has been presented by Verde-Star w13x. For equations

with variable coefficients, the closed-form solution of the first-order equation is known w14]16x. If N y 1 linearly independent solutions of a
homogeneous equation of order N are known, then any other linearly
independent solution can be obtained explicitly in terms of the known
solutions by using the Casoratian Wronskian in discrete domain.
w14, 15, 17, 18x. For a second-order linear homogeneous difference equation
with variable coefficients, explicit solutions in terms of coefficients have
been presented by Buchberger w19x and in w20x through different approaches. Solutions of linear difference equations of unbounded order and
of order N with variable coefficients have been represented in terms of
determinants of submatrices of a single solution matrix by Kittappa w21x.
Work on the existence and construction of closed-form solutions of linear
recurrences with polynomial and rational coefficients has been done by
Petkovsek
w22, 23x. But, in the available open literature, there are no
expressions in terms of coefficients for the complete solution of a linear
difference equation with varying coefficients when the order is 3 or more,
except for cases in which the coefficients have some special properties.
This paper presents explicit solutions in terms of coefficients of linear
difference equations with variable coefficients, for both the unbounded
order case and the Nth-order case.

2. LINEAR DIFFERENCE EQUATION OF

UNBOUNDED ORDER
Consider the linear difference equation
ky1

yk s

b k , i yi q x k ,

kG1

1.

is1

of unbounded order, with integral index k, variable complex coefficients

bk, 1 , . . . , bk, ky1 , and complex forcing term x k . The solution of this equation, which is an expression for y k , k G 1 in terms of only coefficients and
forcing terms, is given by the proposition that follows.
PROPOSITION 1. The solution of difference equation 1. is gi en by
ky1

yk s

is1

ck , i x i q x k ,

k G 1,

2.

81

LINEAR DIFFERENCE EQUATIONS

where
j

kyi

c k , i s bk , i q

bk , kyl 1

js2

l1 , . . . , l j .
l1 , . . . , l jG1
l1ql 2q ??? ql jskyi

ms2

1
m
bky my
ns 1 l n , ky ns 1 l n

3.

for i s 1, . . . , k y 1, k G 2.
Proof. From the difference equation 1., it is clear that its solution is of
the form 2.. Substituting 2. in 1., we obtain, for i s 1, . . . , k y 1, k G 2,
ky1

yk s

ky1 iy1

bk , i x i q

is2 rs1

ky1

ky2

bk , i x i q

bk , i c i , r x r q x k

rs1 isrq1

ky2

ky1

is1

bk , i c i , r x r q x k

is1

ky1

bk , i q

is1

bk , r c r , i x i q bk , ky1 x ky1 q x k .

4.

rsiq1

Comparing 4. with 2., we get

ky1

c k , i y bk , i s

for i s 1, . . . , k y 1, k G 2.

bk , r c r , i

5.

rsiq1

If we can show that c k, i given by 3. satisfies 5., then the proposition will
be proved.
Now using 3., the right-hand side of 5. can be expressed as
ky1

ky1

bk , r c r , i s bk , iq1 biq1, i q

rsiq1

bk , r br , i

rsiq2
kyiy1 ky1

js2

rsiqj

l1 , . . . , l j .
l 1 , . . . , l jG1
l1ql 2q ??? ql jsryi

bk , r br , ryl 1

ms2

1
m
bry my
.
ns 1 l n , ry ns 1 l n

6.

82

RANJAN K. MALLIK

Substituting l 1 s k y r and replacing the index j by j y 1 in 6., we get

ky1

bk , r c r , i

rsiq1

kyiy2

s bk , ky kyiy1. bky kyiy1., ky kyiy1.y1 q

bk , kyl1 bkyl1 , kyl1y kyl 1yi.

l1s1

kyi kyiy jy1 .

l 1s1

js3

bk , kyl 1

l2 , . . . , l j.
l 2 , . . . , l jG1
l 2q ??? ql jskyiyl 1

1
m
bky my
ns 1 l n , ky ns 1 l n

kyi

ms2

bk , kyl 1

js2

l1 , . . . , l j .
l1 , . . . , l jG1
l 1ql 2q ??? ql jskyi

ms2

1
m
bky my
ns 1 l n , ky ns 1 l n

s c k , i y bk , i

7.

for i s 1, . . . , k y 1, k G 2, from 3.. Therefore we conclude that the

expression for c k, i given by 3. obeys 5., which proves the proposition.
3. LINEAR DIFFERENCE EQUATION OF ORDER N
We consider the linear difference equation
N

y kq N s

a k , j y kqNyj q x kqN ,

kG1

8.

js1

of order N N G 2. with variable complex coefficients a k, j , j s 1, . . . , N,

complex forcing term x kq N , and complex initial values y 1 , . . . , yN . The
solution of this equation, which is an expression for y kq N , k G 1 in terms
of only coefficients, initial values, and forcing terms, is given by the
following proposition.
PROPOSITION 2. The solution of difference equation 8. with initial alues
y 1 , . . . , yN is gi en by
N

y kq N s

js1

d k , j yNq1yj q

js2yk

d k , j x Nq1yj q x kqN ,

k G 1, 9a .

83

LINEAR DIFFERENCE EQUATIONS

where
kqjy1

dk , j s

rs1

l1 , . . . , l r .
1Fl 1 , . . . , l rFN
l rGj
l 1ql 2q ??? ql rskqjy1

ms1

a kql m y mns 1 l n , l m

9b .

for j s 2 y k, . . . , N, k G 1.
Proof. Difference equation 8. with initial values y 1 , . . . , yN can be
treated as a special case of 1. in which
yk s x k

for 1 F k F N,

bk , i s 0

for i s 1, . . . , k y 1, 1 F k F N,

bk , i s 0

for i s 1, . . . , k y N y 1, k G N q 2,

bk , i s a kyN , kyi

10 .

for i s k y N, . . . , k y 1, k G N q 1.

Now the solution of 8. with initial values y 1 , . . . , yN is given by Proposition 1, with the expression for c k, i given by 3. having the additional
constraints on bk, i in 10.. The solution can therefore be expressed by
using 10. as
N

yk s

ky1

c k , i yi q

is1

ck , i x i q x k ,

k G N q 1, or

isNq1

y kq N s

c kqN , Nq1yj yj

11 .

js1
0

c kq N , Nq1yj x Nq1yj q x kqN ,

k G 1.

js2yk

Since 3. can be rewritten as

j

kyi

ck , i s

js1

l1 , . . . , l j .
l 1 , . . . , l jG1
l1ql 2q ??? ql jskyi

ms1

bkql m y mns 1 l n , ky mns 1 l n

12 .

84

RANJAN K. MALLIK

for i s 1, . . . , k y 1, k G 2, it is clear from 10. that for each term of the

summation on the right-hand side of 12. to be nontrivial, the conditions
j

1 F l 1 , . . . , l j F N,

k q lj y

ln s i q l j G N q 1

ns1

for j s 1, . . . , k y i

13 .

need to be satisfied. Therefore we get

j

kyi

ck , i s

js1

l1 , . . . , l j .
1Fl 1 , . . . , l jFN
l jGNq1yi
l1ql 2q ??? ql jskyi

ms1

bkql m y mns 1 l n , ky mns 1 l n

14 .

for i s 1, . . . , k y 1, k G N q 1. Substituting
d k , j s c kqN , Nq1yj

a k , j s bkqN , kqNyj

and

15 .

in 11. and 14. respectively, and combining the two results, we obtain the
proposition.
Note that difference equation 8. with initial values y 1 , . . . , yN can be
expressed in vector form as
y kq N
y kq Ny1
..
.
y kq 1

ak , 1
1
0
s
..
.
0

ak , 2
0
1
..
.

ak , 3
0
0
..
.

???
???
???
???

a k , Ny1
0
0
..
.

ak , N
0
0
..
.

y kq Ny1
x kqN
y kq Ny2
0
=
q
,
..
..
.
.
yk
0

k G 1,

16 .

where the N = N matrix on the right-hand side of 16. is the companion

matrix for index k. Defining
yk J w y kqNy1 , y kqNy2 , . . . , y k x ,

x k J w x kqN , 0, . . . , 0 x , 17a .

ak , 1
1
0
Ak J
..
.
0

ak , 2
0
1
..
.

ak , 3
0
0
..
.

???
???
???
???

a k , Ny1
0
0
..
.

ak , N
0
0
..
.

17b .

85

LINEAR DIFFERENCE EQUATIONS

for k G 1, we can rewrite 16. as

ykq 1 s A k yk q x k ,

k G 1,

18 .

where y1 s w yN , yNy1 , . . . , y 1 xT is the initial value vector. The solution of

18. with initial value vector y1 can then be given in matrix form by
ky1

ykq 1 s A k A ky1 ??? A 1 y1 q

A k A ky1 ??? A lq1 x l q x k ,

k G 1. 19 .

ls1

3.1. Product of Companion Matrices

Consider the homogeneous case of difference equation 8. with initial
values y 1 , . . . , yN , in which x kqN s 0 for k G 1. The solution of this
homogeneous equation can be expressed, using 9a., as
N

y kq N s

d k , j yNq1yj ,

k G 1,

20 .

js1

where d k, j , j s 1, . . . , N are given by 9b..

Extending the definition of d k, j to k s y N y 1., . . . , 0, we can express
the solution 20. as
N

y kq N s

d k , j yNq1yj ,

k G y N y 1. ,

21 .

js1

where d k, 1 , . . . , d k, N are given by 9b. for k G 1, and by

d 0, 1
dy1 , 1
..
.

d 0, 2
dy1, 2
..
.

dy Ny1., 1

dy Ny1., 2

???
???
???

d 0, N
dy1, N
..
.

s IN ,

22 .

dy Ny1., N

where I N is the N = N identity matrix, for k s y N y 1., . . . , 0. However, 19. implies that the matrix form of the homogeneous solution is
ykq 1 s A k A ky1 ??? A 1 y1 ,
where yk and A k are defined in 17..

k G 1,

23 .

86

RANJAN K. MALLIK

Comparing 17., 21., and 23., we find that the product of companion
matrices can be expressed as
A k A ky1 ??? A 1

dk , 1
d ky 1, 1
..
.

dk , 2
d ky1, 2
..
.

d ky Nq1, 1

d kyNq1, 2

???
???
???

d k , Ny1
d ky1, Ny1
..
.

dk , N
d ky1, N
..
.

d kyNq1, Ny1

d kyNq1, N
k G 1,

24 .

where A k k G 1. is defined by 17b., and d kyiq1, j , the entry in the ith row
and the jth column 1 F i, j F N . of the product A k A ky1 ??? A 1 , which is
obtained from 9b. and 22., is given by
kyiqj

d ky iq1, j s

rs1

l1 , . . . , l r .
1Fl 1 , . . . , l rFN
l rGj
l 1ql 2q ??? ql rskyiqj

ms1

a kyiq1ql m y mns 1 l n , l m

if i F min k, N . ,

25a .

s1

if j s i , i ) k , k - N,

25b .

s0

if j / i , i ) k , k - N.

25c .

It can easily be shown that the characteristic equation of A k is given by

det A k y lI N . s y1 .

ln y

a k , j l Nyj ,

js1

26 .

which implies
det A k . s y1 .

Nq 1

ak , N .

27 .

Based on this result, we have the following proposition for linearly independent solutions of the homogeneous version of difference equation 8.,
that is, the equation
N

y kq N s

js1

a k , j y kqNyj ,

k G 1.

28 .

87

LINEAR DIFFERENCE EQUATIONS

PROPOSITION 3. Difference equation 28. with a k, N / 0, k G 1, has N

linearly independent solutions expressed as y kq N s d k, j , j s 1, . . . , N for
k G y N y 1., where d k, 1 , . . . , d k, N are gi en by 9b . and 22..
Proof. If is clear from 21. that y kq N s d k, j , j s 1, . . . , N, are N
solutions of difference equation 28. for k G y N y 1.. The Casoratian of
the N sequences  d k, j 4k Gy Ny1. , j s 1, . . . , N is given by the determinant
dk , 1
d kq 1, 1
..
.
d kq Ny1, 1

s y1 .

???
???

dk , N
d kq2, N
..
.

???

? N2 @

y1.
s~
y1.

d kqNy1, N
d kq Ny1, 1
d kq Ny2, 1
..
.
dk , 1

???
???
???

d kqNy1, N
d kqNy2, N
..
.
dk , N

? N2 @
? N2 @

for k s y N y 1 .
kqNy1

y1. Nq1 ai , N

for k G y N y 2 .

is1

/ 0,

29 .

using 22., 24., 27., and the fact that a i, N / 0, i s 1, . . . , k q N y 1,

k G y N y 2.. Therefore  d k, j 4k Gy Ny1. , j s 1, . . . , N, are linearly independent sequences, which implies that y kq N s d k, j , j s 1, . . . , N, are N
linearly independent solutions of the difference equation for k G y
N y 1..
3.1.1. Power of the Companion Matrix for the Constant Coefficient Case
Consider the case of the homogeneous Nth-order difference equation
28. with initial values y 1 , . . . , yN in which a k, j s a j for all k G 1, j s
1, . . . , N, that is, the equation
N

y kq N s

a j y kqNyj ,

kG1

30 .

js1

with initial values y 1 , . . . , yN . We can rewrite 30. as

ykq 1 s Ayk ,

k G 1,

31 .

88

RANJAN K. MALLIK

where yk is defined in 17a., A is defined in 17b. with a k, j replaced by a j ,

and y1 is the initial value vector. From 23., the matrix form of the solution
to 31. with initial value vector y1 is given by
ykq 1 s Ak y1 ,

k G 1.

32 .

Thus the kth power of the N = N companion matrix A gives the solutions
for y kq 1 , . . . , y kqN of the Nth-order homogeneous linear difference equation 30. with constant coefficients a1 , . . . , a N and initial values y 1 , . . . , yN .
Using the result in 24. and 25. for the product of companion matrices,
we obtain

Ak s

???
???

dk , 1
d ky 1, 1
..
.

dk , 2
d ky1, 2
..
.

d ky Nq1, 1

d kyNq1, 2

???

d k , Ny1
d ky1, Ny1
..
.

dk , N
d ky1, N
..
.

d kyNq1, Ny1

d kyNq1, N

k G 1,

33 .

where d ky iq1, j , the entry in the ith row and the jth column of Ak , is
expressed as
d ky iq1, j
kyiqj

rs1

l1 , . . . , l r .
1Fl 1 , . . . , l rFN
l rGj
l1ql 2q ??? ql rskyiqj

ms1

al m

if i F min k, N . , 34a .

s1

if j s i , i ) k, k - N,

34b .

s0

if j / i , i ) k, k - N.

34c .

Our aim is to obtain an alternative expression for d ky iq1, j when i F

min k, N . in terms of powers of the coefficients a1 , . . . , a N .
Now 34a. can be rewritten as

d ky iq1 , j

s ~
N

sj

kyiqj
rs1

l 1 , . . . , l ry1 .
1Fl 1 , . . . , l ry1 FN
l 1ql 2q ??? ql ry1 skyiqjy

al1 ??? al ry 1

.a

35 .

for i F min k, N .. The sum of a l 1 ??? a l ry 1 over all l 1 , . . . , l ry1 satisfying

1 F l 1 , . . . , l ry1 F N,

l 1 q l 2 q ??? ql ry1 s k y i q j y

89

LINEAR DIFFERENCE EQUATIONS

where r s 1, . . . , k y i q j is the same as the sum of

t 1 q t 2 q ??? qt N t
a11 ??? a Nt N
t1 , . . . , t N

over all t 1 , . . . , t N satisfying

t 1 , . . . , t N G 0,

t 1 q 2 t 2 q ??? qNt N s k y i q j y .

Therefore 35. implies, for i F min k, N .,

d ky iq1, j

s ~

sj

t1 , . . . , t N .
t1 , . . . , t N G0
t1q2 t 2q ??? qNt N skyiqjy

t 1 q t 2 q ??? qt N
t1 , . . . , t N

ms1

tm
am

a .

36 .
Replacing t by t y 1 in 36., we get
d ky iq1, j s

t j q t jq1 q ??? qt N

t 1 q t 2 q ??? qt N

t1 , . . . , t N .
t1 , . . . , t N G0
t1q2 t 2q ??? qNt N skyiqj

t 1 q t 2 q ??? qt N
t1 , . . . , t N

ms1

tm
am

37 .

for i s 1, . . . , min k, N ., j s 1, . . . , N, k G 1. This is an alternative way of

expressing 34a. in terms of the coefficients a1 , . . . , a N of the companion
matrix A. The result is consistent with the expression for the combinatorial
power of the companion matrix presented by Chen and Louck w24, Theorem 3.1x, obtained by using the concept of a digraph to represent a matrix.

4. CONCLUSION
The explicit solutions of the linear difference equations presented here
utilize the combinatorial properties of the indices of the coefficients. The

90

RANJAN K. MALLIK

solution of the difference equation of unbounded order results in the

solution of the Nth-order equation, which, in turn, provides expressions
for the product of companion matrices and the positive integral powers of
a companion matrix.

ACKNOWLEDGMENTS
The author thanks his colleague Dr. S. Ponnusamy of the Department of Mathematics,
Indian Institute of Technology, Guwahati, for his valuable suggestions. He also expresses his
gratitude to the anonymous referee for hisrher constructive comments.

REFERENCES
1. H. Poincare,
aux differentielles
ordinaires et aux differences
Sur les equations lineaires

finies, Amer. J. Math. 7 1885., 203]258.

2. C. R. Adams, On the irregular cases of the linear ordinary difference equation, Trans.
Amer. Math. Soc. 30 1928., 507]541.
3. G. D. Birkhoff, Formal theory of irregular linear difference equations, Acta Math. 54
1930., 205]246.
4. G. D. Birkhoff and W. J. Trjitzinsky, Analytic theory of singular difference equations,
Acta Math. 60 1932., 1]89.
5. W. J. Trjitzinsky, Analytic theory of linear q-difference equations, Acta Math. 61 1933.,
1]38.
6. W. J. Trjitzinsky, Laplace integrals and factorial series in the theory of linear differential
and linear difference equations, Trans. Amer. Math. Soc. 37 1935., 80]146.
7. W. J. A. Culmer and W. A. Harris, Jr., Convergent solutions of ordinary linear homogeneous difference equations, Pacific J. Math. 13 1963., 1111]1138.
8. W. A. Harris, Jr., and Y. Sibuya, Note on linear difference equations, Bull. Amer. Math.
Soc. 70 1964., 123]127.
9. G. K. Immink, Asymptotics of analytic difference equations, in Lecture Notes in
Mathematics, Vol. 1085, Springer-Verlag, BerlinrNew York, 1984.
10. J. Wimp and D. Zeilberger, Resurrecting the asymptotics of linear recurrences, J. Math.
Anal. Appl. 111 1985., 162]176.
11. R. J. Kooman and R. Tijdeman, Convergence properties of linear recurrence sequences,
Nieuw Arch. Wisk. 4 . 4 1990., 13]25.
12. A. M. Odlyzko, Asymptotic Enumeration Methods, Handbook of Combinatorics, R. L.
Graham, M. Grotschel,
and L. Lovasz,

Eds.., Chap. 22, pp. 1130]1146, Elsevier, Amsterdam, 1995.

13. L. Verde-Star, Operator identities and the solution of linear matrix difference and
differential equations, Stud. Appl. Math. 91 1994., 153]177.
14. K. S. Miller, Linear Difference Equations, W. A. Benjamin, New YorkrAmsterdam,
1968.
15. C. M. Bender and S. A. Orszag, Advanced Mathematical Methods for Scientists and
Engineers, McGraw-Hill, New York, 1978.

LINEAR DIFFERENCE EQUATIONS

91

16. W. G. Kelley and A. C. Peterson, Difference Equations: An Introduction with Applications, Academic Press, San Diego, 1991.
17. S. J. Farlow, Difference Equations, An Introduction to Differential Equations and Their
Applications, Chap. 7, McGraw-Hill, New York, 1994.
18. S. L. Ross, Explicit Methods of Solving Higher-Order Linear Differential Equations,
Differential Equations, Chap. 4, 3rd Ed., Wiley, New York, 1984.
19. B. Buchberger, O velichinakh Si , opredelennykh rekursie Siq1 s u i Si q i Siy1 , Tech.
Rep. r5-5789, Obedinennogo Instituta Yadernykh Issledovani,
Dubna, May 19, 1971 in
Russian..
20. R. K. Mallik, On the solution of a second order linear homogeneous difference equation
with variable coefficients, J. Math. Anal. Appl. 215 1997., 32]47.
21. R. K. Kittappa, A representation of the solution of the nth order linear difference
equation with variable coefficients, Linear Algebra Appl. 193 1993., 211]222.
22. M. Petkovsek,
Hypergeometric solutions of linear recurrences with polynomial coefficients, J. Symbolic Comput. 14 1992., 243]264.
23. M. Petkovsek,
A generalization of Gospers algorithm, Discrete Math. 134 1994., 125]131.
24. W. Y. C. Chen and J. D. Louck, The combinatorial power of the companion matrix,
Linear Algebra Appl. 232 1996., 261]278.