Queuing Theory: A Case Study to Improve the Quality Services of a

Restaurant
Lakhan Patidar1*, Trilok Singh Bisoniya2, Aditya Abhishek3, Pulak Kamar Ray4
Department of Mechanical Engineering, SIRT-E, Bhopal-462041(India)
Email: [email protected], [email protected],[email protected],
[email protected]

Abstract Queuing theory is the mathematical study of waiting lines, or queues. In queuing theory a
model is constructed so that queue lengths and waiting times can be predicted. The common problem
arises in almost every famous restaurant is that they lose their customers due to a long wait on the
line. This shows a need of a numerical model for the restaurant management to understand the
situation better. This paper aims to show that queuing theory satisfies the model when tested with a
real-case scenario. Authors obtained the data from a restaurant Bapu Ki Kutiya in Bhopal, M.P. in
order to derive the arrival rate, service rate, utilization rate, waiting time in queue and the probability
of potential customers to balk. The collected data is analyzed by using Littles Theorem and M/M/1
queuing model. The arrival rate at Bapu Ki Kutiya, Bhopal during its busiest period of the day is
3.25 customers per minute (cpm) while the service rate is 3.27 cpm during our study period. The
average number of customers in the restaurant is 210 and the utilization period is 0.993.
Keywords: Queuing theory, Littles Theorem, Kendalls notation, Waiting Lines.
I. INTRODUCTION
There are several determining factors for a

considered especially when the restaurant has

restaurant to be considered a good or a bad one.

already succeeded in attracting customers. This

Taste, cleanliness, the restaurant layout and

factor is the customers queuing time. Once we

settings are some of the most important factors.

are being served, our transaction with the service

These factors, when managed carefully, will be

organization may be efficient, courteous and

able to attract plenty of customers. However,

complete: but the bitter taste of how long it took

there is also another factor that needs to be

to get attention pollutes the overall judgments

that we make about the quality of service. In a

lines in Bapu Ki Kutiya, Restaurant at Bhopal,

waiting line system, managers must decide what

M.P India., and The restaurant provides 40 tables

level of service to offer. A low level of service

out of which some have 4 chairs and some have

may be inexpensive, at least in the short run, but

6 chairs. There are 25 waiters working at any

may incur high costs of customer dissatisfaction,

one time. On a daily basis it serves over 800

such as lost future business and actual

customers during weekdays and over 1300

processing costs of complaints. Queuing theory

customers during weekend. This paper seeks to

is the study of queue or waiting lines. Some of

illustrate the usefulness of applying queuing

the analysis that can be derived using queuing

theory in a real case situation.

theory include the expected waiting time in the

II. QUEUING THEORY

queue, the average time in the system, the

expected queue length, the expected number of

In 1908, Copenhagen Telephone Company

customers served at one time, the probability of

requested Agner K. Erlang to work on the

balking customers, as well as the probability of

holding times in a telephone switch. He

the system to be in certain states, such as empty

identified

or full. Waiting lines are a common sight in

conversations and telephone holding time fit into

restaurants especially during lunch and dinner

Poisson

time. Hence, queuing theory is suitable to be

distributed. This was the beginning of the study

applied in a restaurant setting since it has an

of queuing theory. In this section, we will discuss

associated

two common concepts in queuing theory.

queue

or

waiting

line

where

customers who cannot be served immediately

that

the

distribution

number

and

of

telephone

exponentially

A. Littles Theorem

have to queue (wait) for service. Researchers

have previously used queuing theory to model
the restaurant operation [2], reduce cycle time in
a busy fast food restaurant [3], as well as to
increase throughput and efficiency [5]. This
paper uses queuing theory to study the waiting

Littles theorem [6] describes the relationship

between throughput rate (i.e. arrival and service
rate), cycle time and work in process (i.e.
number of customers/jobs in the system). This
relationship has been shown to be valid for a

wide class of queuing models. The theorem

the performance change when the system is

states that the expected number of customers (N)

modified.

for a system in steady state can be determined

B.

using the following equation:

The principle actors in a queuing situation are the

L=T

(1)

Queuing Models and Kendalls Notation

customer and the server. On arrival at a service

Here, is the average customer arrival rate and T

facility, they can start service immediately or

is the average service time for a customer.

wait in a queue if the facility is busy. From the

Consider the example of a restaurant where the

standpoint of analyzing queues, the arrival of

customers arrival rate () doubles but the

customers is represented by the inter arrival time

customers still spend the same amount of time in

between successive customers, and the service is

the restaurant (T). These facts will double the

described by the service time per customer. The

number of customers in the restaurant (L).By the

queuing behavior of customers plays a role in

same logic, if the customer arrival rate ()

waiting line analysis.Human customers may

remains the same but the customers service time

jockey from one queue to another in the hope of

doubles this will also double the total number of

reducing waiting time. They may also balk from

customers in the restaurant. This indicates that in

joining a queue all together because of

order to control the three variables, managerial

anticipated long delay, or they may renege from

decisions are only required for any two of the

a queue because they have been waiting too long.

three variables. Three fundamental relationships

In

can be derived from Littles theorem [5]:

characterized by the following factors:

L increases if or T increases.

most

cases,

queuing

models

can

Arrival time distribution: Inter-arrival times

most commonly fall into one of the following

increases if L increases or T decreases.

T increases if L increases or decreases.

Rust [8] said that the Littles theorem can be

useful in quantifying the maximum achievable
operational improvements and also to estimate

distribution patterns: a Poisson distribution, a

Deterministic

distribution, or a General

distribution. However, inter-arrival times are

most often assumed to be independent and
memory less, which is the attributes of a

be

Poisson distribution.

System capacity (optional): The maximum

Service time distribution: The service time

number of customers in a system can be from

distribution can be constant, exponential,

1 - infinity.

hyper-exponential,

or

Kendall, in 1953, proposed a notation system

general. The service time is independent of the

to represent the six characteristics discussed

inter-arrival time.

above. The notation of a queue is written as:

hypo-exponential

Number of servers: The queuing calculations

A/B/P/Q/R/Z

change depends on whether there is a single

A describes the distribution type of the inter

server or multiple servers for the queue. A

arrival times, B describes the distribution type

single server queue has one server for the

of the service times, P describes the number of

queue. This is the situation normally found in

servers in the system, Q (optional) describes

a Book store where there is a line for each

the maximum length of the queue, R

cashier. A multiple server queue corresponds

(optional) describes the size of the system

to the situation in a bank in which a single line

population and Z (optional) describes the

waits for the first of several tellers to become

queuing discipline.

available.

III.BAPU KI KUTIYAQUEUING MODEL

Queue Lengths (optional): The queue

in

a system can be modeled as having infinite or

The daily number of visitors was obtained

finite

from the restaurant itself. The restaurant has

queue

length.

This

includes

the

customers waiting in the queue.

been recording the data as part of its end of

Queuing discipline (optional): There are

day

several possibilities in terms of the sequence

restaurant manager to find out about the

of customers to be served such as FIFO (First

capacity of the restaurant, the number of

In First Out, i.e. in order of arrival), random

waiters as well as the number of chefs in the

order, LIFO (Last In First Out, i.e. the last one

restaurant. Based on the interview with the

to come will be the first to be served), SIRO

restaurant manager, we concluded that the

(Service in Random Order).

queuing

routine.

We

model

also

that

interviewed

best

illustrate

the

the

operation of Bapu Ki Kutiya is M/M/1.

This means that the arrival and service time

Pn: The probability of having n customers in the

restaurant.

(Poisson

Pn=Po n=(1 ) n .... (3)

process). The restaurant system consists of

L: average number of customers dining in the

only one server. In our observation the

restaurant.

are

exponentially

distributed

restaurant has several waitresses but in the

L=

actual waiting queue, they only have one chef

=
..
1

to serve all of the customers. Figure 1

(4)

illustrates the M/M/1 queuing model.

Lq: average number of customers in the queue.

2
L q=
1
............. (5)
W: average time spent in BKK including the
waiting time.
Fig. 1: M/M/1 Queuing Model

W=

For the analysis of the Bapu Ki Kutiya M/M/1

Queuing Model, the following variables will be

(6)
Wq: average waiting time in the queue.

investigated [6]:
W q=
: The mean customers arrival rate

Lq

...

(7)
: The mean service rate
: /: utilization factor
Probability of zero customers in the restaurant
(Po) is given by
Po=1 ......
(2)

IV. RESULT AND DISCUSSION

The one week analysis is shown by graph below
analyzed by us at dinner slot.

much when compared to the theoretical waiting

time as shown below
W q=

Lq 3 5 c u s t omers
=
=10.7 6 minutes

3 . 25 c p m

Next, calculate the average number of people in

the restaurant using the above calculated values,
Fig.2: One Week Analysis of Number of
Customers arriving in restaurant

L=3 . 25 cpm 6 5 mi n u tes=211 c ustomers

As can be seen in Figure 2, the number of

After

customers on Saturdays and Sundays are double.

customers in the restaurant Bapu Ki Kutiya ,

The busiest period for the restaurant is on

derive the service rate as:

weekend during dinner time. Hence focus our

analysis in this time window. Authors analyzed
the restaurant between 19to 22 hours.
A. Calculation

Our teams conducted the research at dinner time.

calculating

the

average

number

of

( 1+ L) 3. 2 5(1+211)
=
=3. 27 c pm()
L
2 11

Now, calculate Traffic Intensity or utilization

factor
3 .2 5
= =
=0. 993
3 .2 7

There are on average 585people are coming to

With the high utilization rate of 0.993 during

the restaurant in 3 hours time window of dinner

dinner time the probability of zero customers in

time. From this we can derive the arrival rate as:

the restaurant or probability that system is idle

585
=
=3. 25 c ustomer per minute(cpm)
1 80
We also found out from observation and
discussion with manager that each customer
spends 65 minutes on average in the restaurant
(W), the queue length is around 35 people (Lq)
on average and the waiting time is around 12
minutes. It can be shown using (7) that the
observed actual waiting time does not differ by

can be calculated by (2)

PO=1=0.007
The generic formula that can be used to calculate
the probability of having n customer in the
restaurant is as follows:
Pn=( 1 ) n=( 10. 9 93 ) 0. 9 93n =( 0.007 ) 0.9 93 n
Assume that potential customers will start to balk
when they see more than 30 people are already

queuing for the restaurant and the maximum

only half of the number of visitors on weekends.

queue length that a potential customer can

In addition, the number of waiters or waitresses

tolerate is 40 people. As the capacity of the

remains the same regardless whether it is peak

restaurant when fully occupied is 200 people,

hours or off-peak hours. In case the customers

can calculate the probability of 30 people in the

waiting time is lower or in other words we

queue as the probability when there are 240

waited for less than 15 minutes, the number of

people in the system (i.e. 210 in the restaurant

customers that are able to be served per minute

and 30 or more queuing) as follows:

will increase. When the service rate is higher the

Probability of customers going away = P (more

utilization will be lower, which makes the

than 30 people in the queue) = P (more than 240)

probability of the customers going away

people in the restaurant)

decreases.

0. 9 93

n
(.00 7 )

C. Benefits
This research can help Bapu Ki Kutiya to
=4.31%

2 40

Pn=

anticipating if there are many customers in the

n=2 1 1
2 40

P2 112 40=

increase their QOS (Quality of Service), by

n=2 1 1

B. Analysis
The utilization is directly proportional with the
mean number of customers. It means that the
mean number of customers will increase as the
utilization increases. The utilization rate at the
restaurant is very high at 0.993. This, however, is
only the utilization rate during lunch and dinner
time on Saturdays and Sundays. On weekday, the
utilization rate is almost half of it. This is
because the number of visitors on weekdays is

queue. The result of this paper work may

become the reference to analyze the current
system and improve the next system. Because
the restaurant can now estimate of how many
customers will wait in the queue and the number
of customers that will go away each day.
By anticipating the huge number of
customers coming and going in a day, the
restaurant can set a target profit that should be
achieved daily and The formulas that were used
during the completion of the research is
applicable for future research and also could be

use to develop more complex theories.

V.

complexity so that the model can mirror the

CONCLUSION

actual operation of the restaurant more closely.

This study has discussed the application of

queuing theory of Bapu Ki Kutiya Restaurant.
Here authors focused on two particularly
common decision variables as a medium for
introducing and illustrating all the concepts.
From the result authors obtained that the rate at
which customers arrive in the queuing system is
3.25 customers per minute and service rate is
3.27 cpm and utilization rate is 0.993. This
theory is also applicable for the restaurant if they
want to calculate all the data daily. It can be

This study gives a generalized guarantee to

stabilize the system from the problems arisen
like customers balking, reneging, jockeying and
collusion or delay in services by present way of
working of restaurant. In todays world of
accelerating

advancement

in

computer

technology, it will be fruitful to restaurant

manager to install a computer for the proper
control of service facilities and to keep the
previous record so as to make the forecasting
better over good in order to excel in the field.

concluded that the arrival rate will be lesser and

REFERENCES:

the service rate will be greater if it is on

[1] A. K. Kharwat, Computer Simulation: an

weekdays since the average number of customers

Important Tool in the Fast-Food Industry,

is less as compared to those on weekends. The

constraints that were faced for the completion of

Proceedings

of

Winter

Simulation

Conference, IEEE Press (Dec. 1991), pp.

this research were the inaccuracy of result since

some of the data that was just based on
assumption or approximation. Authors hope that

811-815.
[2] M.Laguna and J. Marklund, Business Process

this research can contribute to the betterment of

Modeling, Simulation and Design. Pearson

Bapu Ki Kutiya restaurant in terms of its way

Prentice Hall (2005) ISBN 0-13-091519-X.

of dealing with customers.

[3] J. D. C. Little, A Proof for the Queuing

Future Outcomes

Formula: L= W, Operations Research

This study will develop a simulation model for

the restaurant which will be able to confirm the

(1961) vol. 9(3), pp. 383-387.

[4] K. Rust, Using Littles Law to Estimate

results of the analytical model. In addition, a

Cycle Time and Cost, Proceedings of Winter

simulation

Simulation Conference, IEEE Press (2008).

model

allows

adding

more

[5] Sonali Yadav et.al, Optimization of Quality

Academic Press (2007). ISBN0-12-370523 1.

of Service of Restaurant, Proceedings of

[7] D.M. Brann and B.C. Kulick, Simulation of

NCRTES, PIE, Academic Press (2012).

[6] T. Altiok and B. Melamed, Simulation
Modeling

and Analysis

with

ARENA,

restaurant operations using the Restaurant

Modeling Studio, Proceedings of the Winter
Simulation Conference, IEEE Press (2002),
pp. 1448-145.