Chapter 4: Advanced counting methods

Introduction
Definition 1: If 𝑎 and 𝑏 are integers with 𝑎 ≠ 0, we say that 𝑎 divides 𝑏 if there is an integer 𝑐 such
𝑏
that 𝑏 = 𝑎𝑐 (or equivalently, if 𝑎 is an integer). When 𝑎 divides 𝑏 we say that 𝑎 is a factor or divisor of

𝑏, and that 𝑏 is a multiple of 𝑎. The notation 𝑎 ∣ 𝑏 denotes that 𝑎 divides 𝑏. We write 𝑎 | 𝑏 when 𝑎 does
not divide 𝑏.
Note: Let 𝑛 and 𝑑 be positive integers. How many positive integers not exceeding 𝑛 are divisible by 𝑑?
The positive integers divisible by 𝑑 are all the integers of the form 𝑑𝑘, where 𝑘 is a positive integer.
Hence, the number of positive integers divisible by 𝑑 that do not exceed 𝑛 equals the number of integers
𝑛 𝑛
𝑘 with 0 < 𝑑𝑘 ≤ 𝑛 or with 0 < 𝑘 ≤ 𝑑 . Therefore, there are ⌊𝑑⌋ positive integers not exceeding 𝑛 that

are divisible by 𝑑.

4.1 The principle of Inclusion – Exclusion

Let 𝐴, 𝐵, 𝐶 and 𝐷 be sets. Then
1. |𝐴 ∪ 𝐵| = |𝐴| + |𝐵| − |𝐴 ∩ 𝐵|
2. |𝐴 ∪ 𝐵 ∪ 𝐶| = |𝐴| + |𝐵| + |𝐶| − |𝐴 ∩ 𝐵| − |𝐴 ∩ 𝐶| − |𝐵 ∩ 𝐶| + |𝐴 ∩ 𝐵 ∩ 𝐶|
3. |𝐴 ∪ 𝐵 ∪ 𝐶 ∪| = |𝐴| + |𝐵| + |𝐶| + |𝐷|
− |𝐴 ∩ 𝐵| − |𝐴 ∩ 𝐶| − |𝐴 ∩ 𝐷| − |𝐵 ∩ 𝐶| − |𝐵 ∩ 𝐷|
− |𝐶 ∩ 𝐷| + |𝐴 ∩ 𝐵 ∩ 𝐶| + |𝐴 ∩ 𝐵 ∩ 𝐷| + |𝐵 ∩ 𝐶 ∩ 𝐷|
+|𝐴 ∩ 𝐶 ∩ 𝐷| − |𝐴 ∩ 𝐵 ∩ 𝐶 ∩ 𝐷|.
Where |𝑋| is cardinality of 𝑋 or the number of elements of set 𝑋.
Theorem 1: (Inclusion – Exclusion)
Let 𝐴1 , 𝐴2 , − − −, 𝐴𝑛 be finite sets. Then

|𝐴1 ∪ 𝐴2 ∪ 𝐴3 ∪ − − − 𝐴𝑛 | = ∑ |𝐴𝑖 | − ∑ |𝐴𝑖 ∩ 𝐴𝑗 |

1≤𝑖≤𝑛 1≤𝑖<𝑗≤𝑛

+ ∑ |𝐴𝑖 ∩ 𝐴𝑗 ∩ 𝐴𝑘 | − … + (−1)𝑛+1 |𝐴1 ∩ . . .∩ 𝐴𝑛 |

1≤𝑖<𝑗<𝑘≤𝑛

Note:- 1. For ‘n’ number of sets, we have 2𝑛 − 1 number of terms.

2.The coefficient of a term is -1 if the number of sets in the intersection is even and 1
otherwise.

1
Note: - The inclusion – exclusion principle is used to solve problems that ask for the number of elements
in a set that have none of ‘𝑛’ properties 𝑝1 , 𝑝2 , − − −, 𝑝𝑛 , denoted by 𝑁(𝑝1 ′ 𝑝2 ′ − − − 𝑝𝑛 ′ ) and defined
by
𝑁(𝑝1 ′ 𝑝2 ′ − − − 𝑝𝑛 ′ ) = 𝑁 − |𝐴1 ∪ 𝐴2 ∪ 𝐴3 ∪ − − − 𝐴𝑛 | , where 𝑁 is the no of elements which
belong to the universal set (𝑈) and 𝐴𝑖 ∈ 𝑈 and 𝐴𝑖 has property 𝑝𝑖
Example 1: 12 balls are painted in the following ways: 2 are painted red, 1 is painted blue, 1 is painted
white, 2 are painted red & blue, 1 is painted red & white and 3 are painted red, blue & white. How many
balls are not painted?
Solution:- Let 𝑝1 , 𝑝2 , 𝑝3 denotes the properties that a ball is painted red, blue, white respectively; then
𝑁(𝑝1 ) = 8, 𝑁(𝑝2 ) = 6, 𝑁(𝑝3 ) = 5, 𝑁(𝑝1 𝑝2 ) = 5, 𝑁(𝑝1 𝑝3 ) = 4 ,
𝑁(𝑝2 𝑝3 ) = 3, 𝑁(𝑝1 𝑝2 𝑝3 ) = 3
Then 𝑁(𝑝1 ′ 𝑝2 ′ 𝑝3 ′ ) = 𝑁 − |𝐴1 ∪ 𝐴2 ∪ 𝐴3 |
= 𝑁 − ( 𝑁(𝑝1 ) + 𝑁(𝑝2 ) + 𝑁(𝑝3 ) − 𝑁(𝑝1 𝑝2 −𝑁(𝑝1 𝑝3 ) − 𝑁(𝑝2 𝑝3 ) + 𝑁(𝑝1 𝑝2 𝑝3 ))
= 12– [8 + 6 + 5– 4– 3 + 3] = 2
Example 2: 1. Determine the number of positive integer ‘𝑛’ where 1 ≤ 𝑛 ≤ 100 and 𝑛 is not divisible
by 2, 3, 5.
Solution:- Here 𝑈 = {1, 2, 3, − − −, 100}, 𝑁 = 100,
For 𝑛 ∈ 𝑈, 𝑛 satisfies: 𝑝2 if 𝑛 is divisible by 2.
𝑝3 if 𝑛 is divisible by 3.
𝑝5 if 𝑛 is divisible by 5.
Note: - ⌊𝑥⌋ = is the greatest integer 𝑙 such that 𝑙 ≤ 𝑥.
100 100
Now, 𝑁(𝑝2 ) = ⌊ ⌋ = 50 , 𝐴2 = {2, 4, 6, − − −}, 𝑁(𝑝2 𝑝3 ) = ⌊ ⌋ = 16
2 6
100 100
𝑁(𝑝3 ) = ⌊ ⌋ = 33, 𝐴3 = {3, 6, 9, − − −}, 𝑁(𝑝2 𝑝5 ) = ⌊ 10 ⌋ = 10
3
100 100
𝑁(𝑝5 ) = ⌊ ⌋ = 20, 𝐴5 = {5, 10, 15, − − −}, 𝑁(𝑝3 𝑝5 ) = ⌊ 15 ⌋ = 6
5
100
𝑁(𝑝2 𝑝3 𝑝5 ) = ⌊ 30 ⌋ = 3

Then 𝑁(𝑝2 ′ 𝑝3 ′ 𝑝5 ′ ) = 𝑁 − |𝐴2 ∪ 𝐴3 ∪ 𝐴5 |

= 𝑁 − ( 𝑁(𝑝2 ) + 𝑁(𝑝3 ) + 𝑁(𝑝5 ) − 𝑁(𝑝2 𝑝3 )
− 𝑁(𝑝2 𝑝5 ) − 𝑁(𝑝3 𝑝5 ) + 𝑁(𝑝2 𝑝3 𝑝5 ))
= 100 – (50 + 33 + 20 – 10 − 16 – 6 + 3)
= 26 number are not divisible by 2, 3, 5.

2
2. Using question (1);
i) Find the number of positive integers that are not divisible by 2 nor by 5 but are divisible by 3?
ii) Find the number of positive integers that are not divisible by 5 but are divisible
by 2 &3
Solution:- i) 𝑁(𝑝2 ′ 𝑝3 𝑝5 ′ ) = 𝑁(𝑝3 ) − 𝑁(𝑝2 𝑝3 ) − 𝑁(𝑝3 𝑝5 ) + 𝑁(𝑝2 𝑝3 𝑝5) = 33-16-6+3
ii) 𝑁(𝑝2 𝑝3 𝑝5 ′ ) = 𝑁(𝑝2 𝑝3 ) − 𝑁(𝑝2 𝑝3 𝑝5 ) = 16 − 3 = 13

4.2. Recurrence Relation

Consider the sequence 2, 4, 8, 16, - - - can be defined recursively like
𝑎𝑘+1 = 2𝑎𝑘 , 𝑎1 = 2, 𝑘 ≥ 1 . . . (∗)
The equation 𝑎𝑘+1 = 2𝑎𝑘 in (∗) , which defines one member of the sequence interms of a previous one
is called a recurrence relation.
The equation 𝑎1 = 2 is called an initial (boundary) condition.
Example 3: 1. Let {𝑎𝑛 } be a sequence that satisfies the recurrence relation
𝑎𝑛 = 𝑎𝑛−1 − 𝑎𝑛−2 , for 𝑛 = 2, 3, 4, . . . , 𝑎0 = 3 and 𝑎1 = 5 . Then find 𝑎2 , 𝑎3 , 𝑎4 , 𝑎5
2. Suppose that 𝑓 is defined recursively by 𝑓(0) = 3, 𝑓(𝑛 + 1) = 2𝑓(𝑛) + 3. Then find
𝑓(1), 𝑓(2), 𝑓(3), 𝑓(4)
3. The Fibonacci numbers 𝑓0, 𝑓1 , 𝑓2 , . . . are defined by the equations 𝑓0 = 0, 𝑓1 = 1
and 𝑓𝑛 = 𝑓𝑛−1 + 𝑓𝑛−2 , 𝑓𝑜𝑟 𝑛 = 2,3,4, . . . . Then find 𝑓2, 𝑓3 , 𝑓4 , 𝑓5, 𝑓6 , 𝑓7
Definition 2: The general linear recurrence relation of degree 𝑘 with constant coefficients has the form
𝑐0 𝑎𝑛 + 𝑐1 𝑎𝑛−1 + 𝑐2 𝑎𝑛−2 + . . . + 𝑐𝑘 𝑎𝑛−𝑘 = 𝑓(𝑛), where
𝑐0, 𝑐1 , 𝑐2 , . . . , 𝑐𝑘 are real numbers, 𝑐0, 𝑐𝑘 ≠ 0 and 𝑘 ∈ 𝑁
Note:- When 𝑓(𝑛) = 0, ∀𝑛 ∈ 𝑁, the relation is called homogeneous, otherwise
non-homogeneous.
Example 4: The recurrence relation
a) 𝑝𝑛 = 5𝑝𝑛−1 is linear, homogeneous with degree 1.
b) 𝑓𝑛 = 𝑓𝑛−1 + 𝑓𝑛−2 is linear, homogeneous with degree 2.
c) 𝑎𝑛 = 𝑎𝑛−5 is linear, homogeneous with degree 5.
d) 𝑓𝑛 = 𝑓𝑛−1 + 𝑓 3 𝑛−2 is not linear, it is homogeneous with constant coefficient.
e) 𝑓𝑛 = 2𝑓𝑛−1 𝑓𝑛−2 + 1 is not linear, not homogeneous but it has constant coeffi.
f) 𝐵𝑛 = 𝑛𝐵𝑛−1 is linear, homogeneous but the coefficient is not constant.

3
4.3 Solution of recurrence relation
4.3.1 Homogeneous first order recurrence relation
Form: 𝑎𝑛 = 𝑟𝑎𝑛−1 , 𝑛 ≥ 1, 𝑎0 = 𝐴, 𝑟 is constant.
Its general solution is 𝑎𝑛 = 𝐴𝑟 𝑛
Example 5: Solve the recurrence relation
a) 𝑎𝑛 = 3𝑎𝑛−1 , 𝑛 ≥ 1, 𝑎0 = 5
b) 𝑎𝑛 = 7𝑎𝑛−1 , 𝑛 ≥ 1, 𝑎2 = 98
c) 𝑎𝑛 = 𝑛𝑎𝑛−1 , 𝑛 ≥ 1, 𝑎0 = 1
Solution:- a) Since 𝑟 = 3, 𝐴 = 5, then 𝑎𝑛 = 5(3𝑛 ) is the solution
b) Since 𝑟 = 7, 𝐴 = 𝑎0 = 2 because 𝑎2 = 7𝑎1 𝑎𝑛𝑑 𝑎1 = 7𝑎0 ⟹ 𝑎0 = 2
Then 𝑎𝑛 = 2(7𝑛 ) is the solution
c) Since the coefficient of 𝑎𝑛−1 = 𝑛 , which is not constant, we can not use the above
formula.
𝑎1 = 1, 𝑎0 = 1, 𝑎2 = 2𝑎1 = 2.1, 𝑎3 = 3𝑎2 = 3.2 , 𝑎4 = 4𝑎3 = 4.3.2 = 4!
𝑎5 = 5! . Thus 𝑎𝑛 = 𝑛!
4.3.2 Homogeneous second order recurrence relation
Form: 𝑐0 𝑎𝑛 + 𝑐1 𝑎𝑛−1 + 𝑐2 𝑎𝑛−2 = 0, 𝑛 ≥ 2,. . . (∗)
where 𝑐0 , 𝑐1 , 𝑐2 ∈ ℝ (real no), 𝑐0 , 𝑐2 ≠ 0
First we assume for solutions of the form 𝑎𝑛 = 𝐴𝑟 𝑛 , where 𝐴, 𝑟 ≠ 0 and substituting
in to the above equation (∗), we have
𝑐0 𝐴𝑟 𝑛 + 𝑐1 𝐴𝑟 𝑛−1 + 𝑐2 𝐴𝑟 𝑛−2 = 0
⟹ 𝑐0 𝑟 2 + 𝑐1 𝑟 + 𝑐2 = 0 . . . . . . . (∗∗)
which is called the characteristic equation of the recurrence relation.
Then we may have two roots, 𝑟1 and 𝑟2 are called the characteristic roots.
For these roots we have 3 cases:
Case1:- Distinct real roots
In this case the general solution of the recurrence relation is
𝑎𝑛 = 𝑏1 𝑟1 𝑛 + 𝑏2 𝑟2 𝑛 , where 𝑏1 , 𝑏2 are arbitrary constants.
Example 6:- Solve the recurrence relations
a) 𝑎𝑛 + 𝑎𝑛−1 − 6𝑎𝑛−2 = 0, 𝑛 ≥ 2 , 𝑎0 = −1 , 𝑎1 = 8
b) 𝑎𝑛 = 𝑎𝑛−1 + 2𝑎𝑛−2 , 𝑛 ≥ 2 , 𝑎0 = 2 , 𝑎1 = 7
c) 𝑎𝑛+2 = 𝑎𝑛+1 + 𝑎𝑛 , 𝑛 ≥ 0 , 𝑎0 = 0 , 𝑎1 = 1

4
 Solution:- a) Let 𝑎𝑛 = 𝑐𝑟 𝑛 with 𝑐, 𝑟 ≠ 0. Then substitute this on
𝑎𝑛 + 𝑎𝑛−1 − 6𝑎𝑛−2 = 0, we have 𝑐𝑟 𝑛 + 𝑐𝑟 𝑛−1 − 6𝑐𝑟 𝑛−2 = 0
⟹ 𝑟2 + 𝑟 − 6 = 0
⟹ 𝑟 = 2 𝑜𝑟 𝑟 = −3
Here we have two distinct real roots, so, we can 𝑎𝑛 = 2𝑛 and 𝑎𝑛 = (−3)𝑛 which are both solutions.
They are linearly independent solutions because one is not a multiple of the other, i.e, there is no real
constant 𝑘 such that (−3)𝑛 = 𝑘(2)𝑛 for all 𝑛𝜖ℕ.
Therefore its general solution is the form 𝑎𝑛 = 𝑏1 2𝑛 + 𝑏2 (−3)𝑛 , where 𝑏1 , 𝑏2 are arbitrary constants.
Now to determine 𝑏1 and 𝑏2 , use 𝑎0 = −1, 𝑎1 = 8 .
If 𝑎0 = −1 = 𝑏1 20 + 𝑏2 (−3)0 ⟹ −1 = 𝑏1 + 𝑏2
If 𝑎1 = 8 = 𝑏1 21 + 𝑏2 (−3)1 ⟹ 8 = 2𝑏1 − 3𝑏2
Solve this system of equations, we have 𝑏1 = 1, 𝑏2 = −2
∴ 𝑎𝑛 = 2𝑛 − 2(−3)𝑛 for 𝑛 ≥ 0 is its general solution.
b) Let 𝑎𝑛 = 𝑐𝑟 𝑛 with 𝑐, 𝑟 ≠ 0. Then substitute this on 𝑎𝑛 = 𝑎𝑛−1 + 2𝑎𝑛−2,
we have 𝑐𝑟 𝑛 = 𝑐𝑟 𝑛−1 + 2𝑐𝑟 𝑛−2
⟹ 𝑟2 − 𝑟 − 2 = 0
⟹ 𝑟 = −1 𝑜𝑟 𝑟 = 2
Here we have two distinct real roots, so, we can 𝑎𝑛 = 2𝑛 and 𝑎𝑛 = (−1)𝑛 which are both solutions.
They are linearly independent solutions because one is not a multiple of the other, i.e, there is no real
constant 𝑘 such that (−1)𝑛 = 𝑘(2)𝑛 for all 𝑛𝜖ℕ.
Therefore its general solution is the form 𝑎𝑛 = 𝑏1 2𝑛 + 𝑏2 (−1)𝑛 , where 𝑏1 , 𝑏2 are arbitrary constants.
Now to determine 𝑏1 and 𝑏2 , use 𝑎0 = 2, 𝑎1 = 7 .
If 𝑎0 = 2 = 𝑏1 20 + 𝑏2 (−1)0 ⟹ 2 = 𝑏1 + 𝑏2
If 𝑎1 = 7 = 𝑏1 21 + 𝑏2 (−1)1 ⟹ 7 = 2𝑏1 − 𝑏2
Solve this system of equations, we have 𝑏1 = 3, 𝑏2 = −1
∴ 𝑎𝑛 = 3(2𝑛 ) − 1(−1)𝑛 for 𝑛 ≥ 0 is its general solution.

c) Let 𝑎𝑛 = 𝑐𝑟 𝑛 with 𝑐, 𝑟 ≠ 0. Then substitute this on 𝑎𝑛+2 = 𝑎𝑛+1 + 𝑎𝑛 ,

we have 𝑐𝑟 𝑛+2 = 𝑐𝑟 𝑛+1 + 𝑐𝑟 𝑛
⟹ 𝑟2 − 𝑟 − 1 = 0
1 − √5 1 + √5
⟹𝑟= 𝑜𝑟 𝑟 =
2 2

5
 𝑛 𝑛
1−√5 1+√5
Here we have two distinct real roots, so, we can 𝑎𝑛 = ( ) and 𝑎𝑛 = ( ) which are both
2 2

solutions. They are linearly independent solutions because one is not a multiple of the other, i.e, there is
𝑛 𝑛
1−√5 1+√5
no real constant 𝑘 such that ( ) = 𝑘( ) for all 𝑛𝜖ℕ.
2 2
𝑛 𝑛
1−√5 1+√5
Therefore its general solution is the form 𝑎𝑛 = 𝑏1 ( ) + 𝑏2 ( ) , where 𝑏1 , 𝑏2 are arbitrary
2 2

constants.
Now to determine 𝑏1 and 𝑏2 , use 𝑎0 = 0, 𝑎1 = 1 .
0 0
1−√5 1+√5
If 𝑎0 = 0 = 𝑏1 ( ) + 𝑏2 ( ) ⟹ 0 = 𝑏1 + 𝑏2 ⟹ 𝑏1 = −𝑏2
2 2
1 1
1−√5 1+√5 1−√5 1+√5
If 𝑎1 = 1 = 𝑏1 ( ) + 𝑏2 ( ) ⟹ 1 = 𝑏1 ( ) + 𝑏2 ( )
2 2 2 2
1 1
Solve this system of equations, we have 𝑏1 = − , 𝑏2 =
√5 √5
𝑛 𝑛
1 1−√5 1 1+√5
∴ 𝑎𝑛 = − ( ) + ( ) for 𝑛 ≥ 0 is its general solution.
√5 2 √5 2

Case 2:- Repeated real roots

If 𝑟1 = 𝑟2 = 𝑟, then the general solution of the recurrence relation is
𝑎𝑛 = 𝑏1 𝑟1 𝑛 + 𝑏2 𝑛𝑟2 𝑛 or 𝑎𝑛 = 𝑏1 𝑛𝑟1 𝑛 + 𝑏2 𝑟2 𝑛 , where 𝑏1 , 𝑏2 are arbitrary constants.
Example 7:- Solve the recurrence relation
a) 𝑎𝑛+2 = 4 𝑎𝑛+1 − 4𝑎𝑛 , 𝑛 ≥ 0 𝑎𝑛𝑑 𝑎0 = 1, 𝑎1 = 3
b) 𝑎𝑛 = 6 𝑎𝑛−1 − 9𝑎𝑛−2 , 𝑛 ≥ 2 𝑎𝑛𝑑 𝑎0 = 1, 𝑎1 = 12
c) 𝑎𝑛 = 6 𝑎𝑛−1 − 9𝑎𝑛−2 , 𝑛 ≥ 2 𝑎𝑛𝑑 𝑎0 = 1, 𝑎1 = 6
Solution:- a) Let 𝑎𝑛 = 𝑐𝑟 𝑛 with 𝑐, 𝑟 ≠ 0 and 𝑛 ≥ 0 Then substitute this on
𝑎𝑛+2 = 4𝑎𝑛+1 − 4𝑎𝑛 , we have 𝑐𝑟 𝑛+2 = 4𝑐𝑟 𝑛+1 − 4𝑐𝑟 𝑛
⟹ 𝑟 2 − 4𝑟 + 4 = 0 ⟹ 𝑟 = 4
Thus its general solution is the form 𝑎𝑛 = 𝑏1 2𝑛 + 𝑏2 𝑛2𝑛 , where 𝑏1 , 𝑏2 are arbitrary constants.
Now to determine 𝑏1 and 𝑏2 , use 𝑎0 = 1, 𝑎1 = 3 .
If 𝑎0 = 1 = 𝑏1 20 + 𝑏2 (0)(2)0 ⟹ 1 = 𝑏1
1
If 𝑎1 = 3 = 𝑏1 21 + 𝑏2 (1)(2)1 ⟹ 3 = 2𝑏1 + 𝑏2 ⟹ 𝑏2 = 2
1
∴ General solution 𝑎𝑛 = 2𝑛 + 2 𝑛2𝑛 , 𝑛 ≥ 0

Case 3:- Complex roots

In this case the solution could be expressed in the same way as in case of distinct real roots. Before
getting into the case of complex roots , we recall Demoiver’s theorem:

6
  (𝑐𝑜𝑠𝜃 + 𝑖𝑠𝑖𝑛𝜃)𝑛 = 𝑐𝑜𝑠𝑛𝜃 + isin 𝑛𝜃 ; 𝑛 ≥ 0
 If 𝑧 = 𝑥 + 𝑖𝑦 ∈ ℂ ; 𝑧 ≠ 0, where ℂ is complex number, y

then can write r

y
i) For 𝑥 ≠ 0 , 𝑧 = 𝑟(𝑐𝑜𝑠𝜃 + 𝑖𝑠𝑖𝑛𝜃), where 𝑟 = √𝑥 2 + 𝑦 2
𝜃
x x
ii) for 𝑥 = 0 and 𝑦 > 0,
𝜋 𝜋 𝜋
𝑧 = 𝑖𝑦 = 𝑟(𝑐𝑜𝑠 2 + 𝑖𝑠𝑖𝑛 2 ) = 𝑦𝑖𝑠𝑖𝑛 2

iii) for 𝑥 = 0 and 𝑦 < 0,

−𝜋 −𝜋 𝜋 𝜋
𝑧 = 𝑖𝑦 = 𝑟(𝑐𝑜𝑠 + 𝑖𝑠𝑖𝑛 ) = 𝑟(𝑐𝑜𝑠 2 − 𝑖𝑠𝑖𝑛 2 )
2 2

 In all case 𝑧 𝑛 = 𝑟 𝑛 (cos 𝑛𝜃 + 𝑖 sin 𝑛𝜃), for 𝑛 ≥ 0 by Demoiver’s theorem

10
Example 8:- Determine (1 + 𝑖√3)
10
Solution:- (1 + 𝑖√3) = (𝑟(𝑐𝑜𝑠𝜃 + 𝑖sin𝜃))10 but 𝑟 = 2, 𝑥 = 1, 𝑦 = √3
𝜋 𝜋 10 10𝜋 10𝜋 −1 √3
= (2(𝑐𝑜𝑠 3 + 𝑖sin 3 )) =210 (cos + 𝑖sin ) = 210 ( 2 − 𝑖)
3 3 2

Example 9:- Solve the recurrence relations

a) 𝑎𝑛+2 + 4𝑎𝑛 = 0, 𝑛 ≥ 0 𝑎𝑛𝑑 𝑎0 = −1, 𝑎1 = 4
b) 𝑎𝑛 = 2(𝑎𝑛−1 − 𝑎𝑛−2 ), 𝑛 ≥ 2 𝑎𝑛𝑑 𝑎0 = 1, 𝑎1 = 2
c) 𝑎𝑛+2 + 4𝑎𝑛+1 + 16𝑎𝑛 = 0, 𝑛 ≥ 0 𝑎𝑛𝑑 𝑎0 = 2, 𝑎1 = 10
Solution:- a) Let 𝑎𝑛 = 𝑏𝑟 𝑛 , 𝑤ℎ𝑒𝑟𝑒 𝑏, 𝑟 ≠ 0 𝑎𝑛𝑑 𝑛 ≥ 0.
Then 𝑎𝑛+2 + 4𝑎𝑛 = 0 becomes 𝑏𝑟 𝑛+2 + 4𝑏𝑟 𝑛 = 0
⟹ 𝑟 2 + 4 = 0 ⟹ 𝑎1 = 2𝑖, 𝑎2 = −2𝑖
It follows that the general solution of the recurrence relation is given by
𝑎𝑛 = 𝑏1 (2𝑖)𝑛 + 𝑏2 (−2𝑖)𝑛
𝜋 𝜋 𝑛 𝜋 𝜋 𝑛
⇒ 𝑎𝑛 = 2𝑛 [𝑏1 (𝑐𝑜𝑠 2 + 𝑖𝑠𝑖𝑛 2 ) + 𝑏2 (𝑐𝑜𝑠 2 − 𝑖𝑠𝑖𝑛 2 ) ]
𝑛𝜋 𝑛𝜋 𝑛𝜋 𝑛𝜋
= 2𝑛 [𝑏1 (𝑐𝑜𝑠 + 𝑖𝑠𝑖𝑛 ) + 𝑏2 (𝑐𝑜𝑠 − 𝑖𝑠𝑖𝑛 )]
2 2 2 2
𝑛𝜋 𝑛𝜋
= 2𝑛 [(𝑏1 + 𝑏2 )𝑐𝑜𝑠 + 𝑖(𝑏1 − 𝑏2 )𝑠𝑖𝑛 ]
2 2
𝑛𝜋 𝑛𝜋
= 2𝑛 [𝑐1 𝑐𝑜𝑠 + 𝑐2 𝑠𝑖𝑛 ] , where 𝑐1 = 𝑏1 + 𝑏2 , 𝑐2 = 𝑖(𝑏1 − 𝑏2 )
2 2

But 𝑎0 = −1 = 20 [𝑐1 𝑐𝑜𝑠0 + 𝑐2 𝑠𝑖𝑛0] = 𝑐1 ⟹ 𝑐1 = −1

𝜋 𝜋
𝑎1 = 4 = 21 [𝑐1 𝑐𝑜𝑠 2 + 𝑐2 𝑠𝑖𝑛 2 ] = 2𝑐2 ⟹ 𝑐2 = 2
𝑛𝜋 𝑛𝜋
Therefore, 𝑎𝑛 = 2𝑛 [−𝑐𝑜𝑠 + 2𝑠𝑖𝑛 ] is the solution.
2 2

7
b) Let 𝑎𝑛 = 𝑏𝑟 𝑛 , 𝑤ℎ𝑒𝑟𝑒 𝑏, 𝑟 ≠ 0 𝑎𝑛𝑑 𝑛 ≥ 0. Then 𝑎𝑛 = 2(𝑎𝑛−1 − 𝑎𝑛−2 ) becomes
𝑏𝑟 𝑛 = 2(𝑏𝑟 𝑛−1 − 𝑏𝑟 𝑛−2 )
⟹ 𝑟 𝑛 = 2(𝑟 𝑛−1 − 𝑟 𝑛−2 ) , since divide both side by 𝑏.
⟹ 𝑟 2 − 2𝑟 + 2 = 0
⟹ 𝑟1 = 1 + 𝑖, 𝑟2 = 1 − 𝑖 , Now since x = 1, y = 1, then 𝑟 = √2
Then the general solution is 𝑎𝑛 = 𝑏1 (1 + 𝑖)𝑛 + 𝑏2 (1 − 𝑖)𝑛
𝑛 𝑛 𝑛
⟹ 𝑎𝑛 = (√2) [𝑏1 (𝑐𝑜𝑠𝜋4 + 𝑖 𝑠𝑖𝑛𝜋4) + 𝑏2 (𝑐𝑜𝑠−𝜋
4
+ 𝑖 𝑠𝑖𝑛−𝜋
4
) ]
𝑛
= (√2) [𝑏1 (𝑐𝑜𝑠𝑛𝜋
4
+ 𝑖 𝑠𝑖𝑛𝑛𝜋
4
) + 𝑏2 (𝑐𝑜𝑠𝑛𝜋
4
− 𝑖 𝑠𝑖𝑛𝑛𝜋
4
)]
𝑛
= (√2) [(𝑏1 + 𝑏2 )𝑐𝑜𝑠𝑛𝜋
4
+ (𝑏1 − 𝑏2 )𝑖 𝑠𝑖𝑛𝑛𝜋
4
]
𝑛
= (√2) [𝑐1 𝑐𝑜𝑠𝑛𝜋
4
+ 𝑐2 𝑠𝑖𝑛𝑛𝜋
4
], where 𝑐1 = 𝑏1 + 𝑏2 , 𝑐2 = (𝑏1 − 𝑏2 )𝑖
To find 𝑐1 𝑎𝑛𝑑 𝑐2 , use initial condition, i.e.
0
𝑎0 = 1 = (√2) [𝑐1 𝑐𝑜𝑠 0 + 𝑐2 𝑠𝑖𝑛 0] ⟹ 𝑐1 = 1
1
𝑎1 = 2 = (√2) [𝑐1 𝑐𝑜𝑠𝜋4 + 𝑐2 𝑠𝑖𝑛𝜋4] ⟹ 2 = √2(𝑐1 √2
2
+ 𝑐2 √2
2
)
⟹ 2 = 𝑐1 + 𝑐2 ⟹ 𝑐2 = 1
𝑛
Thus the general solution is 𝑎𝑛 = (√2) [𝑐𝑜𝑠𝑛𝜋
4
+ 𝑠𝑖𝑛𝑛𝜋
4
].
c)Let 𝑎𝑛 = 𝑏𝑟 𝑛 , 𝑤ℎ𝑒𝑟𝑒 𝑏, 𝑟 ≠ 0 𝑎𝑛𝑑 𝑛 ≥ 0.Then 𝑎𝑛+2 + 4𝑎𝑛+1 + 16𝑎𝑛 = 0 becomes
𝑏𝑟 𝑛+2 + 4𝑏𝑟 𝑛+1 + 16𝑏𝑟 𝑛 = 0
⟹ 𝑟 2 + 4𝑟 + 16 = 0
⟹ 𝑟1 = −2 + 2√3 𝑖, 𝑟2 = −2 − 2√3 𝑖
Since 𝑥 = −2 𝑎𝑛𝑑 𝑦 = 2√3, 𝑟 = 4
Then the general solution is 𝑎𝑛 = 𝑏1 (−2 + 2√3𝑖)𝑛 + 𝑏2 (−2 − 2√3𝑖)𝑛

2𝜋 2𝜋 𝑛 −2𝜋 −2𝜋 𝑛
⟹ 𝑎𝑛 = 4𝑛 [𝑏1 (𝑐𝑜𝑠 + 𝑖 𝑠𝑖𝑛 ) + 𝑏2 (𝑐𝑜𝑠 + 𝑖 𝑠𝑖𝑛 ) ]
3 3 3 3

2𝜋 2𝜋 𝑛 2𝜋 2𝜋 𝑛
= 4𝑛 [𝑏1 (𝑐𝑜𝑠 + 𝑖 𝑠𝑖𝑛 ) + 𝑏2 (𝑐𝑜𝑠 − 𝑖 𝑠𝑖𝑛 ) ]
3 3 3 3

2𝑛𝜋 2𝑛𝜋 2𝑛𝜋 2𝑛𝜋

= 4𝑛 [𝑏1 (𝑐𝑜𝑠 + 𝑖 𝑠𝑖𝑛 ) + 𝑏2 (𝑐𝑜𝑠 − 𝑖 𝑠𝑖𝑛 )]
3 3 3 3
2𝑛𝜋 2𝑛𝜋
= 4𝑛 [(𝑏1 + 𝑏2 )𝑐𝑜𝑠 + (𝑏1 − 𝑏2 )𝑖 𝑠𝑖𝑛 ]
3 3

8
 2𝑛𝜋 2𝑛𝜋
= 4𝑛 [𝑐1 𝑐𝑜𝑠 + 𝑐2 𝑠𝑖𝑛 ], where 𝑐1 = 𝑏1 + 𝑏2 , 𝑐1 = (𝑏1 − 𝑏2 )𝑖
3 3

To find the value of 𝑐1 and 𝑐2 , use initial condition, i.e.

𝑎0 = 2 = 40 [𝑐1 cos 0 + 𝑐2 sin 0] ⟹ 𝑐1 = 2
2𝜋 2𝜋 10 −1 2√3
𝑎1 = 10 = 41 [𝑐1 𝑐𝑜𝑠 + 𝑐2 𝑠𝑖𝑛 ]⟹ = 𝑐1 ( 2 ) + 𝑐2 ( )
3 3 4 4
5 −1 √3
⟹2= 𝑐1 + 𝑐
2 2 2

7√3
⟹ 𝑐2 =
3
2𝑛𝜋 7√3 2𝑛𝜋
Hence the general solution is 𝑎𝑛 = 4𝑛 [2𝑐𝑜𝑠 + 𝑠𝑖𝑛 ].
3 3 3

4.4 First and second order non- homogeneous recurrence relations

The general solution for non- homogeneous recurrence relation is
𝑎𝑛 = 𝑎𝑛 (ℎ) + 𝑎𝑛 (𝑝) , where 𝑎𝑛 (ℎ) is homogeneous solution,
𝑎𝑛 (𝑝) is particular solution.
4.4.1 Non- homogeneous first order recurrence relation
Form: 𝑎𝑛 − 𝑘𝑎𝑛−1 = 𝑐𝑟 𝑛 , where 𝑐 is a constant and 𝑛 ∈ ℤ+
 If 𝑟 𝑛 is not a solution of a homogeneous relation 𝑎𝑛 − 𝑘𝑎𝑛−1 = 0, then
𝑎𝑛 (𝑝) = 𝐴𝑟 𝑛 , where 𝐴 is constant.
 If 𝑟 𝑛 is a solution of a homogeneous relation 𝑎𝑛 − 𝑘𝑎𝑛−1 = 0, then
𝑎𝑛 (𝑝) = 𝐵𝑛𝑟 𝑛 , where 𝐵 is constant.
Example 10:- Solve a) 𝑎𝑛 − 3𝑎𝑛−1 = 5(7𝑛 ), 𝑛 ≥ 1, 𝑎0 = 2 − − − − − −(∗)
b) 𝑎𝑛 − 3𝑎𝑛−1 = 5(3𝑛 ), 𝑛 ≥ 1, 𝑎0 = 2 − − − − − −(∗∗)
c) 𝑎𝑛 − 2𝑎𝑛−1 = 5(3𝑛 ), 𝑛 ≥ 1, 𝑎0 = 2 − − − − − −(∗∗∗)
Solution:- a) Let 𝑎𝑛 (ℎ) = 𝐶𝑟 𝑛 , where 𝐶, 𝑟 ≠ 0. Then substitute on 𝑎𝑛 − 3𝑎𝑛−1 = 0.
We have 𝐶𝑟 𝑛 − 3𝐶𝑟 𝑛−1 = 0
⟹ 𝑟 − 3 = 0 (by dividing both side with 𝐶𝑟 𝑛−1)
⟹𝑟=3
Thus 𝑎𝑛 (ℎ) = 𝐶3𝑛
Since 7𝑛 is not a solution for homogeneous, 𝑎𝑛 (𝑝) = 𝐴7𝑛 and substitute on 𝑎𝑛 − 3𝑎𝑛−1 =
5(7𝑛 ). We have 𝐴7𝑛 − 3𝐴7𝑛−1 = 5(7𝑛 )
3 35
⟹ 𝐴 − 7𝐴 = 5 ⟹ 𝐴 = 4

9
 35
Thus 𝑎𝑛 (ℎ) = 7𝑛
4

Hence the general solution of (∗) is the form 𝑎𝑛 = 𝑎𝑛 (ℎ) + 𝑎𝑛 (𝑝)

35
⟹ 𝑎𝑛 = 𝐶3𝑛 + 7𝑛
4
35 −27
To find 𝐶 , use initial condition 𝑎0 = 2 = 𝐶(30 ) + (70 ) ⟹ 𝐶 =
4 4
−27 35
Therefore 𝑎𝑛 = (3𝑛 ) + (7𝑛 ) is the general solution of (∗)
4 4

b) Let 𝑎𝑛 (ℎ) = 𝐶𝑟 𝑛 , where 𝐶, 𝑟 ≠ 0. Then substitute on 𝑎𝑛 − 3𝑎𝑛−1 = 0.

We have 𝐶𝑟 𝑛 − 3𝐶𝑟 𝑛−1 = 0
⟹ 𝑟 − 3 = 0 (by dividing both side with 𝐶𝑟 𝑛−1)
⟹𝑟=3
Thus 𝑎𝑛 (ℎ) = 𝐶3𝑛
Since 3𝑛 is a solution for homogeneous, 𝑎𝑛 (𝑝) = 𝐴𝑛3𝑛 and substitute on 𝑎𝑛 − 3𝑎𝑛−1 = 5(3𝑛 ).
We have 𝐴𝑛3𝑛 − 3𝐴(𝑛 − 1)3𝑛−1 = 5(3𝑛 )
3
⟹ 𝐴𝑛 − 3 𝐴(𝑛 − 1) = 5 ⟹ 𝐴 = 5

Thus 𝑎𝑛 (𝑝) = 5𝑛3𝑛

Hence the general solution of (∗∗) is the form 𝑎𝑛 = 𝑎𝑛 (ℎ) + 𝑎𝑛 (𝑝)
⟹ 𝑎𝑛 = 𝐶3𝑛 + 5𝑛3𝑛
To find 𝐶 , use initial condition 𝑎0 = 2 = 𝐶(30 ) + 5.0. (30 ) ⟹ 𝐶 = 2
Therefore 𝑎𝑛 = 2(3𝑛 ) + 5𝑛3𝑛 is general solution of (∗∗)
c) Exercise
4.4.2 Non- homogeneous second order recurrence relation
Form: 𝑐0 𝑎𝑛 + 𝑐1 𝑎𝑛−1 + 𝑐2 𝑎𝑛−2 = 𝑓(𝑛), 𝑛 ≥ 0, ------------------------- (∗)
where 𝑐0 , 𝑐1 , 𝑐2 are real numbers and 𝑐0 , 𝑐2 ≠ 0 .
Suppose 𝑎𝑛 = 𝑐1 𝑎𝑛−1 + 𝑐2 𝑎𝑛−2 + 𝑓(𝑛) , where 𝑐1 , 𝑐2 are real numbers , 𝑐2 ≠ 0 and 𝑓(𝑛) = 𝑝(𝑛)𝑟 𝑛 ,
say 𝑝(𝑛) is polynomial.
When 𝑟 is not a root of characteristic equation of homogeneous.
There is 𝑎𝑛 (𝑝) = 𝑞(𝑛)𝑟 𝑛 , where 𝑞(𝑛) is polynomial.
When 𝑟 is a root of characteristic equation of homogeneous and its multiplicity is ‘𝑚’, there
is 𝑎𝑛 (𝑝) = 𝑛𝑚 𝑞(𝑛)𝑟 𝑛 , where 𝑞(𝑛) is polynomial.
Example 11:- What form 𝑎𝑛 (𝑝) for 𝑎𝑛 = 6𝑎𝑛−1 − 9𝑎𝑛−2 + 𝑓(𝑛) have:
when 𝑓(𝑛) = 3𝑛 , 𝑓(𝑛) = 𝑛3𝑛 , 𝑓(𝑛) = 𝑛2 2𝑛 𝑎𝑛𝑑 𝑓(𝑛) = (𝑛 + 1)2 3𝑛 ?

10
Solution:- 1st find 𝑎𝑛 (ℎ) ?
Let 𝑎𝑛 = 𝑐𝑟 𝑛 , where 𝑐, 𝑟 ≠ 0 and substitute on homogeneous. We have
𝑐𝑟 𝑛 = 6𝑐𝑟 𝑛−1 − 9𝑐𝑟 𝑛−2
⟹ 𝑟 𝑛 − 6𝑟 + 9 = 0
⟹ (𝑟 − 3)𝑛 = 0 has a single root with multiplicity 2. ∴ 𝑎𝑛 (ℎ) = 𝑏1 3𝑛 + 𝑏2 𝑛3𝑛
2nd find 𝑎𝑛 (𝑝) ?
Since 𝑟 = 3 is a root with multiplicity 𝑚 = 2, but 𝑟 = 2 is not a root of charactrstic equation for
homogeneous.Then
𝑎𝑛 (𝑝) has the form 𝑛2 𝐴3𝑛 if 𝑓(𝑛) = 3𝑛
𝑎𝑛 (𝑝) has the form 𝑛2 (𝑝1 𝑛 + 𝑝2 )3𝑛 if 𝑓(𝑛) = 𝑛3𝑛
𝑎𝑛 (𝑝) has the form (𝑝2 𝑛2 + 𝑝1 𝑛 + 𝑝0 )2𝑛 if 𝑓(𝑛) = 𝑛2 2𝑛
𝑎𝑛 (𝑝) has the form 𝑛2 (𝑝2 𝑛2 + 𝑝1 𝑛 + 𝑝0 )3𝑛 if 𝑓(𝑛) = (𝑛 + 1)2 3𝑛
4.5 Method of undetermined coefficients
The method of undetermined coefficients (uc) is a procedure for finding a particular solution 𝑎𝑛 (𝑝) of a
recurrence relation of the form: 𝑐0 𝑎𝑛 + 𝑐1 𝑎𝑛−1 + 𝑐2 𝑎𝑛−2 + . . . + 𝑐𝑘 𝑎𝑛−𝑘 = 𝑓(𝑛) --------------- (*)
𝑐0, 𝑐1 , 𝑐2 , . . . , 𝑐𝑘 are real numbers, 𝑐0, 𝑐𝑘 ≠ 0 and 𝑘 ∈ 𝑁
The method requires that the function 𝑓(𝑛) in the equation (*) be one of the following three forms:

i) 𝑃𝑘 (𝑛) -is polynomial of degree 𝑘.

ii) 𝑃𝑘 (𝑛)𝑟 𝑛
iii) 𝑃𝑘 (𝑛)𝑟 𝑛 sin 𝑛𝜗 or 𝑃𝑘 (𝑛)𝑟 𝑛 cos 𝑛𝜗 , and the method fails if 𝑓(𝑛) is different from
these three forms

𝑓(𝑛) 𝑎𝑛 (𝑝) (where 𝑐, 𝐴, 𝐴0 , 𝐴1 , 𝐴2 are constants.)

C A
Cn 𝐴0 + 𝐴1 𝑛
C n2 𝐴0 + 𝐴1 𝑛 + 𝐴2 𝑛2
𝐴𝑠𝑖𝑛 𝑛𝜃 or 𝐴𝑐𝑜𝑠 𝑛𝜃 𝐴1 𝑠𝑖𝑛 𝑛𝜃 + 𝐴2 𝑐𝑜𝑠 𝑛𝜃
𝐴 𝑟 𝑛 𝑠𝑖𝑛 𝑛𝜃 or 𝐴 𝑟 𝑛 𝑐𝑜𝑠 𝑛𝜃 𝐴1 𝑟 𝑛 𝑠𝑖𝑛 𝑛𝜃 + 𝐴2 𝑟 𝑛 𝑐𝑜𝑠 𝑛𝜃
𝐴𝑛𝑚 𝑟 𝑛 𝑟 𝑛 (𝐴0 + 𝐴1 𝑛 + − − − + 𝐴𝑚 𝑛𝑚 )

Example 12:- Find the general solution of a) 𝑎𝑛 = 3𝑎𝑛−1 + 2𝑛, 𝑎1 = 3

b) 𝑎𝑛 = 5𝑎𝑛−1 − 6𝑎𝑛−2 + 7𝑛

11
Solution:- a) Let 𝑎𝑛 (ℎ) = 𝐴𝑟 𝑛 , where 𝐴 is constant, 𝑟 ≠ 0. and substitute this on the
homogeneous 𝑎𝑛 = 3𝑎𝑛−1, we have 𝑎𝑛 (ℎ) = 𝐴3𝑛
since 𝑓(𝑛) = 2𝑛 is a polynomial with degree 1, 𝑎𝑛 (𝑝) is a linear function in 𝑛, i.e, 𝑎𝑛 (𝑝) =
𝑐 𝑛 + 𝑑, where 𝑐, 𝑑 are constants.Then
𝑎𝑛 = 3𝑎𝑛−1 + 2𝑛 becomes 𝑐 𝑛 + 𝑑 = 3(𝑐(𝑛 − 1) + 𝑑) + 2𝑛
⟹ 𝑐 𝑛 + 𝑑 = 3𝑐 𝑛 − 3𝑐 + 3𝑑 + 2𝑛
⟹ (3𝑐 + 2 − 𝑐)𝑛 − 3𝑐 = 𝑑 − 3𝑑
⟹ (2𝑐 + 2)𝑛 + (2𝑑 − 3𝑐) = 0
⟹ (2𝑐 + 2) = 1 and (2𝑑 − 3𝑐) = 0
−3
⟹ 𝑐 = −1, 𝑑 =
2
3
Thus 𝑎𝑛 (𝑝) = − 𝑛 − is a particular solution.
2
3
All solutions are the form 𝑎𝑛 = 𝑎𝑛 (ℎ) + 𝑎𝑛 (𝑝) = 𝑏1 3𝑛 − 𝑛 − 2

To find 𝑏1 :- use initial condition 𝑎1 = 3

3 3 11
𝑎1 = 3 = 𝑏1 31 − 1 − 2 ⟹ 3𝑏1 = 3 + 1 + 2 ⟹ 𝑏1 = 6
11 3
∴ 𝑎𝑛 = 3𝑛 − 𝑛 − 2 is a general solution.
6

b) To find 𝑎𝑛 (ℎ)
Let 𝑎𝑛 = 𝑐𝑟 𝑛 , where 𝑐, 𝑟 ≠ 0 and substitute on 𝑎𝑛 = 5𝑎𝑛−1 − 6𝑎𝑛−2 , we have
𝑐𝑟 𝑛 = 5𝑐𝑟 𝑛−1 − 6𝑐𝑟 𝑛−2
⟹ 𝑟 2 − 5𝑟 + 6 = 0
5±√25−24
⟹𝑟= 2
Thus the roots are 𝑟1 = 3, 𝑟2 = 2

𝑎𝑛 (ℎ) = 𝑏1 (3𝑛 ) + 𝑏2 (2𝑛 )

Since 𝑓(𝑥) = 7𝑛 , 𝑎𝑛 (ℎ) = 𝐴(7𝑛 ), where 𝐴 is constant. Then substitute on
𝑎𝑛 = 5𝑎𝑛−1 − 6𝑎𝑛−2 + 7𝑛 , we have 𝐴7𝑛 = 5𝐴7𝑛−1 − 6𝐴7𝑛−2 + 7𝑛
⟹ 𝐴72 = 5𝐴7 − 6𝐴 + 72
⟹ (49 − 35 + 6)𝐴 = 49
49
⟹ 𝐴 = 20
49
Thus 𝑎𝑛 (𝑝) = (7𝑛 )
20
49
𝑎𝑛 = 𝑎𝑛 (ℎ) + 𝑎𝑛 (𝑝) = 𝑏1 3𝑛 + 𝑏2 2𝑛 + 20 (7𝑛 )

12