ADDIS ABABA UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

SCHOOL OF COMMERCE
DEPARTMENT OF LOGISTICS AND SUPPLY CHAIN MANAGEMENT

PROGRAM: Extension
COURSE: MODELING AND IT APPLICATION FOR DECISION MAKING IN SUPPLY CHAIN (LSCM 611-3)

ANALYSIS ON: Queueing Model application (in case of ministry of revenue medium taxpayer
branch office)

Reviewed by: Name ID No.

1. Abel Solomon ----------------------GSE/6836/13
2. Abebe Deressa----------------------GSE/3171/12
3. Amare Seifu--------------------------GSE/5823/13
4. Christian Teshome------------------GSE/2410/13
5. Dereje Kuma--------------------------GSE/7646/13
Section: One

Submitted to instructor: - Dr. Matiwos E.(Ph.D)(Associate. Prof)

Date: - June 20, 2022

Acknowledgement

We would like to express our special thanks and gratitude to instructor Dr Matiwos
Ensermu (Ass. Prof) for providing us this assignment since it made us to add one
additional knowledge on the course of modeling and IT application for decision
making in supply chain by scientifically apprising real world scenarios which
significantly will help us to make sound decisions in our professional life by
scientifically analyzing issues using mathematics in addition to aiding us to be
ready for the near future academic thesis preparation.
Table of Contents

1. Introduction 3

1.1.Observation and problem definition 3

2. Literature Review 6

2.1.Theoretical Framework 6

2.1.2.Elements of waiting line analysis 7

3. Model Construction 11

4. Acquiring input data 12

5. Model Solution 14

5.1. Operating characteristics of the system with four (4) servers 14

5.2. Operating characteristics of the system with five (5) servers 16

6. Discussion 18

7. Conclusion 19

8. Recommendation 19

9. References 20

1
Abstract
Everyone experiences queues in every day to day life seeking service from numerous organizations,
including banks, airlines, telecommunications, and in transports. And this analysis was conducted to on
the observation of queue experienced at the last due dates of monthly tax duty payment days in the
medium tax payers branch office located in kazenchis area which have six serve which four of them
designated to serve customers from private firms and the rest two for serving governmental agencies which
would be idle most of the time. Given the huge number of private taxpayers queue would happen in the
four servers line. On this ground, we analyzed the operating characteristics of the current system with four
servers by collecting data using observational survey method and analyzed through queueing model. On
this ground the current system with four servers operating characteristic analysis implies that the average
time a customer spends in waiting line was 2.2 minute while it takes a customer a total mean of 7.2 minute
to get the service and exit the system. Based on this fact, in order to improve the service performance of
the firm during this peak season, we have considered to use one server from governmental agencies
service desk to provide service for private firm’s taxpayers. Based on this fact, the number of servers who
server private firm was increased from four server to five server, which had decreased the average time a
customer spends in queue 2.2 minute to half minute (0.51 mini) while at the same time reduced the average
time a customer spends in the system from 7.2 minute to 5.5 minute. On this ground, when the number of
servers increase by taking other variables constant and without making an additional investment, the
service capability of the system had improved. Hence, we recommend the company to use one of the idle
servers available for serving governmental agency taxpayers to serve taxpayers from private companies to
cope up with the queue issue on those times.

2
 1. Introduction
All of us have experienced the annoyance of having to wait in line. Unfortunately, this
phenomenon continues to be common in congested, urbanized, “high-tech” communities. We
wait in line in our cars in traffic jams or at toll booths; we wait on hold for an operator to pick up
our telephone calls; we wait in line at supermarkets to check out; we wait in line at fast-food
restaurants; and we wait in line at stores and post offices. We, as customers, do not generally like
these waits, and the managers of the establishments at which we wait also do not like us to wait,
since it may cost them business. Why then is there waiting? The answer is simple: There is more
demand for service than there is facility for service available. Why is this so? There may be
many reasons; for example, there may be a shortage of available servers, it may be infeasible
economically fora business to provide the level of service necessary to prevent waiting, or there
may be a space limit to the amount of service that can be provided. Generally, these limitations
can be removed with the expenditure of capital, and to know how much service should then be
made available, one would need to know answers to such questions as “How long must a
customer wait?” and “How many people will form in the line?” Queueing theory attempts to
answer these questions through detailed mathematical analysis John F. S, James M. T, Donald G.
& Carl M.H (2018).
There are many valuable applications of queueing theory including traffic flow (vehicles,
aircraft, people, communications), scheduling (patients in hospitals, jobson machines, programs
on a computer), and facility design (banks, post offices, amusement parks, fast-food restaurants).
Most real problems do not correspond exactly to a mathematical model, and increasing attention
is being paid to complex computational analysis, approximate solutions, simulation, and
sensitivity analyses. Customers arrive, wait for service, receive service, and then leave the
system. Some customers may leave without receiving service, perhaps because they grow tired
of waiting in line or perhaps because there is no room to enter the service facility in the first
place John F. S, James M. T, Donald G. & Carl M.H (2018).

3
1.1 Observation and problem definition
As we have seen the application of queuing model on real life activities we face. The first step in
the management science process begins with identification of a problem that exists in the system
(organization) or subject organization. On this ground, we have observed the Ethiopian ministry
of revenue medium taxpayers branch office system located in Addis Ababa around Kazanchis
environment. In doing so, we have observed that the system is characterized by multiple server
system and subjected to waiting line issue during last 5 days of monthly tax payment days. To
elaborate the issue in depth, the ministry of revenue medium taxpayers’ branch office collects
taxes from medium level taxpayers. And the taxes to be collected include monthly and annual tax
duties. From this, the tax authority will allow taxpayers to pay their monthly tax duties of last
month until the end of the following month. To illustrate, the April month tax duty would have to
be paid until the end of May otherwise punishment would follow. In such cases we have
observed that, during the last 6 days of tax payment periods long waiting line problem would be
experienced by taxpayers. The organization has around six servers that serve the customers
parallels but from the available servers 2 of them only give service to government office
taxpayers and the rest 4 servers serve private companies taxpayers. In this case, because of the
vast number of private taxpayers the four servers would get busy, and our focus would be on
them. For this reason, when we say customers, we are referring to private company taxpayers.
Based on this fact, due to the long waiting line, customers either forced to wait in the line to pay
the monthly duties or decide to leave the waiting line in anticipation of paying the tax with the
respective punishment amount in the beginning of the following month where waiting line will
not be experienced much. This negatively impacts both parties. In the first dimension, the
organization would be negatively affected by this system if the customer decided to quit the line
since the anticipated tax to be collected wouldn’t be achieved. And this would impact go further
to the economy since physical policy is followed in our country in which budgets are estimated
in anticipation of collected taxes and other revenue sources. So the deviation in the collected tax
would result in deviation in the amount of fund planned to be used in economy sectors and
public infrastructures with the constraint of time of collection of the taxes even if the duty have
to be paid with the respective financial punishment. On the other hand, customers also negatively
affected by this scenario in various ways. Customer would experience frustration by long waiting
line and loss of precious resource (time). And this trend also directly affects taxpayers since it
4
affects public infrastructures in which they would be served in. For this reason, this study would
try to analyze the problem using management science process through application of queuing
model and forward implications based on mathematical assumptions which would be used to
improve the service system performance of the organization.

1.2 Objective
The objective of this analysis is to apprise the current system performance in terms of customer
service lead-time and recommending the possible alternative way of improving the service
performance of the current system with no or minimum additional investment (cost).
1.2.1 specific objectives
 To analyze the current system with four servers operating characteristics
 To analyze the change in operating characteristics by changing the number
of servers from four to five server

5
 2. Literature review
2.1 Theoretical framework
Queueing theory is one of the branches of applied mathematics which studies and models the
waiting lines. Danish mathematician Erlang A. K. (1878–1929), who published his first paper
entitled “The Theory of Probability and Conversations” in 1909 [1], is considered as the father of
queueing theory. Further, when analyzing the history, it can be observed that a viable queueing
theory was developed by French Mathematician Poisson S. D. (1781–1840), who created a
probability distribution function for the total outcomes of independent trials. He used statistical
approach for these distributions which can be applied to the situations where excessive demands
are to be fulfilled on a limited resource. During the late 1800s, all telephone calls used to be
switched manually to the recipient by an operator. Each customer used to call the operator first
and the operator used to fix the call for the customer. In this process, telephone companies were
facing problem to appoint more operators. Callers who were unable to reach to an operator may
simply hang up for several minutes with frustration and might think that it was a busy time for
the operators. On the other hand, some would be waiting their turn to talk to the operator. And
some others would call repeatedly thinking that the operator would be sufficiently annoyed by
repeated calls to serve them next. These type of behavior of the customers caused problems for
traffic engineers because they affected the level of demand for service from an operator. A call
which was not reached to the operator could be lost and could be effectively out of the system.
To overcome this situation and to reduce the number of switchboards in an area, the most
important application of queuing theory was developed. Those callers who repeatedly try for the
operator increased demands on the system by appearing several requests. Poisson's formula was
meant only for the repeated callers. Kendall, D. G. (1951)presented a paper that opened a general
review of some points in congestion theory to enhance the study for a single server queue where
input is Poisson and service time is generally distributed. In Kendall, D. G. (1953), he further
extended the study the stochastic processes for the theory of queues and their analysis by the
method of the imbedded Markov chain. Bernard W. Taylor (2013) implied some of the elements
of waiting line analysis as described below.

6
2.1.2 Elements of waiting line analysis

Waiting lines form because people or things arrive at the servicing function, or server, faster than
they can be served. However, this does not mean that the service operation is understaffed or
does not have the overall capacity to handle the influx of customers. In fact, most businesses and
organizations have sufficient serving capacity available to handle their customers in the long run.
Waiting lines result because customers do not arrive at a constant, evenly paced rate, nor are they
all served in an equal amount of time. Customers arrive at random times, and the time required to
serve them individually is not the same. Thus, a waiting line is continually increasing and
decreasing in length (and is sometimes empty), and it approaches an average rate of customer
arrivals and an average time to serve the customer in the long run. Generally, elements of waiting
line analysis are, arrivals, servers, and the waiting line structure.
Decisions about waiting lines and the management of waiting lines are based on these averages
for customer arrivals and service times. They are used in queuing formulas to compute operating
characteristics. From this, John F. S, James M. T, Donald G. & Carl M.H (2018) in their book
called “fundamentals of queueing theory” described the operating characteristics of queuing
system in detail as follows.
a. Arrival Pattern of Customers
In usual queueing situations, the process of arrivals is stochastic (random), and it is thus
necessary to know the probability distribution describing the times between successive customer
arrivals (inter arrival times). A common arrival process is the Poisson process which is a
common process for modeling arrivals to a queueing system. Intuitively, the process can be
thought of describing events that occur “randomly” in time. It is also necessary to know whether
customers can arrive simultaneously (batch or bulk arrivals), and if so, the probability
distribution describing the size of the batch.
Another factor is the manner in which the pattern changes with time. An arrival pattern that does
not change with time (i.e., the probability distribution describing the input process is time-
independent) is called a stationary arrival pattern. One that is not time-independent is called
nonstationary. An example of a system with a nonstationary arrival pattern might be a
restaurant where more customers tend to arrive during the lunch hour than during other times of
the day. But our model would assume stationary arrival pattern.

7
It is also necessary to know the reaction of a customer upon arrival to the system. A customer
may decide to wait no matter how long the queue becomes, or, if the queue is too long, the
customer may decide not to enter the system. If a customer decides not to enter the queue upon
arrival, the customer is said to have balked. A customer may enter the queue, but after a time
lose patience and decide to leave. In this case, the customer is said to have reneged. In the event
that there are two or more parallel waiting lines, customers may switch from one to another, that
is, jockey for position. These three situations are all examples of queues with impatient
customers.

b. Service Patterns
Much of the previous discussion concerning the arrival pattern is appropriate in discussing
service. Most important, since service times are typically stochastic, a probability distribution is
needed to describe the sequence of customer service times. Service may also be single or batch.
One generally thinks of one customer being served at a time by a given server, but there are
many situations where customers may be served simultaneously by the same server, such as a
computer with parallel processing, sightseers on a guided tour, or people boarding a train. The
service process may also depend on the number of customers waiting for service. A server may
work faster if the queue is building up or, on the contrary, may get flustered and become less
efficient. The situation in which service depends on the number of customers waiting is referred
to as state-dependent service. Service, like arrivals, can be stationary or nonstationary with
respect to time. For example, learning may take place, so that service becomes more efficient as
experience is gained. The dependence on time is not to be confused with dependence on state.
The former depends on how long the system has been in operation (regardless of the state of the
system), while the latter depends on the number of customers in the system (regardless of how
long the system has been in operation). Of course, a queueing system can be both nonstationary
and state dependent.
c. Number of Servers
The number of servers is an important characteristic of a queueing system and represents a
fundamental trade-off – adding servers incur extra cost to the business but can substantially
reduce delays for customers. Thus, the choice of the number of servers from single server or
multiple servers is often a critical decision. And from this as a rule of thumb for the trade-off

8
between the number of servers and the customer delays. Another decision is the configuration of
the queuing lines.
For Multi-server system, there are several possible configurations.
 In the first case, the servers are fed by a single queue.
 In the second case, each server is fed by its own queue.
 Hybrid situations can also occur.
It is generally preferable for a multi-server queueing system to be fed by a single line. Thus,
when specifying the number of parallel servers, we typically assume that the servers are fed by a
single line. Also, it is generally assumed that the servers operate independently of each other.
These characters are in alignment with our subject system.
d. Queue Discipline
Queue discipline refers to the manner in which customers are selected for service when a queue
has formed. A common discipline in everyday life is first come first served (FCFS).However,
there are many other disciplines. Some other queue disciplines are: Last come, first served
(LCFS), which is applicable to many inventory systems, as it is easier to reach the nearest items
which are the last in; random selection for service (RSS) in which customers are selected
randomly from the queue independent of their arrival times; processor sharing (PS) in which the
server processes all customers (or jobs) simultaneously but works at a slower rate on each job
based on the number in the system (this is common in computer systems); polling, in which a
single server serves multiple queues by taking customers from the first queue, then customers
from the second, and so forth in a cycle (a traffic light is a kind of polling system); and a variety
of priority schemes where some customers receive preference in terms of being selected for
service.

e. System Capacity
In some systems, there is a physical limitation to the amount of space for customers to wait, so
that when the line reaches a certain length, no further customers are allowed to enter until space
becomes available. These are referred to as finite queueing situations; that is, there is a finite
limit to the maximum system size. A queue with limited waiting room can be viewed as one
where a customer is forced to balk if it arrives when the queue size is at its limit. On the other
hand, there is system which doesn’t have maximum system size called infinite queuing.

9
 f. Number service stages
A queueing system could have only a single stage of service, or it could have several stages. An
example of a multistage queueing system is a physical examination procedure where each patient
must proceed through several stages, comprising medical history; ear, nose, and throat
examination; blood tests; electrocardiogram; eye examination; and so on. In some multistage
queueing processes, recycling or feedback may occur (Figure 1.3). Recycling is common in
manufacturing processes, where quality control inspections are performed after certain stages,
and parts that do not meet quality standards are sent back for reprocessing.

Conceptual representation of the model

10
 3. Model Construction

Model Assumption
Description of operating characteristics of the queuing system
1. Arrival pattern
The arrival process is stochastic (random), and the inter-arrival has an exponential probability
distribution with a mean arrival rate of customer arrival per unit time. An arrival pattern is a
stationary arrival pattern which does not change with time. This means the arrival pattern of
taxpayers doesn’t change with the change in time with in working hours. A customer may have
balked without entering the system, but it was assumed no balking. But there is no reneging and
jockeying for position
2. Number of Servers and queuing discipline
There are four (4) servers who serve each taxpayer (customer) which is multi-server system with
single queue feed. Also, the servers operate independently of each other, and they serve
customers using the rule of first come first served (FCFS).
3. System Capacity and service stage
There is no physical limitation to the amount of space for customers to wait, so the model
follows infinite queueing situations. And a queueing system have only a single stage of service
which means the available servers provide the same service and each arrived customer would be
served by each server and exit or there is not another stage after each server. And service time
has an exponential probability distribution with a mean service rate which is greater than average
arrival rate. And there is no unusual server behavior.

Notation for queueing model

As we have described before customers awaiting service form one single line and then proceed
to the first available server. An example of such a multichannel, single-phase waiting line is
found familiar today in many banks. A common line is formed and the customer at the head of
the line proceeds to the first free teller Barry R., Ralph M. S, JR, &Michael E. H (2012). And the
following formula was implied by Bernard W. Taylor (2013) for calculating the operating
characteristics of multiple servers, single line queuing model.

11
  The average number of customers in the queuing system

 The average time a customer spends in the queuing system (waiting and being served)

 The average number of customers in the queue

 The average time a customer spends in the queue, waiting to be served

 The probability that a customer arriving in the system must wait for service (i.e., the
probability that all the servers are busy) is

 The probability that there are no customers in the system (all servers are idle)

 The probability of n customers in the queuing system is

Using the multiple-server model formulas, we compute the operating characteristics for the
service organization based on the following data.

12
 4. Acquiring Input Data
Arrival rate vs service rate
The data was collected through survey method specifically direct observation. And it was
collected for two days in which most queues were expected to be observed. In this case, we have
used the Ethiopian calendar and since the April month tax duties must had to be paid until May
08/2022which fall under weekend (Sunday) that no service was available, the due date of the tax
was shortened by 1.5 days since on May 07 (Saturday) only half day service was available. On
this ground, we are convinced to collect data on May 05 and 06/2022 (Thursday and Friday). In
doing so, the following data was collected.

Customer arrival data (at 27thdate) or three days Customer arrival data (at 28thdate) or two
before tax payment due date days before tax payment due date
Arrival rate on May 05,2022 Arrival rate on May 06,2022
Observation time in No of customer Observation time No of customer
local time arrived per hour in local time arrived per hour
3:00 to 4:00 am 42 3:00 to 4:00 am 48
4:01 to 5:00 am 37 4:01 to 5:00 am 36
5:01 to 6:00 am 21 5:01 to 6:00 am 31
Lunch Break Lunch Break
7:30 to 8:30 pm 39 7:30 to 8:30 pm 46
8:31 to 9:30 pm 34 8:31 to 9:30 pm 38
9:31 to 4:30 pm 19 9:31 to 4:30 pm 29
Total 192 Total 228
Mean 32 Mean 38

The mean arrival rate was calculated for the data collected on the two days or for (twelve hours)
as shown in the above table. In this case total observation (total customer arrived) in two days
(12 hours) is 420 customers (192+228). Total observation hour for two days is 12 hours (6 hrs+ 6
hrs.).
Mean (average) customer arrival rate(λ)= Total observations(∑ 0)
Total observation hours(n)

420
λ= =35 Per hr.
12
The service rate was constant by assuming that the service server wouldn’t be varied over time
13
which denoted by
μ = on average each server would serve 12 customers per hour
c = there are 4 servers (windows) that serve customers
Based on this information’s, in the following sections we analyze the operating characteristics of
the current system with four servers and then we recommend an alternative model that
maximizes the customer service capability of the system without costing the company.

5. Model solution
5.1 Operating characteristics of the system with four (4) servers
First, we will analyze the operating characteristics of the current system with four servers
λ = on average customer 35 customers arrive per hour
μ = on average each server would serve 12 customers per hour
c = there are 4 servers (windows) that serve customers
 The probability that there are no customers in the system (all servers are idle)

Po=0.04266 probability that no customer are in the service organization

 Average number of customers in the queuing system

L = 4.195customers, on average, in the service department

 The average time a customer spends in the queuing system (waiting and being served)
14
 W = 0.12 hr. (7.2 min.) average time in the service department per customer
 The average number of customers in the queue

Lq = 1.28 customers, on average, waiting to be served

 The average time a customer spends in the queue, waiting to be served

Wq = 0.0366 hr. (2.2 min.) average time waiting in line per customer
 The probability that a customer arriving in the system must wait for service (i.e., the
probability that all the servers are busy) is

Pw= 0.475probability that a customer must wait for service i.e., that there aresix
customers in the system.

5.2 Operating characteristics of the system with five (5) servers

15
In the above section we have analyzed the operating characteristics for four server, single line
queuing system. As implied in the problem definition section, the organization has another two
servers who serve the customers from governmental agencies only however because of there are
limited number of government firms who would be served by these two windows, the servers

 The probability that there are no customers in the system (all servers are idle)

𝑃𝑜 = 0.05117𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦𝑡ℎ𝑎𝑡𝑛𝑜𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑎𝑟𝑒𝑖𝑛𝑡ℎ𝑒𝑠𝑒𝑟𝑣𝑖𝑐𝑒firm

 Average number of customers in the queuing system

L = 3.219 customers, on average, in the service department

 The average time a customer spends in the queuing system (waiting and being served)

W = 0.092 hr.(5.5 min.) average time in the service department per customer
 The average number of customers in the queue

Lq = 0.30 (0) customers, on average, waiting to be served

 The average time a customer spends in the queue, waiting to be served

16
 Wq = 0.00857 hr. (0.51 min.) average time waiting in line per customer
 The probability that a customer arriving in the system must wait for service (i.e., the
probability that all the servers are busy) is

Pw = 0.216 probability that a customer must wait for service i.e., that there are three
customers in the system.

6. Discussion
We examine the operational features of the current system with four servers based on this
information in the sections that follow.
 The probability that there are no customers in the system (all servers are idle)
Po=0.04266 for four servers but Operating characteristics of the system with five (5)
servers The probability that there are no customers in the system (all servers are idle)
𝑃𝑜 = 0.05117𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑛𝑜 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑎𝑟𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 firm
Thus this implies four serves is less than five serve
 Average number of customers in the queuing system for four serve is 4.195
customers, on average, in the service department and for five serve is 3.219
customers, on average, in the service department. This is show that server five is
better than server four
 The average time a customer spends in the queuing system (waiting and being served)
For server four is 0.12 hr. (7.2 min.) average time in the service department per
customer but for serve five is 0.092 hr.(5.5 min.) average time in the service
department per customer this indicate serve five is take less time that is efficient than
serve four.
17
  The average number of customers in the queue four server is 1.28 customers, on average,
waiting to be served but for five serve is 0.30 (0) customers, on average, waiting to be
served this show that five server is less customer four server , this implies five server is
better than four serve
 The average time a customer spends in the queue, waiting to be served for four server is
0.0366 hr. (2.2 min.) average time waiting in line per customer but for five server is
0.00857 hr. (0.51 min.) average time waiting in line per customer this show that server five take
less time than server four.

 The probability that a customer arriving in the system must wait for service (i.e., the
probability that all the servers are busy) is 0.475 probability that a customer must wait for
service for four serve but for five server is 0.216 probability that a customer must wait for service
i.e., that there are three customers in the system. This indicate that the probability of four server is
greater that server five.

7. Conclusion
We have seen the Addis Ababa-area branch office network for the Ethiopian ministry of
revenue's medium taxpayers. During the last five days of the monthly tax payment days, we
have seen that the system is characterized by a multiple server system and is prone to
waiting line issues. The organization has about six servers that serve customers
concurrently, but of the servers that are available, two only provide service to taxpayers
who work for government offices, and the remaining four servers serve taxpayers who
work for private enterprises. We would concentrate on the four servers in this scenario
because of the enormous quantity of private taxpayers. As the discussion implies the four
servers compare and contrast with five servers the capacity show that the server five is
better efficiency than server four. The customer serviceability of five servers is better than
four servers.

8. Recommendation
For the medium taxpayers in the Ethiopian ministry of revenue, we have observed the
branch office network in the Addis Ababa region. We have observed that the system is
characterized by a multiple server system and is prone to waiting line troubles over the last

18
five days of the monthly tax payment days. The company has around six servers that can
serve users at once, but only two of those servers can serve customers who work for
government agencies; the other four servers can only serve customers who work for private
businesses. In this case, given the vast number of private taxpayers, we would concentrate
on the four servers. Through the use of the queuing model and forward implications based
on mathematical presumptions, this study will attempt to examine the issue using
management science methods in order to enhance the organization's service system
performance. As per discussion state that addition of one more server from the private to
government rectifies the waiting time and increase better serviceability. We recommend
that medium taxpayers in the Ethiopian ministry of revenue to add one more serve to
government for better performance.

19
 9. References

Barry R., Ralph M. S, JR, &Michael E. H (2012). Quantitative Analysis for Management, 11th
edit.
Bernard W. Taylor (2013).Introduction to Management Science11 th edit. Virginia Polytechnic
Institute and State University

Kendall, D. G. (1951). Some Problems in the Theory of Queues, Journal of the Royal Statistical
Society. Series B (Methodological), 13(2), 151-185.

Kendall, D. G. (1953). Stochastic Processes Occurring in the Theory of Queues and their
Analysis by the Method of the Imbedded Markov Chain, the Annals of Mathematical Statistics,
24(3), 338-354

John F. S, James M. T, Donald G. & Carl M.H (2018).fundamentals of queueing theory5 thedit.
Wiley series in probability and statistics.

20