See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/319846030
Finite Difference Equations
Chapter · January 2005
DOI: 10.1007/0-387-27645-9_3
CITATIONS READS
0 48
3 authors, including:
Robby Robson
Eduworks
84 PUBLICATIONS 636 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Intelligent Tutoring Systems View project
Knowledge Management View project
All content following this page was uploaded by Robby Robson on 06 March 2018.
The user has requested enhancement of the downloaded file.
3
Finite Difference Equations
3.1 Linear Difference Equations
A difference equation is the discrete analog of a differential equation. Al-
though differential equations are typically studied earlier in a mathematical
curriculum, there are many respects in which the theory of difference equa-
tions is simpler. A finite difference equation has the general form
(3.1) sn = Φ(sn−1 , sn−2 , . . . , sn−k , n),
where Φ is a fixed complex-valued function and the integer k is called the
order of the equation. The initial value problem for (3.1) is the problem
of finding a sequence sn that satisfies (3.1) for a given function Φ and a
fixed initial value vector S0 = (sk−1 , . . . , s1 , s0 )T . An initial value problem
has only one solution, because (as in Chapter 2) for any fixed k initial values
s0 , . . . , sk−1 , the k th term is specified by sk = Φ(sk−1 , sk−2 , . . . , s0 , k), and
all successive values are similarly found. From this we obtain the following
result.
Theorem 3.1.1 (Existence and uniqueness theorem). Every initial
value problem for a finite difference equation has a unique solution.
The essential ingredient in the proof of this result is the assumption that
Φ is a function, that is, that Φ returns exactly one value for a given input.
The function Φ is called a linear function in the variables sn−1 , . . . , sn−k
if there exist k + 1 complex-valued functions g1 , . . . , gk and ψ defined on
34 3. Finite Difference Equations
the natural numbers such that for all n ≥ k,
Φ(sn−1 , sn−2 , . . . , sn−k , n) = g1 (n)sn−1 + · · · + gk (n)sn−k + ψ(n) ,
where ψ is called the input or forcing function. When Φ is a linear
function, then (3.1) is called a linear difference equation. For the special
case in which the coefficients g1 , . . . , gk are constants, (3.1) becomes
sn = c1 sn−1 + · · · + ck sn−k + ψ(n) for c1 , . . . , ck ∈ C ,
and it is called a constant coefficient equation.
In this chapter we study linear constant coefficient difference equations,
and we will simply refer to them as difference equations or recurrences.
Every such recurrence can be written in the form
(L) sn − c1 sn−1 − c2 sn−2 − · · · − ck sn−k = ψ(n) for n ≥ k ,
for some complex-valued function ψ and constants c1 , . . . , ck , where ck is
non-zero. The homogeneous linear constant coefficient equation (HL) stud-
ied in Chapter 2 is the special case in which the input function ψ(x) is the
zero function.
3.1.1 First–order equations
A first–order recurrence has the form
(L1) sn = λsn−1 + ψ(n) for all n ≥ 1 ,
where the function ψ is defined for all positive natural numbers and λ = 0.
When we consider the associated initial value problem for fixed s0 = α0 ,
we have
s0 = α0 ;
s1 = λs0 + ψ(1) = λα0 + ψ(1) ;
s2 = λs1 + ψ(2) = λ2 α0 + λψ(1) + ψ(2) ;
s3 = λs2 + ψ(3) = λ3 α0 + λ2 ψ(1) + λψ(2) + ψ(3);
..
.
and this gives the following form for the solution.
Theorem 3.1.2. The initial value problem
(3.2) sn = λsn−1 + ψ(n), s0 = α0 ,
always has the unique solution sn with general term
n
n
(3.3) sn = α0 λ + ψ(i)λn−i
i=1
(where the sum is defined to be zero when n = 0).
View publication stats
