CE 207

Applied Mathematics for Engineers

Course Teacher: Dr. Md. Hadiuzzaman
Room No. 544, Civil Building
(Lecture 5)
 Poisson Distribution

• Poisson distribution is a special limiting case of the binomial

distribution where: n→, p→0. Binomial calculations then
become impractical.

Simeon D. Poisson (1781-1840)

Quote from Poisson:

Life is good for only two things: to
study mathematics & to teach it
 Poisson Distribution (contd.)
• Physical meaning: a large number of trials (n going to infinity),
and the probability of the event occurring by itself is pretty
small (p approaching zero).
• The combined effect (np) is finite.
• Substituting λ/n for p, and letting n→, the binomial
distribution becomes the Poisson distribution:
• X which takes one of the values 0,1,2,… is Poisson random
variable with parameter , if its probability mass function is

i
PX = i = e −
, i = 0,1,..........
i!
Poisson Distribution (Contd.)
0.1 0.5

1 1
0.8 0.8
0.6
0.6
0.4 0.4
0.2
0.2
0
0
0

10

12

10

12
1 6

1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0

10

12
0

10

12

4
 Poisson Distribution (contd.)
• Examples:

▪ Number of wrong telephone numbers that are

dialed in a day.
▪ Number of misprints in a book.
▪ Number of accidents on a highway in a given
period of time.
▪ Number of earthquakes in a given period of time.
 Poisson Distribution (contd.)
• Expectation of the Poisson distribution:
E(X) = 
The only thing we have to know to specify the
Poisson distribution is the mean number of
occurrences.

• Variance of the Poisson distribution:

Var(X) = 
 Poisson Distribution (contd.)
Poisson distribution applies to events that are rare in time .
If X represents the number of events in time t, and  represents
the rate of occurrence of events, then the probability for X=i in
time t is given by:

PX = i = e −t
( t )i
, i = 0,1,..........
i!

Here t (used instead of ) is the parameter of the Poisson

random variable X.
Expected no. of events in time t, E[X] = t
 Poisson Distribution (contd.)

Problem 1:
Rutherford, Geiger, and Bateman (1910) counted the number of
-particles emitted by a film of polonium in 2608 successive
time intervals of one-eighth of a minute.

Data is given in next page.

Does the data follow Poisson distribution?
Determine values of n and p.
 Poisson Distribution (contd.)
Problem 1 (contd.) No. of -particles in No. of times Total No. of -
each time interval observed particles
• Here, n= total no. of time intervals
0 57 0
• Let X be no. of particles per interval 1 203 203

• Calculation of  2
3
383
525
766
1575
 = Mean no. of particles per interval 4 532 2128
5 408 2040
= 10097/2608 = 3.87 6 273 1638
7 139 973
• Probability for X=i in each interval: 8 45 360

p = PX = i = e −3.87 (3.87 )i 9 27 243

i! 10 10 100
11 4 44
• Expected value of no. of intervals 12 0 0
13 1 13
where X=i is given as :
( )
14 1 14
i
3.87 >14 0 0
np = 2608  e −3.87 2608 10097
i!
 Poisson Distribution (contd.)
No. -particles Observed Expected
0 57 54
Problem 1 (contd.): 1 203 210
2 383 407
Compare actual (observed) 3 525 525
data with expected value 4 532 508
calculated assuming Poisson 5 408 394
6 273 255
distribution
7 139 140
8 45 68
9 27 29
10 10 11
11 4 4
12 0 1
13 1 1
14 1 1
Over 14 0 0
Total 2608 2608
 Poisson Distribution (contd.)

Problem 2:
Average number of claims handled daily by insurance company
is 5.
What proportion of days have less than 3 claims?
Probability that there will be 4 claims in exactly 3 days of the
next 5 days?

Let X represent no. of claims handled daily. It is a Poisson

random variable since probability of claim on a given day is
really small.

Given that, E(X)=5, we have  = 5.

 Poisson Distribution (contd.)

Problem 2 (contd.):
The probability that there will be less than 3 claims on a day is
given by:
P  X  3 = P  X = 0 + P  X = 1 + P  X = 2
( ) ( ) ( )
0 1 2
5 5 5
= e −5 + e −5 + e −5 = 0.125
0! 1! 2!

In other words, 12.5% of days will have less than 3 claims.

Probability of having 4 claims on a day is:

PX = 4 = e −5
(5 )4
 0.175
4!
 Poisson Distribution (contd.)

Problem 2 (contd.):
The probability that there will be 4 claims on 3 days out of 5 days
can be calculated applying binomial distribution. Here, n=5 and
p=0.175.

5 4 3
C3 p (1 − p ) = (0.175)3 (1 − 0.175)2
5 3 2

3  2 1
 0.0367
 Poisson Distribution (contd.)
SUM OF POISSON RANDOM VARIABLES

If X and Y are independent Poisson random variables with

parameters  and , then the random variable Z defined by
Z=X+Y, is also a Poisson random variable with parameter +.

The above proposition is obvious considering that a Poisson

Distribution refers to counts of events scattered at random in
space or time.
 Poisson Distribution (contd.)
Problem 3:
Both cars and trucks pass an observer at random points in time.
On average there are 300 cars per hour. The average time
Between lorries is 5 min. Determine the probability that exactly
6 vehicles (car+lorry) pass the observer in 1 min.

Assume Poisson Distribution

Mean rate for trucks = 60/5 = 12 per hour.
Mean rate for vehicles = 300+12 = 312 per hour
= 5.2 per min. 
Probability = (5.2) e −5.2
6
= 0.151
6!
 Moment Generating Functions of Random Variable

Moment generating function (t) of random variable X is defined

as:

 (t ) = E e =  
tX
etx f (x ) dx if X is continuous
−

=  etx p(x ) if X is discrete

x

d
 ( )
tX  d tX 
 (t ) = E e = E  e  = E XetX ( )  
dt  dt 
Putting t = 0,  (0) = E X 
Moment Generating Functions (contd.)

Similarly,  (t ) =  (t ) =

d
dt
d
dt
(
E Xe tX )
 
= E X 2 e tX
Putting t = 0,  (0 ) = E X  2

In general, the nth derivative of  (t) evaluated at t = 0

equals E X n 
For independent random variables X and Y with moment
generating functions X (t ) and Y (t ),
 X +Y (t ) = E et ( X +Y )  = E etX E etY  =  X (t )Y (t )
 Poisson Distribution Function
If X is Poisson with mean , then:
PX = i + 1 e −  i +1 (i + 1)! 
= =
PX = i e  (i )!
− i
i +1
PX = 0 = e − 
PX = 1 = PX = 0

PX = 2 = PX = 1
2
......................................

PX = i + 1 = PX = i
i +1
 Uniform Distribution
A random variable X is uniformly distributed over the interval
[,] if its probability density function is given by:
f (x ) =
1
if   x  
 −
=0 otherwise
Uniform distribution arises in practice when we suppose a
certain random variable is equally likely to be near any
particular value in the interval [,].
Let [a,b] be a subinterval of [,]. Then
Pa  X  b = 
bdx
a  −

b−a
=
 −
Uniform Distribution (contd.)
f (x ) =
1 1
= if 0  x  100
100 − 0 100
=0 otherwise

FX ( x ) =
x
if 0  x  100
100
 Uniform Distribution (contd.)
Problem 1:
Buses arrive at a specified stop at 15 min. intervals starting at 7am.
That is, they arrive at 7, 7:15, 7:30, 7:45 am etc. If a passenger arrives
at a stop at a time uniformly distributed between 7 & 7:30 am, find
the probability that he waits (a) less than 5 min. (b) at least 12 min.

Let X denote the time in min. past 7 am when passenger comes. He will
have to wait less than 5 min. if he comes between 7:10-7:15 am or
between 7:25-7:30 am. Since X is a uniform random variable in the
interval (0,30), probability for (a) is given by:
15 − 10 30 − 25 1
P10  X  15+ P25  X  30 = + =
30 − 0 30 − 0 3
Similarly, probability for (b) is given by:
3 − 0 18 − 15 1
P0  X  3+ P15  X  18 = + =
30 − 0 30 − 0 5
 Uniform Distribution (contd.)
The mean [,] of a uniform random variable X is

 x   +
E X  =  xf ( x )dx =  
 
dx =
−   −
  2

The variance is given as

Var ( X ) = E X − (E X ) =
2 2 (  −  )2

12