Queuing Theory

Queue: Systematic arrangement of group of objects or persons who requires services.

Arrival: The process or the way in which the customers are entering into service.
Arrival Pattern: The gap between any two consecutive arrivals if noticed will fit in certain
fashion or pattern. This pattern of arrival time often follows one of the following distributions:
 Poison Distribution
 Exponential Distribution
 Erlang Distribution
 General pattern
Queue example:
1. Passengers waiting at railways station booking counter.
2. Vehicles lined up at the petrol bunk.
3. Machines lined up at service station.
4. Trucks in lined to be unloaded.
5. Airplanes lined up on a runway waiting for the permission to take off.

When demand for serice is more than the capacity of service facility the queue forms.
In mathematical terms we define the situation as :
λ>μ
where , λ -- represents mean rate of arrival
μ -- represents mean rate of service

Cost Associated with Waiting Queue

1. Cost of offering service.
2. Cost induced due to delay in offering service or due to keeping customer waiting.
Formulae:
𝑻𝒐𝒕𝒂𝒍 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒄𝒖𝒔𝒕𝒐𝒎𝒆𝒓𝒔 𝒂𝒓𝒓𝒊𝒗𝒆𝒅
1. Mean rate of arrival ( λ ) :
𝑻𝒐𝒕𝒂𝒍 𝒕𝒊𝒎𝒆 𝒕𝒂𝒌𝒆𝒏

𝑻𝒐𝒕𝒂𝒍 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒄𝒖𝒔𝒕𝒐𝒎𝒆𝒓𝒔 𝒔𝒆𝒓𝒗𝒆𝒅

2. Mean rate of service (μ) :
𝑻𝒐𝒕𝒂𝒍 𝒕𝒊𝒎𝒆 𝒕𝒂𝒌𝒆𝒏

𝟏
3. Mean inter arrival ( ) : It represents the time gap between 2 consecutive arrivals.
λ

𝟏
4. Mean service time ( ): The time taken to serve each customer.
μ

𝛌
5. Steady period : <𝟏
μ

𝛌
6. Busy Period / Traffic Intensity ( 𝝆 ) =
μ

𝟐 𝛒 𝟐
7. Length of a queue ( 𝑳𝒒 ) =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒

- Length of queue represents number of customers waiting to receive services.

𝝀 𝛒
8. Length of a system ( 𝑳𝒔 ) : or
(μ−𝝀) 𝟏−𝛒
- It represents the total number of customers waiting in a queue to receive the service
Plus customers who are actually receiving the services.
𝝀 𝐋𝐪
9. Waiting time in a queue ( 𝑾𝒒 ) : or
μ(μ−𝝀) 𝛌
- It represents the total time spent on an average in a queue to receive the service.
𝟏 𝟏
10. Waiting time in a system ( 𝑾𝒔 ) : or ( 𝐖𝐪 + )
(μ−𝝀) μ
11. Probability of system being empty : 𝛒𝟎 = (𝟏 − 𝛒)
12. Probability of system being busy : 𝛒𝐛 = (𝟏 − 𝛒𝟎 )
𝛌 𝐧
13. Probability of more than 1 unit in system : 𝛒𝐧 =( )
𝛍
 Queuing Model :
1. Kendall-Lee Notation:
- Queuing model can be defined by five most significant features.
- D.G. Kendall in 1953 has noticed 3 of them in the form of (a/b/c) and later A.M. Lee
in 1966 added 2 more in the form (d/e).
- ‘a’ represents arrival pattern
o Arrival pattern cay be :
i. Poison (M)
ii. Exponential (M)
iii. Erlang ( Er )
iv. General Pattern (Gr)
- ‘b’ represents service pattern/distribution
It can be poison , exponential, erlange or general pattern.
- ‘c’ represents number of service channels,
- ‘d’ represents limit of customers
- ‘e’ represents queue discipline or order of service
It can be :
1. LIFO (Last In First Out)
2. FIFO (First In First Out)
3. SIRO (Service In Random Order)
4. Priority service
i. Preemptive priority rule
ii. Non preemptive priority rule
Problems:
1. In a bank, cheques are cashed at a single teller machine/ counter. Customers arrive at
the counter in a poison manner at an average rate of 30 customers per hour. The teller
takes on an average a minute and half to cash a cheque. The service time has been
shown to be exponentially distributed.
i) Calculate the percentage of time the teller is busy.
ii) Calculate the average time customer is expected to wait.
Sol: Kendall’s Notation
(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
 - Mean rate of arrival (λ) that is number of customers on an average arrives at a unit of
time, represented by λ = 30 per hour
- Mean service time that is the time taken to render a service per customer is 1 and
half minute.
1 3
i.e., 1+ = minutes
2 2
1 3
therefore , = min
μ 2
1 3 1 1
=2∗ = hours
μ 60 40

The value of μ = 40
30 3
I. Busy Period = = of an hour
40 4
Therefore , 75% of the time the teller machine is busy.
𝝀
II. Waiting time in a queue = 𝑾𝒒 =
μ(μ−𝝀)
𝟑𝟎 𝟑
= = 𝟒𝟎 of an hour.
40(40−𝟑𝟎)
𝟑 𝟗
Converting into minutes : Wq= ∗ 𝟔𝟎 = 𝟐 min
𝟒𝟎
Therefore, 4.5 min is the waiting time in a queue.
2. The mean arrival rate to the service center id 3 per hour. Assuming the mean service
time is 10 minutes. Poison arrival rate and exponential serving, determine the
following.
i) Busy period/ Utilization factor?
ii) Expected number of units in queue.
iii) Expected time the customer has to wait in a system.
iv) Expected number of units in a system
v) Probability of 2 units in a system.
Sol: Kendall’s Notation
(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
- Mean rate of arrival (λ) that is number of customers on an average arrives at a unit of
time, represented by λ = 3 per hour
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 10 min
μ
1 10 1
= = hours
μ 60 6
The value of μ = 6
𝛌 3 1
I. Busy Period = = = of an hour
μ 6 2
Therefore , 50% of the time the teller machine is busy.
𝟐 𝛒𝟐
II. Length of the queue 𝑳𝒒 =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒
𝟑𝟐
𝑳𝒒 = 𝟔(𝟔−𝟑) = 𝟏𝟐
𝟏
III. Waiting time in a system = 𝑾𝒔 =
(μ−𝝀)
𝟏 𝟏
= = 𝟑 of an hour.
(6−𝟑)
𝟏
Converting into minutes : Ws= 𝟑 ∗ 𝟔𝟎 = 𝟐𝟎 min

Therefore, 4.5 min is the waiting time in a queue.

𝝀 𝟑
IV. Length of system 𝑳𝒔 = = =1
(μ−𝝀) (6−𝟑)
𝛌 𝟐
V. Probability of 2 units in a system = 𝛒𝟐 = ( )
𝛍
𝟑 𝟐 𝟏
= (𝟔) = 𝟒
Therefore, probability of 2 units in a system is 25%.
 3. The customers arrives at a box office window, being managed by single individual,
according to a poison input process with the mean rate of 30 per hour. The time
required to service a customer has an exponential distribution with the mean of 90
sec. Find the average waiting time of a customer. Also determine the average number
of customers in a system and average queue length.

Sol : Kendall’s Notation

(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)

- Mean rate of arrival (λ) that is number of customers on an average arrives at a unit of
time, represented by λ = 30 per hour
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 90 sec
μ
1 90 1
= = hours
μ 60∗60 40
The value of μ = 40

𝝀
I. Waiting time in a queue = 𝑾𝒒 =
μ(μ−𝝀)
𝟑𝟎 𝟑
= = 𝟒𝟎 of an hour.
40(40−𝟑𝟎)
𝟑 𝟗
Converting into minutes: 𝑾𝒒 = ∗ 𝟔𝟎 = 𝟐 min = 4.5 min
𝟒𝟎

II. Length of the System 𝑳𝒔 =

𝝀
(μ−𝝀)

𝟑𝟎 𝟑𝟎
𝑳𝒔 = (𝟒𝟎−𝟑𝟎 )
= 𝟏𝟎
=3
 𝛒 𝟐 𝟐
III. Length of the queue 𝑳𝒒 =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒
𝟑𝟎𝟐
𝑳𝒒 = 𝟒𝟎(𝟒𝟎−𝟑𝟎) = 𝟗𝟒 = 2.25
𝟏
IV. Waiting time in a system = 𝑾𝒔 = ( 𝐖𝐪 + )
μ
𝟑 𝟏
𝑾𝒔 = ( + )= 6 min
40 𝟒𝟎
4. At a certain petrol bunk customers arrive according to poison process with an average
tie of 5 min between arrivals. The service time is exponentially distributed with a
mean time of 2 min. Find average queue length, find if the steady state exist, average
number of customers in a system. If the waiting time in a queue is 4 min a second
pump will be opened. In such a case what should be the arrival rate for the second
petrol bunk.

Sol: Kendall’s Notation

(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
1
- Mean inter arrival rate is given that is = 5 min = 12 per hour
λ
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 2 min
μ
1 2 1
= = hours
μ 60 30
The value of μ = 30

𝛒 𝟐 𝟐
I. Length of the queue 𝑳𝒒 =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒
𝟏𝟐𝟐 𝟒
𝑳𝒒 = 𝟑𝟎(𝟑𝟎−𝟏𝟐) = 𝟏𝟓 = 0.267
 II. Length of the System 𝑳𝒔 =
𝝀
(μ−𝝀)

𝟏𝟐 𝟏𝟐 𝟐
𝑳𝒔 = (𝟑𝟎−𝟏𝟐 )
= =
𝟏𝟖 𝟑
𝟏
III. Waiting time in a queue = 𝑾𝒒 =4 min = of an hour
𝟏𝟓
𝐋𝐪 𝛌′
=
𝛌 𝛍(𝛍−𝛌′ )
𝟏 𝛌′
=
𝟏𝟓 𝟑𝟎(𝟑𝟎−𝛌′ )

𝛌′ = 20 per hour

IV. Steady state existency :

𝛌
<𝟏
μ

𝟏𝟐
<𝟏
30
0.4 < 1
Therefore, steady state exist

5. In a bank, every 15 min 1 customer arrives for cashing the cheque. The staff in the only
payment counter takes 10 min to serve a customer on an average. Write down and
Kendall-Lee notation and find out
i) Average queue length
ii) Waiting time in a system
iii) Increase in arrival rate to justify 2nd counter where the waiting time of a customer
is 15 min, the management will add one more counter.
Sol : Kendall’s Notation
(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
 1
- Mean inter arrival rate is given that is = 15 min = 4 per hour
λ
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 10 min
μ
1 10 1
= = hours
μ 60 6
The value of μ = 6

𝛒 𝟐 𝟐
I. Length of the queue 𝑳𝒒 =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒
𝟒𝟐
𝑳𝒒 = 𝟔(𝟔−𝟒) = 𝟒𝟑

𝟏
II. Waiting time in a system = 𝑾𝒔 =
(μ−𝝀)
𝟏 𝟏
= = 𝟐 of an hour.
(6−𝟒)
𝟏𝟓
III. Waiting time in a queue = 𝑾𝒒 =15min = of an hour
𝟔𝟎
𝐋𝐪 𝛌′
=
𝛌 𝛍(𝛍−𝛌′ )
𝟏𝟓 𝛌′
=
𝟔𝟎 𝟔(𝟔−𝛌′ )

𝛌′ = 3.6 per hour

6. In a bank, with single sever, there are 2 chairs for waiting customers. On an average
one customer arrives every 10 min and each customer takes 5 min for getting served.
I. Find out the probability that an arrival will get a chair to sit down.
II. Probability that an arrival will have to stand.
III. Expected waiting time of a customer.
Sol : Kendall’s Notation
(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
1
- Mean inter arrival rate is given that is = 10 min = 6 per hour
λ
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 5 min
μ
1 5 1
= = hours
μ 60 12

The value of μ = 12

I. Probability of system being empty 𝛒𝟎 = (𝟏 − 𝛒)

λ
= (1 - μ )
6 1
= (1 - 12) =2

II. Probability of system being busy 𝛒𝐛 = (𝟏 − 𝛒𝟎 )

1
= (1 - )
2
𝟏
=𝟐

𝟏
IV. Waiting time in a system = 𝑾𝒔 =
(μ−𝝀)
𝟏
= = of an hour =10 min.
𝟔
 7. Arrival at a telephone both as considered to be poison with an average time of 10 min
between 1st arrivals to next arrival. The length of telephone call is assumed to be
distributed exponentially with the mean 3 min.
I. What is the probability that the person arriving at both will have to wait?
II. The department will install 2nd booth when convinced that an arrival would
expect waiting for at least 3 mins per call. By how much the flow of an arrival
increase in order to justify a second booth.
III. What is the average length of queue that forms from time to time?
Sol: Kendall’s Notation
(a/b/c) (d/e)
(M/M/1) (∞ , FIFO)
1
- Mean inter arrival rate is given that is = 10 min = 6 per hour
λ
- Mean service time that is the time taken to render a service per customer is 10
minutes.
1
therefore , = 3 min
μ
1 3 1
= = hours
μ 60 20
The value of μ = 20

Probability of system being empty 𝛒𝟎 = (𝟏 − 𝛒)

λ
= (1 - μ )
6 14
= (1 - 20) =20

I. Probability of system being busy 𝛒𝐛 = (𝟏 − 𝛒𝟎 )

14
= (1 - )
20
𝟔
= 𝟐𝟎
 𝟑
II. Waiting time in a queue = 𝑾𝒒 =3 min = of an hour
𝟔𝟎
𝐋𝐪 𝛌′
=
𝛌 𝛍(𝛍−𝛌′ )
𝟑 𝛌′
=
𝟔𝟎 𝟔(𝟔−𝛌′ )

𝛌′ = 10 per hour
𝟐 𝛒 𝟐
III. Length of the queue 𝑳𝒒 =
𝝀 or
μ(μ−𝝀) 𝟏−𝛒
𝟔𝟐 𝟗
𝑳𝒒 = 𝟐𝟎(𝟐𝟎−𝟔) = 𝟕𝟎 = 0.1285