VOL. 5, NO.
7, JULY 2010
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
2006-2010 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
SOLVING DIFFERENCE EQUATIONS BY FORWARD DIFFERENCE
OPERATOR METHOD
1
Odior A. O.1, Charles-Owaba O. E.2 and Fadare D. A.3
Department of Production Engineering, University of Benin, Nigeria
Department of Industrial and Production Engineering, University of Ibadan, Nigeria
3
Department of Mechanical Engineering, University of Ibadan, Nigeria
E-Mail:
[email protected]ABSTRACT
In this paper a forward difference operator method was used to solve a set of difference equations. We also find
the particular solution of the nonhomogeneous difference equations with constant coefficients. In this case, a new operator
call the forward difference operator r,s, defined as r,s yn = r yn+1 - s yn, was introduced. Some of the properties of this
new operator were also investigated.
Keywords: forward difference, particular solution, nonhomogeneous, difference equations, constant coefficients.
INTRODUCTION
In Numerical Analysis, we use some linear
operators such as shift exponential operator E, with Efj =
fj+1, forward difference operator , with fj = fj+1 fj,
and backward difference , with fj = fj fj1. These
operators are used in some aspects of Numerical Analysis,
particularly in interpolation, quadratures, difference
equations, and so forth. (Odior, 2003; Lambert, 1973;
Phillips et al. 1980).
Under the forward difference operator , the
linear difference equations are written in one of the
following forms
P ()yn = 0, (homogeneous)
(1)
P ()yn = fn (nonhomogeneous)
(2)
Where P is a polynomial.
Prototype results towards the development of the
basic theory of the global behaviour of solutions of
nonlinear difference equations of order greater than one.
The techniques and results are also extremely useful in
analyzing the equations in the mathematical models of
various biological systems and other applications (Kalabu
and Kulenovi, 2003; senior, et al. 2001).
THEOREM
To solve linear difference equations for general
solution, we introduce a set of theorems which are used for
the general solution, (Hosseinzadeh and Afrouzi, 2007).
Theorem 3
Let yh be a solution for (1) and yp be a particular
solution for (2), then yc = yh + yp is a solution for (2) too.
Finite differences
Given a special function of f(x), with an
argument x proceeding at equal intervals, we could
generate a Table of logarithms or trigonometric functions
or a Table of special computation for certain calculation
(Odior, 2003). Given a certain function of x, say, with
values f (a), f (a+b), f (a+2h),.., the difference between
the consecutive values of x, denoted by h is called the
interval differencing.
If f(x1) = f(x)
f (x2) = f(x + h) and
f (x2) - f(x1) = f(x)
then f(x) = f(x + h) - f(x)
(3)
Where is called difference operator and f(x) is
called the first difference of f(x). The second difference of
f(x) is given thus:
[ f (x)] = 2f(x) = f(x + h) - f(x)
(4)
Building a difference Table
Build up a difference Table for the function, f (x)
= x3 + 3x2 - 2x + 5, taking x = 0 as the initial value, and h
= 2.
Solution: We now build up the Table as follows:
Theorem 1
This is the principle of superposition. Suppose
that y1, y2, ..., ym are the solutions of the homogeneous
difference Equation (1), then any linear combinations of
them is a solution for same equation too.
Theorem 2
Suppose that the complex-valued function, yn =
y1 + iy2 be a solution of equation (1), then functions y1, y2
are also solutions for the equation.
89
VOL. 5, NO. 7, JULY 2010
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
2006-2010 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
p
p
p
f p = E p f 0 = f 0 + f 0 + 2 f 0 + 3 f 0 + .....
I
2
3
This is obtained from,
f x + nh = E n f x = (1 + ) f x
n
n
n
n
= f x + f x + 2 f x + 3 f x + .....
3
2
1
(5)
Direction of errors in a Table of differences
The effect of a single error in a Table of
differences from the point of error becomes considerably
magnified. This is illustrated in the difference Table
bellow, in which error is made in the value f4.
The Operators E and
The Newton-Gregory forward difference formula for fp is
given as:
So, it is clear that the error made in the value f4 will continue to affect some of the other values
generated until it is pronounced in all the values.
Difference equation
A difference equation is an equation involving
the differences between successive values of a function of
an integer variable. It can be regarded as the discrete
version of a differential equation. For example the
difference equation f (n+1) - f (n) = g (n) is the discrete
version of the differential equation f' (x) = g (x). We can
see difference equation from at least three points of views:
as sequence of number, discrete dynamical system and
iterated function. It is the same thing but we look at
different angle. Difference equation is a sequence of
numbers that are generated recursively using a rule to
relate each number in the sequence to previous numbers in
the sequence as presented in Figure-1, {1, 1, 2, 3, 5, 8, 13,
21,}. The Sequence {1, 1, 2, 3, 5, 8, 13, 21,.} is called
Fibonacci sequence, generated with rule y (k+2) = y (k+1)
+ y (k) for k = 0, 1, 2, 3,, and initial value y(0) = y0 = 1.
90
VOL. 5, NO. 7, JULY 2010
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
2006-2010 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Figure-1. Sequence of difference equation.
Sequence {3, 5, 7, 9, .} has rule y(k+1) = 2y(k) + 3 for
k = 0, 1, 2, 3,,.
Both sequences have initial value y (0) = y0 = 0.
MATERIALS AND METHODS
We now consider the homogeneous and
nonhomogeneous difference equations with constant
coefficients with application of the approach considered
by Hassan and Afrouzi1 (2008).
Let yn = rn be a solution for equation (1), we have
P(r - 1) = 0.
(6)
Where Equation (3) is the corresponding characteristic
equation to Equation (1).
Remarks:
The following remarks can be drawn from the
forgoing as presented below:
Remark 1: All roots of the characteristic equations may
be distinct real values, either some of them equal or some
of them are conjugate complex number.
(i) If r1, r2, ., rk be the distinct real roots to the
n
rkn will be solutions of the homogeneous equations.
Then the above functions are the fundamental solutions of
the homogeneous equation.
(ii) If r1 = r2 = ... = rm = r be the repeated roots of the
characteristic equation (3), then the fundamental solutions
of the homogeneous equation are: rn, nrn, n2rn, ., nm-1rn
and they are linearly independent, (Lambert, 1973).
(iii) If r1,2 = i be two conjugate complex roots, the
fundamental solutions of the homogeneous equation are,
y1 = (2 + 2 )
cosn, y2 = (2 + 2 )
sinn,
Where = tan- (/).
1
Fundamental solution
(1): Find the fundamental solutions of the following
homogeneous equation
(4 +2)yn = 0.
n
n
and y4 = 2 2 sin
4
4
n
y1 = 1, y2 = n, y3 = 2 2 cos
(2): Solve the following difference equation and find the
fundamental solutions
(5 + 6)(322 + 56 + 25)yn = 0
Solution: The roots of the corresponding characteristic
equation are r1 =
r2, 3 =
1
,
5
1
(1 i ) . This gives the solutions of the equation to
8
be:
n
n
1 n
) , y2 = 2 2 cos
and y3 = 2 2 sin
4
4
5
5n
y1 = (
characteristic equations, then the functions r1 , r2 ,.,
Solution: The characteristic equation is (r - 1)2(r2 - 2r + 2)
= 0. This polynomial equation has one double root; (r = 1)
and two complex conjugate
roots (r = 1 i), therefore the fundamental solutions
become:
5n
(3): Find the fundamental solutions of the following
homogeneous equation
(4 +2)yn = 0.
Solution: The characteristic equation is (r - 1)2 (r2 - 2r +
2) = 0. This polynomial equation has one real double root
r = 1 and two complex conjugate
roots r = 1 i, therefore the fundamental solutions may
be written as follow:
n
n
y1 = 1, y2 = n, y3 = 2 cos
and y4 = 2 2 sin
.
4
4
n
2
We need to prove the accuracy of the following equalities
in our operations.
n 1
(i)
f
j =0
= fn
91
VOL. 5, NO. 7, JULY 2010
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
2006-2010 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
Each of the above identities is used for finding
particular solution of non-homogeneous difference
equations with constant coefficients.
n 1
1
fn = f j
j =0
(ii)
n 1
For
the
proof
of
j =0
n 1
n 1
j=0
j =0
= fn ,
we
consider
Solution:
n 1
f j = fn .
yn = cosn and so yn =
j =0
For
the
n 1
f
j =0
proof
j=0
f
j =0
n 1
1
fn = f j ,
j =0
f f
n 1
and
of
n 1
j =0
we
consider
= fn ,
Let yp = 1 (cos n ) =
n 1
1
(cos n )
cos j = 2 sin(
j =0
2n 1
1
) cos
2
2 2
(5): Find the particular solution of the following difference
equation
= fn
2 xm = cos(
n 1
1
fn = f j .
j =0
xm =
(4): Find the particular solution of the following difference
equation
yn = cosn
f j f j = fn .
fj =
j =0
m +1
)
3
Solution:
The solution is obtained thus:
1
m +1
1 m1
( j + 1) m1 1
(2 j + 1)
1
m
cos(
) = cos
= ( + sin
) = 1 m cos
3
j =0
3
2
6
2
3
j =0
1
1
1
1
f n L )) s
(
(L
rm , sm rm 1, sm 1
r1, s1
RESULTS AND DISCUSSIONS
We define the forward difference operator r,s as
follow
yp=
r,syn = ryn+1 - syn = (rE - s)yn
To find the particular solution of a given
nonhomogeneous difference equation such as:
Where yn is the approximate value function y(x) at point
xn [ x0 , xm ] , then two operators (r,s) and (rE - s) are
equivalent.
Difference Operations in Vector Space of
Operator r,s. Some basic difference operations in the
vector space of difference operator are defined as:
(i) r1,s +r2,s r1 + r2,s
(ii) r,s1 + r,s2 r,s1 + s2
(iii) r1,s - r2,s r1 - r2,s
(iv) r,s1 - r,s2 r,s1 - s2
(v) r1,s1 r2 ,s2 r2,s2 r1,
(vi)
r1,s1
r 2, s 2
r1,s1 (
1
r 2, s 2
)(
1
r 2, s 2
) r1, s1
Solution of nonhomogeneous difference equation with
constant coefficients
A general form of nonhomogeneous difference
equation with order m can be written in the form
P (E)yn =
f n , P( )yn = f n ,
P( ) =
fn
m
j =1
rj , sj
fn =
2, 1 2,3yn = 2nn2,
We have: Let yp =
=
1
(22 n n )
2,1 2,3
1
1
1
1
32
368
(
(2n n 2 )) =
(2n (n 2 8n + 28)) = 2n ( n 2 n +
)
2,1 2,3
2,1
3
9
27
CONCLUSIONS
We have been able to establish the first
difference, second difference, third difference, and forth
difference of a given polynomial function. The shift
operator E method for solving the non-homogeneous
difference equations with constant coefficients is a new
method which can be used to solve all types of
nonhomogeneous difference equations with constant
coefficients.
(7)
This can as well be written as:
92
VOL. 5, NO. 7, JULY 2010
ISSN 1819-6608
ARPN Journal of Engineering and Applied Sciences
2006-2010 Asian Research Publishing Network (ARPN). All rights reserved.
www.arpnjournals.com
REFERENCES
Hassan H., Afrouzi1 G. A. 2008. On forward (r, s)difference operator (r, s) and solving difference equations.
J. Applied Math. Sciences. 2(60): 3005-3013.
Hosseinzadeh H., Afrouzi G. A. 2007. On forward rdifference operator and solution of nonhomogeneous
difference equations. IMF J. 2(40): 1957-1968.
Kalabu S., Kulenovi M. R. S. 2003. On rate of
convergence of solutions of rational difference equation of
second order. Dept of Math, University of Rhode Island,
Kingston, RI 02881-0816.
Lambert J. D. 1973. On computation methods in ordinary
differential equations. John Wiley and Sons, New York.
Odior A. O. 2003. Mathematics for science and
engineering students. (1) Third Edition. Ambik Press,
Benin City, ISBN 978-027-289-5.
Phillips G. M., Taylor P. J. 1980. On theory and
applications of numerical analysis. 5th Edition. Academic
Press.
Senior T. B. A., Legault S., Volakis J. L. 2001. A novel
technique for the solution of second-order difference
equations. Antennas and Propagation. IEEE Trans. 49(12):
1612-1617.
93
