CSE-801: Stochastic Systems

Random Variables

Muhammad Shahzad
School of Electrical Engineering & Computer Science
National University of Sciences & Technology (NUST)
Pakistan
What will we cover in this lecture?
In this lecture, we will cover:
Definition of a Random Variable

Discrete Random Variables [Bernoulli, Binomial, Geometric, Poisson,

Discrete Uniform]

Continuous Random Variables [Exponential, Uniform, Gaussian, Erlang,

Gamma]

2
Definition of a Random Variable
Random Variable [m-w.org]
: a variable that is itself a function of the result of a statistical
experiment in which each outcome has a definite
probability of occurrence

3
Definition of a Random Variable
A random variable is a mapping from an outcome s of a
random experiment to a real number

4
Definition of a Random Variable

Random
Sample Space X(s) Random
Experiment
Variable
head

0 1 R
tail

Sx

5
Definition of a Random Variable

X(s)

1 2 3 4 5 6 R

Random
Experiment
Sample Space, S Sx
Random Variable

Image courtesy of www.buzzle.com/

6
Definition of a Random Variable
More than one outcomes can be mapped to the same real
number

X(s)

0 1

Random Sx
Experiment
Random Variable
Sample Space

Image courtesy of www.buzzle.com/

7
Types of Random Variables
Discrete random variables: have a countable (finite or infinite)
image
Sx = {0, 1}
Sx = {, -3, -2, -1, 0, 1, 2, 3, }
Continuous random variables: have an uncountable image
Sx = (0, 1]
Sx = R
Mixed random variables: have an image which contains
continuous and discrete parts
Sx = {0} U (0, 1]

We will mostly focus on discrete and continuous random

variables

8
Discrete Random Variables

9
Probability Mass Function
The Probability Mass Function (pmf) or the discrete probability
density function provides the probability of a particular point in
the sample space of a discrete random variable (rv)
For a countable SX={a0, a1, , an}, the pmf is the set of
probabilities

pX(ak)

SX={a0=0, a1=1, , a5=5},

0 1 2 3 4 5 X

10
Properties of a PMF
P1:

P2:

pX(ak)

SX={a0=0, a1=1, , a5=5},

0 1 2 3 4 5 X

11
Cumulative Distribution Function (CDF) of Discrete
Random Variable
The Cumulative Distribution Function (CDF) for a discrete rv is
defined as:

pX(x) FX(x)

pmf
CDF

1 2 3 4 X 1 2 3 4 X

12
Cumulative Distribution Function (CDF) of Discrete
Random Variable
CDF can be used to find the probability of a range of values in a
rvs image:

pX(x) FX(x)

pmf
CDF

1 2 3 4 X 1 2 3 4 X

13
Properties of a CDF
F1:

F2:

pX(x) FX(x)

pmf
CDF

1 2 3 4 X 1 2 3 4 X

14
Properties of a CDF

F3:

F4:

pX(x) FX(x)

pmf
CDF

1 2 3 4 X 1 2 3 4 X

15
Expected Value of a Discrete Random
Variable
The expected value, expectation or mean of a discrete rv is the
average value of the random variable

What is the average value of a random variable whose image is

SX={1, 6, 7, 9, 13}?

16
Expected Value of a Discrete Random
Variable
What is the average value of the following random variable
whose image is SX={1, 6, 7, 9, 13}?

If your answer is 7.2 then you assumed that all of the values in the
rvs image have equal weights

17
Expected Value of a Discrete Random
Variable
Expectation assigns appropriate weights to each value in the rvs
image
Expectation is computed by using the probability of each point in
the rvs image as the weight of that point
Hence, the expectation of the random variable given below is:

pX(x)
0.4
0.3
0.2
0.1

1 6 7 9 13
18
Expected Value of a Discrete Random
Variable
Mathematically, the expected value of a discrete random
variable is:

In some cases, the expected value does not converge

In such cases, we say that the expected value does not exist

19
Variance of a Random Variable
Variance of a rv is a measure of the amount of variation of a rv
around its mean

Intuitively, which of the following discrete rvs has a higher

variance?

pX(x) qX(x)

E{X}=3.87 E{X}=5.2
0.5 0.4
0.4
0.2
0.1
0.033
1 6 7 9 13 X 1 6 7 9 13 X

20
Variance of a Random Variable
Intuitively, which of the following discrete rvs has a higher
variance?
qX(x) has a higher variance because it varies more around its mean
than pX(x)

pX(x) qX(x)

E{X}=3.87 E{X}=5.2
0.5 0.4
0.4
0.2
0.1
0.033
1 6 7 9 13 X 1 6 7 9 13 X

21
Variance of a Random Variable
Mathematically, the variance of a discrete rv is defined as:

22
 Variance of a Random Variable

pX(x) qX(x)

E{X}=3.87 E{X}=5.2
0.5 0.4
0.4 var{X}=15.36

var{X}=9.91 0.2
0.1
0.033
1 6 7 9 13 X 1 6 7 9 13 X

23
Standard Deviation of a Random Variable
In many scenarios, we use the square root of the variance
called its standard deviation

sX = var {X }

24
 Standard Deviation of a Random Variable

pX(x) qX(x)
E{X}=3.87 E{X}=5.2
0.5 0.4
0.4 X=3.92

X=3.14 0.2
0.1
0.033
1 6 7 9 13 X 1 6 7 9 13 X

sX = var {X } = 9.91 = 3.14 sX = var {X } = 15.36 = 3.92

25
Discrete Uniform Random Variable
A discrete uniform rv, D, has a finite image and all the elements
of the image have equal probabilities

Pr{D=k}

1/n

x1 x2 x3 xn
k

What do you think are the expected value and standard

deviation of this random variable?

26
Bernoulli Random Variable
A Bernoulli Random Variable is defined on a single event A
This rv is based on an experiment called a Bernoulli trial
The experiment is performed and the event A either happens or does not
happen
Thus the sample space of a Bernoulli rv is binary
Bernoulli Random
Variable
Sample Space X(s)
Bernoulli Trial: Toss a

A=head
coin

Ac = Not head 0 1 R
= tail

Sx
27
Bernoulli Random Variable
A Bernoulli Random Variable is defined on a single event A
This rv is based on an experiment called the Bernoulli trial
The experiment is performed and the event A either happens or does not
happen
Thus the sample space of a Bernoulli rv is binary
Bernoulli Random
Variable
Pakistan cricket team

X(s)
plays Australia

A=Pak wins

R
Ac=Pak losses
0 1

Sample Space
Image Courtesies of http://www.tribuneindia.com and
images.google.com

28
Bernoulli Random Variable
A Bernoulli Random Variable is defined on a single event A
This rv is based on a the experiment called a Bernoulli trial
The experiment is performed and the event A either happens or does not
happen
Thus the sample space of a Bernoulli rv is binary
Of course the Pr{A} = 0 for this
experiment Bernoulli Random
Variable
Pakistan cricket team

X(s)
plays Australia

A=Pak wins

R
Ac=Pak losses
0 1

Sample Space
Image Courtesies of http://www.tribuneindia.com and
images.google.com 29
Bernoulli Random Variable
Typical examples of Bernoulli rvs in communication:
Transmit a bit over a wireless channel
Outcomes:
0 > bit is received error-free
1 > bit received is not received error-free => bit is received with errors
Transmit a packet over the Internet
Outcomes:
0 > packet is received
1 > packet is not received => packet is lost en-route due to congestion

30
Bernoulli Random Variable
Sample space of a Bernoulli rv, I, is binary
Both outcomes are mapped to real numbers,
Traditionally: I(A) = 1 and I(Ac) = 0 are used to represent a Bernoulli rvs
outcomes
The pmf of I is:
Pr{I = 1} = p
Pr{I = 0} = 1 p

Pr{I=k}

p
1-p

I
0 1

31
Bernoulli Random Variable
Example:
Consider the experiment of a fair coin toss. What is the expected
value and the variance of this Bernoulli rv?

Pr{I=k}

0.5

I
0 1

32
Discrete Random Variables
Example:
Consider the experiment of a fair coin toss. What is the expected
value and the variance of this Bernoulli rv?
Since the coin toss is fair: Pr{I = 1} = 0.5 and Pr{I = 0} = 0.5
E{I} = (1)x(0.5) + (0)x(0.5) = 0.5
var{I} = (10.5)2x(0.5) + (00.5)2x(0.5) = 0.25

Pr{I=k} I=0.5

I=0.5
0.5

I
0 1

33
Bernoulli Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error
0.1. What is the expected value and the variance of this Bernoulli
rv?

Pr{I=k}
0.9

0.1
I
0 1

34
Bernoulli Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error 0.1. What is the
expected value and the variance of this Bernoulli rv?
The event of interest here is a bit-error:
=> Pr{I = 1} = 0.9 and Pr{I = 0} = 0.1
E{I} = (1)x(0.9) + (0)x(0.1) = 0.9
var{I} = (10.9)2x(0.9) + (00.9)2x(0.1) = 0.09
Note that the variance of this pmf is smaller than the variance of the coin toss
pmf
Pr{I=k} I=0.3
0.9
I=0.9

0.1
I
0 1
35
Binomial Random Variable
Consider a collection of n independent Bernoulli trials

A Binomial Random Variable is the total number of occurrences

of an event A in this independent Bernoulli collection
Send n bits, count the number of bits that are received with errors
Send n packets, count the number of packets that are not lost

36
Binomial Random Variable
If Ij(A)=1 and Ij(Ac)=0, j=1,2,,n, are used to represent the
outcomes of the Bernoulli trials then the Binomial Random
Variable, X, is

So what is the image of X?

37
Binomial Random Variable
If Ij(A)=1 and Ij(Ac)=0, j=1,2,,n, are used to represent the
outcomes of the Bernoulli trials then the Binomial Random
Variable, X, is

So what is the image of X?

SX = {0, 1, 2, 3, , n}

38
Binomial Random Variable
Pr{Ij(A)=1} = p and Pr{Ij(A)=0} = 1-p

Then a Binomial rv X is defined as:

And the pmf of a binomial rv is

39
Binomial Random Variable
The pmf of a Binomial rv X is:

This pmf gives the probability that exactly k out of a total of n

Bernoulli trials were successes

Be very careful about the definition of a success

40
Binomial Random Variable

Pr{X=k}

Image courtesy of Wikipedia article on Binomial Distribution

41
Binomial Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error p.
Find the probability that a packet of n bits is received with one
or more errors?

42
Binomial Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error p.
Find the probability that a packet of n bits is received with one
or more errors?

Pr{Ij = 1} = 0.1 and Pr{Ij = 0} = 0.9, j=1,2,,n

Then the probability that a packet is received with errors is

43
Binomial Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error p.
Find the probability that a packet of n bits is received with one
or more errors?

44
Binomial Random Variable
Example:
Consider a binary symmetric channel with probability of bit-error p.
Find the probability that a packet of n bits is received with one
or more errors?

There is an easier way to compute the same probability by noting

that:

45
Connection Between Bernoulli and Binomial
RVs

I(A)=1, I(AC)=0 n Bernoulli trials

S SI SX

A
Pr{I=1}=p
Pr{A}
Pr{I=0}=1-p

Bernoulli Trial Bernoulli RV Binomial RV

46
Normal Approximation of a Binomial RV
When n becomes large, computations using the binomial
formula become unmangeable

For large n, the normal approximation of a binomial rv is

This approximation is also called the Laplace Approximation

Another approximation of binomial is called the Poisson

approximation which will be discussed later

47
Geometric Random Variable
A Geometric Random Variable, M, is the number of Bernoulli
trials until the first occurrence of an event A

The experiment is stopped as soon as event A is observed

The image of a Geometric random variable is infinite but

countable
SM = {1, 2, 3, }

48
Geometric Random Variable

Also called the modified

The pmf of a Geometric Random Variable, Z, is

geometric pmf
OR

Depending on whether the success trial is included in the total

count or not

49
Geometric Random Variable

Image courtesy of Wikipedia article on Geometric Distribution

50
Connection between Geometric and Binomial
RVs
Can we find the probability of a Geometric random variable
using the Binomial random variable?

Geometric

Binomial

51
Connection between Geometric and Binomial
RVs
Can we find the probability of a Geometric random variable using
the Binomial random variable?

52
Connection between Geometric and Binomial
RVs
In k Bernoulli trials, there are k ways in which you can have 1
success and k-1 failures

Since the Binomial random variable counts successes and

failures, it sums and considers all the k outcomes together

53
Connection between Geometric and Binomial
RVs
For k=4, a Binomial random variable X jointly considers the
outcomes 0001, 0010, 0100, 1000
Pr{X = 1} = Pr{(0001) U (0010) U (0100) U (1000)}
Pr{X = 1} = Pr{0001} + Pr{0010} + Pr{0100} + Pr{1000}
Since the underlying Bernoulli trials are independent:
Pr{X = 1} = (k)Pr{one success in k trials}

For the Geometric random variable, we are only interested in

one of these outcomes, 0001
=> Pr{Z = 1} = Pr{X = 1} / k

54
Memoryless Property
It can be shown that the Geometric rv satisfies the memoryless
property

The memoryless property is satisfied when:

This property is also called the Markov Property

55
Memoryless Property
The memoryless property is satisfied when:

For the Geometric rv, RHS of the above equation is:

Thus to prove that the Geometric rv satisfies the memoryless

property, we need to show that

56
Memoryless Property
To prove that the Geometric rv satisfies the memoryless
property, we need to show that

Lets expand the LHS using the definition of conditional

probability

57
Memoryless Property
Continued from last page

58
Memoryless Property
It can also be shown that the Geometric rv is the only discrete
random variable that satisfies the memoryless property

Because of the memoryless property, the Geometric rv can be

thought of as the number of failures between two successes

OR the inter-arrival time between successes

59
Poisson Random Variable
A Poisson Random Variable, N, is the number of occurrences or
arrivals of an event A in a time interval of fixed length
E.g.: The number of packets that arrive at a wireless access point per
second

The Poisson rv assumes that:

The average number of arrivals per time interval, denoted by , is known
The arrivals are independent of each other

Poisson time interval

Poisson arrivals

60
Poisson Random Variable
A good way of understanding Poisson rv is through the Binomial
rv

Let us divide the fixed Poisson time interval into n very small
sub-intervals of length t
t is so small that only one arrival is possible within each sub-interval

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals

61
Poisson Random Variable
Now the experiment can be treated as Binomial rv where a
success corresponds to the presence of an arrival in a sub-
interval

Given that is the average arrival rate, what is the probability of

success (success corresponds to an arrival)?

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals

62
Poisson Random Variable
Now the experiment can be treated as Binomial rv where a
success corresponds to the presence of an arrival in a sub-
interval

Given that is the average arrival rate, what is the probability of

success (success corresponds to an arrival)?
Pr{success} = p = /n

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals

63
Poisson Random Variable
Now the experiment can be treated as Binomial rv with
parameter /n

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals
64
Poisson Random Variable
Lets take the limit when n tends to infinity

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals
65
Poisson Random Variable
Thus for large n, the Binomial pmf approaches the Poisson pmf

This is called the Poisson approximation of Binomial random

variable

Poisson time interval

1 2 3 4 n
t

t Poisson arrivals
66
Poisson Random Variable
In general, the pmf of a Poisson random variable is

where is the average arrival rate (arrivals/time interval)

67
Poisson Random Variable

Pr{N=k}

k
Image courtesy of Wikipedia article on Poisson Distribution

68
End of Discrete Random Variables
At this point, we conclude our discussion on examples of
discrete random variables

We will cover some other aspects of discrete rvs after looking at

examples of continuous rvs

69
Continuous Random Variables

70
Continuous Random Variables
Recall that a Continuous random variable has an uncountable
image
Sx = (0, 1]
Sx = R

71
Probability Density Function (pdf) of a
Continuous Random Variable
The Probability Density Function (pdf) of a continuous random
variable, fX(x), is a continuous function of the x Sx

Properties of pdf:
f1:

f2:

fX(x)

SX=(0, n]

0 n x

72
Probability Density Function (pdf) of a
Continuous Random Variable
f3:

fX(x)

0 n x
a b

73
Cumulative Distribution Function (CDF) of a
Continuous Random Variable
CDF of a continuous rv is computed by integrating the pdf

Conversely, we also have:

FX(x)

x
74
Properties of the CDF of a Continuous
Random Variable
F1:

F2:

F3:

FX(x)

75
Properties of the CDF of a Continuous
Random Variable
F4:

FX(x)

x
76
Properties of the CDF of a Continuous
Random Variable
F4:

for small values of .

FX(x)

a-/2 a a+/2 x
77
Properties of the CDF of a Continuous
Random Variable
F5:

FX(x)

FX(b)

Pr{a<x<b}

FX(a)

x
a b
78
Expected Value of a Continuous Random
Variable
The expected value of a continuous random variable is:

fX(x) fX(x) E{X}

E{X}

x x
79
Variance of a Continuous Random Variable
The variance of a continuous random variable is:

X > Y

fX(x) X fY(y) Y

x y

80
Exponential Random Variable
An exponential random variable measures the inter-arrival time
between two occurrences of an event
E.g., the inter-arrival time of packets arriving in a routers queue

Recall that a Poisson distribution counts the total number of

arrivals

In that sense, the exponential rv is the inter-arrival time

between two Poisson arrivals
1 2 3 4 5 6

Poisson arrivals
81
Exponential Random Variable

inter-arrival time =60

1 2 3 4 60 t

inter-arrival time =20

1 2 3 4 60 t

inter-arrival time =6

1 2 3 4 60 t

82
Exponential Random Variable
Exponential rv is the inter-arrival time between two Poisson
arrivals

Remembering the definition of a Poisson rv, what should be the

average (expected) inter-arrival time for an exponential rv, E?

1 2 3 4 5 6

Poisson arrivals
83
Exponential Random Variable
Exponential rv is the inter-arrival time between two Poisson
arrivals

Remembering the definition of a Poisson rv, what should be the

average (expected) inter-arrival time for an exponential rv, E?
E{E} = 1/

1 2 3 4 5 6

Poisson arrivals
84
Exponential Random Variable
An exponential rv is the inter-arrival time between two Poisson
arrivals
Lets look at a sub-interval (0,t] of the Poisson window
On average, how many arrivals will take place in the (0,t]
subinterval?

t t t t t t

1 2 3 4 5 6

T
85
Exponential Random Variable
On average, how many arrivals will take place in the (0,t]
subinterval?
Arrivals in the (0,t] sub-interval = t/T

In other words, t arrivals will take place in a (0,t] subinterval per

window T

t t t t t t

1 2 3 4 5 6

T
86
Exponential Random Variable
t arrivals will take place in a (0,t] subinterval per window T

t t t t t t

1 2 3 4 5 6

87
Exponential Random Variable
Considering a window of T=1 unit time, the probability density
and cumulative distribution functions of an exponential rv are:

Like the Poisson rv, the Exponential rv is characterized by a

single parameter

1 2 3 4 5 6

Poisson arrivals 88
Exponential Random Variable
fE (x)

x
Image courtesy of Wikipedia article on Exponential Distribution

89
Exponential Random Variable
Exponential rv is a limiting case of a geometric rv

Like the geometric rv, the exponential rv also possess the

memoryless Markov property (see textbook, Pg. 113)

It can be shown that exponential is the only continuous

distribution with the memoryless property

90
Exponential Random Variable
Memoryless property

91
Erlang Distribution
Consider a system that has r different stages each of which is
exponentially distributed
When r phases of this system have identical exponential
distributions then the resulting probability distribution is called
an r-state Erlang
The density and distribution functions of an Erlang rv, R, are:

input Stage 1 Stage 2 Stage r output

EXP() EXP() EXP()
92
Gamma Distribution
If we denote the number of stages r as , and let this
parameter take non-integral values, we get the Gamma, ,
distribution given below:

and are called shape and scale parameters, respectively

93
Gamma Distribution

Image courtesy of Wikipedia article on Gamma Distribution

94
Gamma Distribution
If we plug in =1 in the above function, we obtain the
exponential rv

95
Gamma Distribution
The gamma function has the following well-known and useful
properties:

96
Uniform Distribution
A uniform rv has a uniform distribution over an interval [a, b]

This distribution is also called the Rectangular distribution

97
Uniform Distribution

fU (x)

1/(b-a)

a b x

98
Gaussian or Normal Distribution
Gaussian or Normal Distribution is by far the most famous
continuous distribution

It has the famous bell-shaped curve which can be used to

approximate many other distributions

99
Gaussian Distribution
The Gaussian rv is defined as:

The Gaussian distribution is completely characterized by its mean

and variance and is generally written as

A Gaussian distribution with zero mean and unit variance is

called a standard normal distribution

100
Gaussian Distribution
fN (x)

Image Courtesy of Wikipedia

101
Gaussian Distribution
The CDF of a Gaussian rv does not have a closed form

Tables of CDF values are provided for each(, ) pair

102
Gaussian Distribution
Gaussian distribution does not deviate from its mean by more
than 3 standard deviations 99.7% of the times

Image Courtesy of Wikipedia

103
Quiz#1
While rolling a die twice and counting their sum of outcomes,
What is the probability of getting 7 while one of the outcome is
atleast 2?

104