CS 2336

Discrete Mathematics

Lecture 8
Counting: Permutations and Combinations

1
 Outline
• Definitions
• Permutation
• Combination
• Interesting Identities

2
 Definitions
• Selection and arrangement of objects appear in
many places
 We often want to compute # of ways to
select or arrange the objects
• Examples :
1. How many ways to select 2 people
from 5 candidates ?
2. How many ways to arrange 7 books
on the bookshelf ?
3
 Definitions
• In most textbooks, we use the word
combination  selection

An r-combination of n objects is an unordered

selection of r objects from the n objects

• Example :
{ c, d } is a 2-combination of { a, b, c, d, e }

4
 Definitions
• In most textbooks, we use the word
permutation  arrangement

An r-permutation of n objects is an ordered

arrangement of r objects from the n objects

• Example :
cabd is a 4-permutation of { a, b, c, d, e }

5
 Definitions
• Further, we define the following notation:

C(n, r) denotes the number of r-combinations

of n distinct objects

P(n, r) denotes the number of r-permutations

of n distinct objects

• What are the values of C(n, n), C(n, 1), C(3, 2),
and P(3, 2) ?
6
 Test Your Understanding

• Why are the following equalities correct ?

1. P(n, r) = P(r, r)  C(n, r)

2. P(n, n) = P(n, r)  P(n – r, n – r)
3. C(n, r) = C(n, n – r)

7
 Permutation
• In fact, there is a formula for P(n, r) :

P(n, r) = n (n – 1)(n – 2) … (n – r + 1)

• Proof :
P(n, r) = # ways to get r of n objects in some order.
There are n ways to choose the 1st object,
n – 1 ways to choose the 2nd object, … ,
n – r + 1 ways to choose the rth object
 Result follows from rule of product
8
 Examples
• Ex 1 : How many ways to select a first-prize, a
second-prize, and a third-prize winners
from 100 different people ?
• Ex 2 : How many ways can n people be ordered
to form a ring ?
2 1

1 3

3 2

The above are considered the same

(as relative order is the same)

9
 With Indistinguishable Objects
• How many different strings can be made by re-
ordering the letters of the word “SUCCESS” ?
• Answer :
First, suppose that all the 7 letters are distinct.
Then, there will be 7! different strings.
Now, if we make the two Cs indistinguishable, we
will only have 7!/2! different strings.
Further, if the three Ss are indistinguishable, the
number of different strings becomes (7!/2!)/3!
10
 With Indistinguishable Objects
• In general, if there are n objects, with
n1 indistinguishable objects of type 1,
n2 indistinguishable objects of type 2,
…,
nk indistinguishable objects of type k,

 the number of n-permutations is :

n!
n1! n2! … nk!

11
 Examples
• If we have 5 dashes and 8 dots, how many
different ways to arrange them ?
...___._._ ...

• If we can only use 7 symbols of them, how many

different arrangements are there ?
._._._.

12
 Examples
• Show that for any positive integer k,

(k!) ! is divisible by k! (k – 1)! ?

• For instance, when k = 3,

(k!) ! = (3!) ! = 6! = 720
k! (k – 1)! = (3!) 2! = 62 = 36

13
 With Unlimited Repetitions
• Suppose that there are n distinct objects, each
with unlimited supply

• How many r-permutations are there ?

That is, how many ways to get a total of r objects
from them, and then form an arrangement ?

• Answer : nr

14
 Examples
• Ex 1 : Consider all numbers between 1 and 1010
(i) How many of them contain the digit 1 ?
(ii) How many of them do not ?

• Ex 2 :
(i) How many bit strings of length n are there ?
(ii) How many contain even number of 0s?

15
 Combination
• Recall that
P(n, r) = P(r, r)  C(n, r)

• Thus, we have

C(n, r) = n (n – 1)(n – 2) … (n – r + 1) / r!

• Alternatively, we can express C(n, r) as :

C(n, r) = n! / ( (n – r)! r! )

16
 Examples
• Consider a hexagon where no three diagonals
meet a one point

• How many diagonals are there ?

• How many intersections between the diagonals ?
• How many line segments are the diagonals
divided by their intersections ?
17
 Examples
• In how many ways can we select 3 numbers from
1, 2, …, 300, such that their sum is a multiple of 3 ?

• Hint :
When the sum is a multiple of 3, what special
property does the 3 numbers have ?

• Answer : 1003 + 3  C(100, 3)

18
 Examples
• Five pirates have discovered a treasure box
They decided to keep the box in a locked room,
so that all the locks of the room can be opened
if and only if 3 or more pirates are present

• How to do so ? How many locks do they need ?

(Each pirate may possess keys to different locks)

19
 With Unlimited Repetitions
• Suppose that there are n distinct objects, each
with unlimited supply

• How many r-combinations are there ?

That is, how many ways to get a total of r objects
from them, and the ordering is not important ?

• Answer : C(n – 1 + r, r) [Why?]

20
 With Unlimited Repetitions
• Imagine we have a box for each type of objects

1 2 3
… n

• A particular r-combination is equivalent to

throwing a total of r balls into these boxes

1 2 3
… n

21
 With Unlimited Repetitions
• To represent one of the r-combination, we may
use a list of n – 1 bars and r stars, where
 the bars are used to mark off n different boxes
 the stars are used to indicate how many balls
in each box
• For instance, suppose n = 5, r = 6

1 2 3 4 5 **|*| | |* **
22
 With Unlimited Repetitions
• Using the bars-and-stars representation, we see
that
 each r-combination corresponds to a unique
representation (with n – 1 bars and r stars), and
 each representation (with n – 1 bars and r stars)
corresponds to a unique r-combination

 # of r-combinations = # of representations
= C(n – 1 + r, r)

23
 Examples
• Ex 1 : Suppose that a cookie shop has four
different kinds of cookies. How many
different ways can 6 cookies be chosen ?

• Ex 2 : How many solutions does the equation

x + y + z = 11
have, if x, y, z are non-negative integers ?

• Ex 3 : What if x, y, z are positive integers in Ex 2 ?

24
 Interesting Identities
Pascal’s Identity :
C(n, r) = C(n – 1, r) + C(n – 1, r – 1)
• Proof (by combinatorial arguments):
To select r of n objects, there are in two cases :
1. Get the first object, and then get r – 1 objects
from the remaining n – 1 objects ;
2. Do not get the first object, and get r objects
from the remaining n – 1 objects
 In total, C(n – 1, r – 1) + C(n – 1, r) ways
25
 Interesting Identities
Binomial Theorem :
n
(x + y)n = 
r=0
C(n, r) xn – r yr

• Proof (by combinatorial arguments):

The terms in (x + y)n must be of the form xn – r yr.
To obtain the term xn – r yr, x is chosen n – r times
from the n occurrences of (x + y) in the product,
so that y will be automatically chosen r times
 the number of ways is exactly C(n, r)
26
 Examples
• Ex 1 : What is the expansion of (x + y)4 ?

• Ex 2 : What is the coefficient of x12 in (2x – 3y)25 ?

n
• Ex 3 : What is the value of r= 0C(n, r) ?
n
• Ex 4 : What is the value of 
r=0
(–1) r C(n, r) ?

n
• Ex 5 : What is the value of 
r=0
2 r C(n, r) ?

27
 Interesting Identities
Vandermonde’s Identity :
r
C(m + n, r) =  C(m, r – k) C(n, k)
k=0

• Proof (by combinatorial arguments):

To select r items from m + n distinct objects,
we may assume that among these objects,
m are white and n are black. The selection may
start by selecting k black objects, and then the
remaining from white objects. As k can vary
from 0 to r, this gives the result.
28
 Example
n
• Can you simplify k= 0 C(n, k)2 ?

• Answer :
Observe that
n n

k=0
C(n, k) 2 =  C(n, n – k) C(n, k)
k=0

By setting m = n and r = n in Vandermonde’s

identity, we get the desired value as C(2n, n)

29