CSM 166 Chapter 4 - Lecture - Slides

CSM 166: Discrete Mathematics
for Computer Science

FINITE DIFFERENCE EQUATIONS - RECURRENCE RELATIONS
Isaac Afari Addo <addoisaacafari@gmail.com>

National Institute for Mathematical Science (NIMS) - Ghana
Department of Mathematics, KNUST
Kumasi-Ghana.
09/07/2021
Content
Introduction
Classification of Recurrence Relations
CSM 166: Discrete Mathematics for Computer Science Chapter 4: Difference Equation-Recurrence Relation, 09/07/2021
Introduction
Definition 1 (Recurrence relation)

A recurrence relation for a sequence a0 , a1 , , . . . is a
relation that defines an in terms of a0 , a1 , , . . . an−1 .
The formula relating an to earlier values in the

sequence is called the generating rule.
The assignment of a value to one of the a’s is

called an initial condition.
Introduction
Example 1
The Fibonacci Sequence 1, 1, 2, 3, 5, . . . is a
sequence in which every number after the first
two is the sum of the preceding two numbers.
The initial conditions are a0 = a1 = 1 and the

generating rule is
an = an−1 + an−2 ; n≥2
Solution to a Recurrence Relation
A solution to a recurrence relation is an explicit
formula for an in terms of n .
The fundamental method for finding the

solution of a sequence defined recursively is by
using iteration.
This involves starting with the initial values of

the sequence and then calculates successive
terms of the sequence until a pattern is observed.
An explicit formula for the sequence and then

uses mathematical induction to prove its validity.
Example 2
Find a solution for the recurrence relation
(
a0 = 1
an = an−1 + 2, n≥1
Solution to Example 2
Listing out some terms of the sequence:
a0 = 1
a1 = 1 + 2
a2 = 1 + 4
a3 = 1 + 6
a4 = 1 + 8
a1 = 1 + 10
A guessed formular is an = 2n + 1 , n ≥ 0 and thus

needs to be proven using mathematical
induction.
Example 3
Consider the arithmetic sequence
an = an−1 + d, n≥1
where a0 is the initial value. Find an explicit

formula for an .
Solution to Example 3
Listing out the first 4 terms of the sequence:
a1 = a0 + d
a2 = a0 + 2d
a3 = a0 + 3d
a4 = a0 + 4d
a5 = a0 + 5d
A guessed formular is an = a0 + nd and thus
needs to be proven using mathematical

induction.
Exercise A:
1. Consider the geometric sequence
an = ran−1 , n≥1
Where a0 is the initial value. Find an explicit

formula for an .
2. Find a solution to the recurrence relation
(
a0 = 0
an = an−1 + (n + 1), n≥1
Exercise B: A function yn is defined recursively
as follows:

y1 = 3

y2 = 7

yn = 3yn−1 − 2yn−2 forn ≥ 3

Find an explicit formula or solution for yn in

terms of n.
Classification of Recurrence
Relations
A recurrence relation is of first order if yn is
defined only in terms of yn−1 .
It is of second order if yn is defined in terms of

yn−1 and yn−2 , and so on.
A recurrence relation of the form

yn = a1 yn−1 + a2 yn−2 + · · · + ak yn−k is called a linear
homogenous recurrence relation of order k
If the ai are constants, then the above equation is

said to have constant coefficients.
Classification of Recurrence
Relations
Linear recurrence relations have the following

important properties:
1. multiplying any solution by a constant gives

another solution,
2. adding two or more solutions give another

solution.
First-Order Recurrence Relations
First-order recurrence relations are of the form:

(
yn = ayn−1
y0 = c
where a and c are constants.
First-order recurrence relations are solved by
iteration:
yn = ayn−1
= a(ayn−2 )
= a2 (ayn−3 )
= ......
= an−1 y1
= a n y0
Using the initial condition, we have yn = can ,

n ∈ Z+
Second-Order Recurrence
Relations
Second-order recurrence relations are of the
form:

y = ayn−1 + byn−2 for n ≥ 2
 n


y 0 = c1 (1)


y 0 = c0

Assuming a, b, c0 and c1 are constants and a trial

function yn = ct n to solve the relation above.
Using this assumption, yn−1 = ct n−1 and

yn−2 = ct n−2
Second-Order Recurrence
Relations
Subtituting these into the recurrence relation (2:
ct n = act n−1 + bct n−2
Dividing through by ct n−2 :
t 2 = at + b ⇒ t 2 − at − b = 0 (2)
(2) is called the auxiliary or characteristic
equation of the recurrence relation.
Roots of the Characteristic
Equation
The characteristic equation is a quadratic

equation whose roots may be:
I two distinct real roots t = t1 and t = t2
II repeated real roots t = t0
III two complex roots t = t1 and t2 = t1
Solution to the Recurrence
Relation
CASE I
Since yn = t1n and yn = t2n are solutions of the
linear recurrence relation then another solution
(the general solution) is
yn = At1n + Bt2n (3)
where A and B are arbitrary constants.
A and B are determined using the initial values

y0 = c0 and y1 = c1 .
Relation
CASE II
Since t = t0 is the repeated root of the
characteristic equation, yn = t0n is a solution of
the recurrence relation as well as the linearly
independent solution yn = nt0n .
Thus a general solution:

yn = At n + Bnt n (4)
where A and B are arbitrary constants.
A and B are determined using the initial values

y0 = c0 and y1 = c1 .
Relation I
CASE III
The complex roots of the auxiliary equation with
real coefficients occur in conjugate pair. i.e if
t1 = u + iv then t2 = u − iv with v 6= 0
By the general rule, the solution
yn = At1n + Bt2n
= A(u + iv)n + B(u − iv)n (5)
Relation II
In polar form
u + iv = r(cos θ + i sin θ)
u + iv = r(cos θ − i sin θ)
And by DeMoivre’s Theorem
[ρ(cos θ ± i sin θ]n = ρ n (cos nθ ± i sin nθ)
Thus
yn = Aρ n (cos nθ + i sin nθ) + Bρ n (cos nθ − i sin nθ)

= (A + B)ρ 2 (cos nθ) + i(A − B)ρ 2 sin nθ
Relation III
If we substitute A = B = 12 , then yn = ρ n cos nθ is a

particular solution.
Similarly taking A = − 12 i and B = 12 i, then

yn = ρ n sin nθ is also a particular solution.
Thus the general solution is
yn = Aρ n sin nθ + Bρ n cos nθ (6)

p
where ρ = u2 + v2 and θ = tan−1 21 .
Example 4
Solve the following

yn = 3yn−1 − 2yn−2 for n ≥ 2

iy2 = 7

y1 = 3


yn = 6yn−1 − 9yn−2 for n ≥ 1

ii y1 = 3

y0 = 5

iii yn + 2yn−1 + 2yn−2 = 0

Solution:
The Characteristic equation is t 2 − 6t + 2 = 0
Exercise C: Solve the following difference
equations:
1. yn+1 − ayn = 0, y0 = 1
2. yn = −3yn−1
3. yn+2 + 2yn = 0
4. yn − 2yn−1 + 2yn−2 = 0
5. yn + 4yn−1 + 8yn−2 = 0; y1 = −1
6. yn − 4yn−1 + 8yn−2 = 0; y2 = 1, y3 = −2
7. yn+2 + 2yn+1 + 4yn = 0
End of Lecture
Questions...???
Thanks

CSM 166 Chapter 4 - Lecture - Slides

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CSM 166 Chapter 4 - Lecture - Slides

Uploaded by

Copyright:

Available Formats

CSM 166: Discrete Mathematics

for Computer Science

Isaac Afari Addo <addoisaacafari@gmail.com>

Classification of Recurrence Relations

Definition 1 (Recurrence relation)

The formula relating an to earlier values in the

The assignment of a value to one of the a’s is

The initial conditions are a0 = a1 = 1 and the

an = an−1 + an−2 ; n≥2

The fundamental method for finding the

This involves starting with the initial values of

An explicit formula for the sequence and then

A guessed formular is an = 2n + 1 , n ≥ 0 and thus

where a0 is the initial value. Find an explicit

A guessed formular is an = a0 + nd and thus

needs to be proven using mathematical

Where a0 is the initial value. Find an explicit

Find an explicit formula or solution for yn in

It is of second order if yn is defined in terms of

A recurrence relation of the form

If the ai are constants, then the above equation is

Linear recurrence relations have the following

1. multiplying any solution by a constant gives

2. adding two or more solutions give another

First-order recurrence relations are of the form:

where a and c are constants.

Using the initial condition, we have yn = can ,

Assuming a, b, c0 and c1 are constants and a trial

Using this assumption, yn−1 = ct n−1 and

Subtituting these into the recurrence relation (2:

ct n = act n−1 + bct n−2

Dividing through by ct n−2 :

equation of the recurrence relation.

The characteristic equation is a quadratic

I two distinct real roots t = t1 and t = t2

II repeated real roots t = t0

III two complex roots t = t1 and t2 = t1

yn = At1n + Bt2n (3)

where A and B are arbitrary constants.

A and B are determined using the initial values

Thus a general solution:

A and B are determined using the initial values

By the general rule, the solution

And by DeMoivre’s Theorem

[ρ(cos θ ± i sin θ]n = ρ n (cos nθ ± i sin nθ)

yn = Aρ n (cos nθ + i sin nθ) + Bρ n (cos nθ − i sin nθ)

If we substitute A = B = 12 , then yn = ρ n cos nθ is a

Similarly taking A = − 12 i and B = 12 i, then

yn = Aρ n sin nθ + Bρ n cos nθ (6)

iii yn + 2yn−1 + 2yn−2 = 0

You might also like