OSU5509-Unit 01 Session 3: Poison Probability Distribution

SESSION 3
POISSON PROBABILITY DISTRIBUTION

Content

Learning Outcomes

Introduction

3.2 Characteristics of the Poisson Distribution

3.3 Queueing Process
3.4 Poisson Probability Formula
3.5 Poisson Probability Distribution Table
3.6 Expected Value and Standard Deviation
3.6 Calculation of Poisson Values using MS Excel

Summary

Review Questions

Learning Outcomes

At the end of this session, you will be able to:

• Understand the Poisson process and characteristics of the Poisson

distribution.
• Compute the Poisson probabilities using different methods.
• Evaluate the expected value and standard deviation in a Poisson
distribution.

Introduction

Before apply the binomial distribution in determination of the probability there

are several requirements that should be satisfied. If the required conditions are
not satisfied, we can examine other distributions which can be useful in
determination of probabilities of discrete random variable. In this session we
introduce the Poisson distribution.

In binomial distribution, issues or problems should have two outcomes that

success or failure in a single trial. However, there are many situations where
we cannot count both the number of successes and failures. For instance,

34
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

consider that manager of a supermarket needs to know the number of

customers arrive to the supermarket within a specific hour. In this case we can
count the number of customers come to the supermarket but we can’t count
the number do not come. Therefore, in such cases where both successes and
failures cannot count, it is impossible to apply the binomial distribution. Then
we have to look for another application of discrete probability distribution
such as the Poisson distribution.

The Poisson distribution is very common in practical life. The probabilities of

events that take place within a specific period of time or space are considered
as the Poisson distribution. This would include arrivals of telephone calls or
letters or accidents within a specific period of time, number of mistakes per
page, numbers of defective items in each package, etc. This session will
help you in identifying Poisson behavior in practice and to evaluate its
probabilities.

3.1 Characteristics of the Poisson distribution

The Poisson distribution describes the probabilities of events that take place
within a specific period of time or space. This would include arrivals of
telephone calls or accidents or typing mistakes within a specific period of time
or space. Then required to count the number of telephone calls received in an
hour or the number of accidents were happened in an hour or number of
mistakes per page.

The possible outcomes will be 0,1,2,3,4,5,………….like we studied in the

binomial distribution as successes. Now we need to find the probabilities of
these outcomes. As an example, assume that we need to find the probability
that number of telephone calls received in an hour or else to find the
probability that number of typing mistakes per page.

The Poisson distribution can be used if the following assumptions are

satisfied. Here we use the term “interval” to refer to either a time interval or
space, depending on the context of the issue.

1. λ is the average number of occurrences in an interval.

As an example, customers arrive to a shop at the rate of 2 per hour
(λ = 2) or number of telephone calls received in an hour is 8 (λ = 8).
2. The probability of occurrence of an event (x), p, in an interval is the
same (constant) for all sub intervals in the same size.
For example, the probability of number of typing mistakes per page
is same in any page of the report.

35
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

3. Number of occurrences in one interval has no influence on any non-

overlapping intervals.
For example, number of customers arriving to the supermarket
between 10.00 a.m. and 11.00 a.m. has no influence on the number
of customers between1.00 p.m. and 2.00 p.m.
4. The probability of occurrence of more than one event in tiny sub
interval is negligible.

Accordingly, you can observe that the Poisson distribution depends on a single
parameter called, λ.

3.2 Queueing Process

Queues are a common feature in life and form a major component in

management. Some examples are the bus queue, cinema queue, the queue at
the post office. To study the behavior of the queue it is necessary to
understand the pattern of arrivals. That is, for instance how many will arrive
this hour, how many will arrive the next hour and so on. To understand this
pattern, data on number of arrivals per hour was recorded. The data collected
for a period of 500 hours and the results are explained in the table below.

Number of
arrivals per Frequency Probability
hour
0 69 0.138
1 137 0.274
2 130 0.260
3 90 0.180
4 46 0.092
5 20 0.040
6 7 0.014
7 1 0.002
Total ∑ = 500 ∑=1

36
Copyright © 2020, The Open University of Sri Lanka
 OSU5509-Unit 01 Session 3: Poison Probability Distribution

PROBABILITY

0.3-

0.2-

0.1-

0 1 2 3 4 5 6 7 NUMBER OF ARRIVALS

According to the table, out of 500 hours, in 69 cases there were no arrivals, in
137 cases there was one arrival, in 130 cases there were two arrivals and so
on. These figures are expressed in the second column of the table. The third
column evaluates the probability. For instance in 69 out of 500 cases there
were no arrivals. Therefore, the probability of no arrival in 69/500 = 0.138.
In 137 out of the 500 hours there was just one arrival therefore probability of
one arrival is 137/500= 0.274. The curve shown above is a graph of
probability against number of arrivals. This curve explains how the
probabilities of being one or two or three arrivals are distributed. It shows that
the highest probabilities are when the number of arrivals is one or two.
Attempts were then made to fit a mathematical model to the curve. The best
model happen to be

(λ)𝑥 𝑒 −λ
; where λ is the average rate of arrival.
𝑥!

This model is referred to as the Poisson probability distribution. This model in

fact could be mathematically proved as well.

Example 3.1: At a super market, customers arrive at an average rate of 3 per

hour. What is the probability that just two customers will arrive at the next
hour.

Solution:
Average rate of arrival, λ = 3 per hour

(λ)𝑥 𝑒 −λ (3)2 𝑒 −3
Probability of just two arrivals = =
𝑥! 2!

37
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

−3
9 × 2.72
= = 0.226
2×1

There is a 22.6% chance to arrive 2 customers at the next hour.

3.3 Poisson Probability Formula

The probability for any discrete value in the Poisson distribution can be found
by using the following formula.

Poisson Probability Formula

(λ)𝑥 𝑒 −λ
P(x) =
𝑥!

Where x = number of occurrences in an interval

λ = expected number of occurrences
e = 2.718281828 - mathematical constant that is
the base of the natural logarithm

Now you know that the probability of given number of arrivals within a
specified time period has a Poisson distribution. But in fact this event just need
not be arrivals, the event could be accidents in one week or births per day or
number of telephone calls in a specified time. However, the event that we are
talking about should be a random event. The arrivals that take place or an
accident or a birth or a telephone call are all random events. If we ask the
number of times the clock chimes per day. This is not a random event because
the number of times is always 24.

However, this interval always not only be a time period. It is even valid for a
specified period of space. Poisson distributions are also explained by number
of accidents within a stretch of 10 Km or number of defective bulbs in a case
of 500 bulbs or number of defects in 10 meters of material or number of
mistakes in a page.

Therefore in general terms, we can say that the probability of a given number
of events in a specified period of time or space has a Poisson form probability
distribution.

38
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

Though we are given that the average of 3 customers arrives to the

supermarket per hour, we do not expect to find 3 customers arriving in every
hour. The actual arrivals follow a Poisson distribution with the expected value
of λ. The expected value or mean for the Poisson distribution is;
E[x] =  = λ

For an example, at a supermarket, customers arrive at an average of 3

customers per hour. Assume that you need to find the probability that 5
customers arrive to the super market at half an hour period.

Now you know that the probability of a given number of arrivals within an
hour follows a Poisson distribution and it is given by,
(λ)𝑥 𝑒 −λ
P(x) = , where λ (3) is the average number of arrivals per
𝑥!
hour.

However, it is not necessary that the time period be exactly one hour. The
probability of a given number of arrivals within any period of two hours has
also a Poisson distribution and it is the same for any time period.

In this case, first we need to find the average number of customers arrive to
the supermarket at half an hour period.

average number of customers arrive per an hour =3

average number of customers arrive per half an hour(½ hour) = 3/2

As the average, λ = 3 per hour, it should be the same unit which is ½ hour not
the 30 minutes.

now new λ = 3/2

Hence, the probability of 5 customers arrive at ½ hour period is,
(3/2)5 𝑒 −3/2
P(x=5) =
5!

Note: Scientific calculators can be used to find these probabilities. Check the
following links.
https://edu.casio.com/support/qsg/pdf/991EX_570EX/06_CASIO_QuickStartGuide_f
x-991EX_fx-570EX_DISTRIBUTION.pdf
https://www.youtube.com/watch?v=EnRz9vSS-jo

39
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

Example 3.2: Assume that a fast food restaurant expect 3 customers every 5
minute as an average.

a. What is the probability that five customers will enter the restaurant in a
10 minute period?
b. Find the probability that more than 3 customers will enter the
restaurant in a 10 minute period?

Solution:
The process of number of customers arrive to the restaurant according to a
Poisson distribution with the average of 3 (λ = 3) customers per 5 minute
period. The probability of each number of customers arrive to the restaurant
can be calculated by using the Poisson formula given in the above. However,
first we need to find the average per 10 minute period as that is the issue/s
interested.

To determine the interested probability, the average of 3 (λ = 3) customers per

5 minute needs to be converted to the new average for a 10 minute period.

average number of customers arrive to the restaurant in 5 minute = 3

average number of customers arrive to the restaurant in 10 minute = 3×2 = 6

a. the probability of number of customers arriving to the restaurant in a

given time period can be computed using the above formula.
Therefore, the probability that five customers will enter the restaurant
in a 10 minute period is,

(λ)𝑥 𝑒 −λ
P(x = 5) =
𝑥!

(6)5 𝑒 −6
=
5!

= 0.1606

b. The probability that more than 1 customer will enter the restaurant in a
10 minute period is,

P(P( x > 1) = 1 – P(x ≤ 1)

= 1 – [P( x = 0) + P( x = 1)]
= 1- [0.0025 + 0.0149]
= 0.9826

40
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

3.4 Poisson Probability Distribution Table

As presented in the binomial distribution, a probability table exists for the

Poisson distribution too. We learnt that there are probability tables for
different sample sizes in binomial distribution, Poisson tables provide
probabilities for different λ values. The full Poisson table is attached in the
appendix.

Example 3.3: (using the Poisson probability distribution table) An

insurance salesman sells life insurances on the average 2 policies per week.
Using the Poisson distribution, calculate the probabilities that in the next week
he will sell

a. 4 policies
b. 3 or more policies but less than 5 policies

Solution:
Average of the random variable
In this case, the average number of policies sells per week, λ = 2.

Average for the event of interest

As it is interested of calculating the probabilities for the next week, the

corresponding interval is also a week, then average number of policies sells in
the next week is also the same, λ = 2.

Events of interest
a. In this case, he sells 4 policies, it is required to find the probability
value corresponding to the λ= 2 and x=4. Using the Poisson
probability table, under the column heading λ= 2, find the value of x
down the left hand column.

P(x=4) = 0.0902

The probability that selling 4 policies in the next week is 9%.

b. In this case, it is required to find the probability that selling 3 or more

policies
but less than 5 policies. We can state this issue as follows.
P ( 3≤ x < 5) = P(x=3) + P(x=4)
Then using the Poisson table, when λ = 2 corresponding to x=3 and
x=4,
P ( 3≤ x< 5) = 0.1804 + 0.0902
= 0.2706

41
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

The probability that selling 3 or more but less than 5 policies in the
next week is 27.06%.

A part of the Poisson probability table for λ values from 2 to 3 are given
below.

λ
x 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3
0 0.1353 0.1225 0.1108 0.1003 0.0907 0.0821 0.0743 0.0672 0.0608 0.0550 0.0498
1 0.2707 0.2572 0.2438 0.2306 0.2177 0.2052 0.1931 0.1815 0.1703 0.1596 0.1494
2 0.2707 0.2700 0.2681 0.2652 0.2613 0.2565 0.2510 0.2450 0.2384 0.2314 0.2240
3 0.1804 0.1890 0.1966 0.2033 0.2090 0.2138 0.2176 0.2205 0.2225 0.2237 0.2240
4 0.0902 0.0992 0.1082 0.1169 0.1254 0.1336 0.1414 0.1488 0.1557 0.1622 0.1680
5 0.0361 0.0417 0.0476 0.0538 0.0602 0.0668 0.0735 0.0804 0.0872 0.0940 0.1008
6 0.0120 0.0146 0.0174 0.0206 0.0241 0.0278 0.0319 0.0362 0.0407 0.0455 0.0504
7 0.0034 0.0044 0.0055 0.0068 0.0083 0.0099 0.0118 0.0139 0.0163 0.0188 0.0216
8 0.0009 0.0011 0.0015 0.0019 0.0025 0.0031 0.0038 0.0047 0.0057 0.0068 0.0081
9 0.0002 0.0003 0.0004 0.0005 0.0007 0.0009 0.0011 0.0014 0.0018 0.0022 0.0027
10 0.0000 0.0001 0.0001 0.0001 0.0002 0.0002 0.0003 0.0004 0.0005 0.0006 0.0008
11 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0001 0.0001 0.0001 0.0002 0.0002
12 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0001
13 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
14 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

3.5 Expected Value (mean) and Standard Deviation

The mean or average of the Poisson distribution is defined as E[x] = λ. This

mean value, λ is used to specify the Poisson distribution. Therefore we should
know the mean value before calculating the probabilities.

Expected Value(mean)

 = E(x) = λ.

where x = number of occurrences in an interval

λ = expected number of occurrences

42
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

As defined for any discrete probability distribution, the standard deviation of

the Poisson distributed variable is also can be calculated by using the
following equation.

x =√∑[𝑥 − E(𝑥)]2 P(𝑥)

Accordingly, the standard deviation for a Poisson distribution is given by the

following formula.

Standard Deviation of the Poisson Probability

Distribution

 = √𝛌

The mean and variance of the Poisson distribution are the same.

Example 3.4:(using the Poisson probability table) At a busy junction,

vehicles are passing through the junction at an average rate 380 vehicles per
hour in day time. Accordingly,

a. Find the probability that none passes in a given minute.

b. What is the expected number of vehicles passing in five minutes and
its standard deviation?

Solution:
a. The number of vehicles passes through the junction follows a Poisson
distribution with an average rate of 360 vehicles per hour.

It is required to find the probability for a given minute which is equal to 1/60
unit (hour).

average number of vehicles per hour = 360

average number of vehicles per minute,  =360/60 =6

We are interested about the probability that none passes in a given

minute. Using the Poisson table, when λ = 6 and x = 0, the probability,

P(x=0) = 0.0025

43
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

There is a very small probability, 0.25% that none of the vehicles

passes in a given minute at that busy junction.

b. Now we know that,

expected no. of vehicles passing through at each one minute = 6
So, expected no. of vehicles passing at each 5 minute = 6×5 = 30

Standard deviation  = √𝛌
= √30 = 5.4772

The large standard deviation is that in some 5 minute there will be less
no. vehicles and other case there will be more vehicles passing through
the junction.

Example 3.5: (how to answer)

Arrivals of customers to a bank ATM machine during the peak period at the
rate of 90 customers per hour.

a. Find the probability that 5 or more customers will arrive at the ATM
machine in a 6 minute interval during the peak period.
b. What is the most likely number of customers will arrive at the ATM
machine in 10 minute period interval during the peak period.
c. Find the probability that no customers will arrive to the ATM in a
given 30 minute time.

Solution:

Arrivals of customers to a bank follow a Poisson distribution with,

average rate, = 90 customers per hour

= 90/60 = 1.5 customers per minute

Assume that x is the number of customers arrive at the ATM machine

a. P(x ≥ 5) at 6 minute interval

The average rate of arrivals in 6 minute interval , λ = 6 × 1.5 = 9

Therefore,

P(x ≥ 5) = 1- P(x < 5)

= 1- [P(x = 0) + P(x = 1) + P(x = 2) + P(x = 3) + P(x = 4)]

44
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

Using the Poisson distribution table when λ = 9,

P(x ≥ 5) = 1 – [0.0001 +0.0011 +0.0050 +0.0150 +0.0337)]

=0.0549

Then, the probability that 5 or more customers will arrive at the ATM
machine in a 6 minute interval during the peak period is 0.0549.

b. The most likely number of arrivals in 10 minute interval = 10 × 1.5 =

15= 15 customers

c. The average rate of arrivals in 30 minute interval , λ = 30 × 1.5 = 45

P(x = 0) = 0

There is no chance to arrive 0 customers to the ATM in a given 30

minute time.

3.6 Calculation of Poisson Values using MS Excel 2016

The probabilities for number of cars passing through at each five minute,
when the average is 6 are computed using the MS Excel 2016 as given below.
The Poisson distribution is also graphically presented using the MS Excel too.

45
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

Steps in MS Excel 2016

1. Open a blank worksheet.

2. Make 1st column as x and enter values from 0 to 20.
3. Bring the cursor to the cell corresponds to the x=0 in the next column.
4. From Formulas click on fx function. Select the Statistical category. From there select the
POISSON.DIST.
5. Place cell with the required x value and enter the mean. Enter False to select the non-
cumulative probabilities.
6. Copy the function for all x values.
7. Select data and graph using Insert>Column.

Note: refer the link below for more details.

https://www.youtube.com/watch?v=CRroD7Mdv0o

Activity 3.1 – Poisson Distributions

1. At a Post Office, customers arrive at an average rate of 5 per hour. Find

the Probability that in the next hour not more than two arrivals will take
place using the Poisson formula?

2. A factory manager observes that, the average number of machine break

down per day is 4. What is the probability of no break? Use the Poisson
probability table to find the probability.

3. The average number of mistakes made by a typesetter per page is 4.

What is the probability of three typing mistakes in the next page? Use the
MS Excel to find the probability.

Summary

This session introduces the Poisson distribution which is a discrete probability

distribution. The Poisson distribution deals with the situations where the trials
are independent but we can count only the number of successes. This
probability estimate could be evaluated mathematically using the formula.
However, the probability table and software like MS Excel can be used to
calculate the probabilities.

46
Copyright © 2020, The Open University of Sri Lanka
OSU5509-Unit 01 Session 3: Poison Probability Distribution

Review Questions

1. A computer based translation company claims that their system has an

error rate at 4 per 500 words of translation. The company randomly select
a 1500 word passage. If the company claim is accurate,
a. Find the probability that no errors will be found.
b. Find the probability that more than 15 errors will be found
c. If 18 errors are found from the 1500 word passage, what is your
opinion about the company claim?

47
Copyright © 2020, The Open University of Sri Lanka