Journal of Research in Management

Vol. 2, No. 1, 2019, pp. 1 - 7

Published by Indonesian Research Society
Published online 30th March, 2019 at http://irs-managementstudies.com/index.php/irs

Analysis of Single Channel-Multi Phase Queue Model

SYAFIQ ANNAS PURWANDA1, SUCI INDRIATI2

1,2
Management Department, Faculty of Economic and Business, Universitas Jenderal Soedirman, Indonesia

Abstract This research is a type of research based on case studies, where is the research with
USING a descriptive approach to explain variables what will be studied. The purpose of this
program was to find out the Queuing System Model of the Office of Central Java Province in
Pemalang District, to find the size pattern of service at UP3AD Central Java Province, and
to test the arrival pattern Poisson distribution and whether Service Time is distributed
Exponential. This research is conducted by observing the rate of arrival of taxpayers per
minute and also the time of service for taxpayers per minute. This study was tested using
the Kolmogorov-Smirnov Conformity Test which was assisted by SPSS and Microsoft Excel
applications.

Keywords Arrival Rate, Service Time, Steady State, Poisson distribution, Exponential distribution

INTRODUCTION Province in Pemalang responded positively

Queuing theory is a probabilistic study of by improving the quality of its services
waiting line events, namely a waiting line for through the acceleration of the payment
customers who need services from an process as outlined in frame posters installed
existing system. The occurrence of queues is on the walls of the service room.
due to limited service sources, which are The flow of completing tax payments goes
generally associated with limited servers due through several different stages. At the one-
to budget / financial problems. If the number year tax service counter, there are some
of available servers is limited, a long queue counters, including registration and
occurs, so the queue decides to leave the appointment counters, payment counters or
queue. This condition causes a loss for a cashiers and delivery counters, where an
company because of losing the customer, officer serves each of the counters. From the
and the company must provide sufficient queuing structure found in the UP3AD Office
servers, with the consequence that the of Central Java Province in Pemalang
company incurs greater costs (Sinalungga, Regency, the authors conclude that the
2008: 238). Queue theory basically cannot queuing model used is a single channel that
directly solve a problem with the queue. goes through several stages ( Multi-Phase ).
However, this queuing theory has the
advantage of being able to provide valuable
information in decision making such as how LITERATURE REVIEW AND
to balance the waiting time in the queue HYPOTHESES
which is done by predicting some According to Siagian (1987) "Queue" is
characteristics of the queue, for example, the defined as a waiting line from customers
average time needed in the queue (Tarliyah (units) that require services from one or more
and Dimyati, 1999: 349). services (service facilities). Queue according
At the Regional Assets Revenue and to Heizer and Render (2005) is a line of
Empowerment Service Unit (UP3AD) of people or goods that are waiting to be
Central Java Province in Pemalang Regency served. According to Bronson (quoted from
especially in the one-year motor vehicle tax Fajar, 2012) the queuing process (Queuing
approval service, there are many people Process) is a process related to the arrival of
queuing to pay motor vehicle tax, on the a customer in a service facility the busy one,
other hand the service counters and seats in the end, the customer leaves the facility
are limited, so Taxpayers complain about the due to having to wait in a line (queue).
length of time they wait. For this situation and Chase et al. (2008) in his book mentioned
condition, the UP3AD Office of Central Java that understanding and managing about


Correspondence to : [email protected] Received: Jan, 7, 2019

Revised: Jan, 23, 2019
Accepted: Feb, 1, 2019
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

queues is one of the essential things in conducting interviews with the head of the
operations management to regulate several Central Java UP3AD Office in Pemalang
schedules, job design, inventory and so on, Regency. Second, the process of research is
and discuss some fundamental problems in about observing the rate of arrival of
the queue and apply standard formulas to taxpayers whose time intervals are
solve queuing problems that is. The formula determined every 5 minutes, besides This
makes it easy for managers to analyse step includes observing the time of service
service needs which then establish service given to each taxpayer, where each taxpayer
facilities that are appropriate for certain is subject to a standard time set at the
conditions. The purpose of the queuing UP3AD Office in Central Java Province in
model is to minimise the total costs, both the Pemalang Regency for 20 minutes. All of
direct costs of providing service facilities these data are obtained based on the results
(servers) and the indirect costs that arise due of direct observations made by researchers
to customers having to wait to be served. If a at the UPPD Office in Central Java Province
queuing system has a service facility (server) in Pemalang District. Then after the data is
that exceeds the required amount, then the obtained, the researcher processes it
excessive budget/investment is needed, and scientifically by using methods that are
vice versa if the number of service facilities following queuing theory in the available
(servers) is less than what is needed then the literature. After getting the data, it is then
service will be delayed. reviewed and examined, the results of which
Some Classifications of queuing systems can be used as information to provide a
according to Hillier and Lieberman (quoted better solution for the queuing system at the
from Subagyo, et al., 1999) Commercial UPPD Office in Central Java Province in
service systems, industrial-business service Pemalang District. Research can be done if
systems, or transport services systems, there is a framework that is used as a guide
social service systems. The characteristics of that can be used to explain the concept of the
the queuing system according to Heizer and research itself. The mindset will make it
Render (2005) have three components, easier for readers to understand a study
namely Arrival (Input), Queuing Discipline, clearly and concisely about what the
and Service. According to Heizer and Render researchers are doing.
(2005) in service characteristics, there are 2 Third process is about gathering
(two) important things, namely Queue information about the procedure for making
System Design and Service Time one-year motor vehicle tax payments by
Distribution. In the design of the queuing conducting interviews with the head of the
system, services are classified according to Central Java UP3AD Office in Pemalang
the number of existing channels, namely the Regency. Fourth step is the process of
number of cashiers and the number of observing the rate of arrival of taxpayers
stages. The design of the queuing system whose time intervals are determined every 5
itself can be reduced to 4 types, namely: minutes, besides In this step, the process of
Single queue system, Multiple line queue observing the time of service given to each
system, One-stage system, Multiple stage taxpayer, where each taxpayer is subject to a
system. standard time set at the UP3AD Office in
According to Heizer and Render (2005), Central Java Province in Pemalang Regency
the primary component of the queue for 20 minutes. All of these data are obtained
measures the performance of the queue. If based on the results of direct observations
the queue is analysed, it will get a measure of made by researchers at the UPPD Office in
the performance of a queue system. The Central Java Province in Pemalang District.
queue model will be beneficial for measuring Then after the data is obtained, the process
the performance of the queue. With the contune with using methods that are
queuing model, managers are following queuing theory in the available
accommodating in deciding to balance literature. After getting the data, it is then
service costs by using queue costs. reviewed and examined, the results of which
can be used as information to provide a
better solution for the queuing system at the
METHODS UPPD Office in Central Java Province in
The first step of research is gathering Pemalang District.
information about the procedure for making This research conduct by using
one-year motor vehicle tax payments by quantitative analysis techniques based on
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

case studies, whose research is carried out significance (2-tailed) or p-value of 0.506.
using a descriptive approach to explain what The Asymptotic significance (2-tailed) value
variables will be studied. The method of data is greater than the value of α (0.05) or the
collection in this study used interviews and value p> α then H_0 is accepted, or it can be
observation. Based on the results of said that the number of arrivals at the
interviews with the Head of the Office of payment counter is a 5-minute Poisson
UP3AD Central Java Province in Pemalang distributed interval.
District, obtained results in the number of
service counters located in the Office of Submission Counter (Third Phase)
UP3AD Central Java Province in Pemalang Data on arrival rates taken at the UP3AD
Regency, the main tasks in each service Office in Central Java Province in Kabupaten
counter, the design of the queuing system in Pemalang at the delivery counter observed
the UP3AD Office of Central Java Province in on Monday 03 December 2018 with a time
Pemalang Regency, as well as the standard interval between taxpayers is 5 minutes.
service time given to taxpayers. Observation Based on the results of the analysis using the
is done by observing the arrival rate data by SPSS program, the results of the Asymptotic
observing the arrival rate every 5 minutes at significance (2-tailed) or p-value are obtained
the registration counter and setting, the at 0.713. The Asymptotic significance value
payment counter and the delivery counter. In (2-tailed) is higher than the value of α (0.05)
addition to data on the rate of arrival of or the value of p> α, H_0 is accepted, or it
researchers also observed the time of service can be said that the number of arrivals at the
provided by the Office of UP3AD Central delivery window is a 5-minute interval
Java Province in Pemalang Regency to each Poisson distribution.
taxpayer who will pay motor vehicle tax.
Service Time Distribution Test
RESULTS AND DISCUSSION Registration and Determination
Arrival Distribution Test Counters (First Phase)
Registration and Determination Test of service time distribution was tested
Counters (First Phase) using the Kolmogorov-Smirnov test. Based
Data on arrival rates taken at the UP3AD on the results of the analysis using the SPSS
Office in Central Java Province in Pemalang program the results of the Asymptotic
District at the registration and determination significance (2-tailed) or p-value are 0,000.
counters observed on Monday 03 December Small Asymptotic significance (2-tailed) value
2018 with a time interval between taxpayers of α (0.05) or p> α then H_0 is rejected, or it
is 5 minutes. The results of the observations can be said Service Time per taxpayer in the
were tested using SPSS where H_0 is the first phase that is at the registration counter is
data of the number of arrivals at the not exponentially distributed.
registration counter and the determination of For the results of the analysis of service time
the intervals of 5 minutes is Poisson distribution tests on Tuesday-Friday in the
Distribution. Based on the results of the first phase (Registration Counters) can be
analysis using the SPSS program the result seen in appendix 10. Based on the results of
of the Asymptotic significance (2-tailed) or p- the analysis with the SPSS program it can be
value is 0.638. The Asymptotic significance concluded that the service time per taxpayer
value (2-tailed) is greater than the value of α in the first phase is at the registration counter
(0.05) or the value of p> α, then H_0 is and determination, not exponential
accepted, or it can be said that the number of distribution.
arrivals at the registration counter is a 5-
minute Poisson distributed interval. Payment Counter (Second Phase)
Test of service time distribution was tested
Payment Counter (Second Phase) using the Kolmogorov-Smirnov test. Based
The arrival rate data taken at the UP3AD on the results of the analysis using the SPSS
Office in Central Java Province in Pemalang program in table 4.10, the results of the
District at the payment counter or cashier Asymptotic significance (2-tailed) or p-value
observed on Monday 03 December 2018 with are 0,000. Small Asymptotic significance (2-
a time interval between taxpayers is 5 tailed) value of α (0.05) or p> α, H_0 is
minutes. rejected, or it can be said that the Service
The data in this study were tested using Time per taxpayer in the second phase is at
SPSS showing the results of the Asymptotic the delivery counter not exponentially
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

distributed. Based on the results of the FCFS queue discipline, and also unlimited
analysis with the SPSS program it can be queue capacity.
concluded that the service time per taxpayer
in the second phase is at the payment Size of Queue System Performance
counter not exponentially distributed. Registration and Determination
Counters (First Phase)
Submission Counter (Third Phase) Based on the results of tests conducted on
Test of service time distribution was tested Monday, December 3, 2018, at the
using the Kolmogorov-Smirnov test Based on registration and determination counters, the
the results of the analysis using the SPSS results of the value λ = 0.01359, the variance
program in table 4.13, the results of the is 382.7982, and for the value of µ =
Asymptotic significance (2-tailed) or p-value 0.01717062. Then after the values of λ and µ
are 0,000. Small Asymptotic significance (2- are obtained, we can find the measure of the
tailed) value of α (0.05) or p value> α, H_0 is effectiveness of the performance of the
rejected, or it can be said that Service Time queuing system at the registration counter
per taxpayer in the third phase is at the using the M / G / 1 equation, namely :
delivery counter not exponentially distributed.
Based on the results of the analysis with the Table 1. Calculation of registration and
SPSS program, it can be concluded that the determination counters
service time per taxpayer in the third phase is N use
No. Information
at the delivery counter. values
1 Steady state 0.79147
The amount of
Determination of Each Phase Queue 1,369
2 mandatory tax payer in the
Model system
tax payer
Registration and Determination The amount of
0.578
Counters 3 mandatory tax enqueue
tax payer
Based on the results of the distribution test, it ( Lq )
can be seen that the rate of arrival is Poisson Time Wait for 100,736
distribution, while for the time the service is 4 mandatory tax payer in the seconds
not exponentially distributed (General system ( WS )
Time Wait for 42.5313
distribution) or in conditions like this it is said
5 mandatory tax payer in the seconds
to be normal. At this counter only has one queue ( Wq )
server to service each taxpayer who comes,
then for the queue discipline used is the
Payment Counter (Second Phase)
FCFS queue discipline, and also unlimited
Based on the results of tests conducted on
queue capacity.
Monday, December 3, 2018, at the payment
counter, the value of λ = 0.01359, the
Payment Counter variance is 400.784, and for the value of µ =
Based on the results of the distribution test, it 0.1764. To measure the performance of the
can be seen that the rate of arrival is Poisson queue in the first phase the same calculation
distribution, while for the time the service is is used, namely by calculating Ls, Lq, Ws,
not exponentially distributed (General Wq on Tuesday, Wednesday, Thursday and
distribution) or in conditions like this it is said FridayTo measure the performance of the
to be normal. At this counter only has one queue in the first phase used the same
server to service each taxpayer who comes, calculation, namely by calculating Ls, Lq, Ws,
then for the queue discipline used is the Wq on Tuesday, Wednesday, Thursday, and
FCFS queue discipline, and also unlimited Friday. The results of the whole calculation
queue capacity. can be seen in the following table.
Submission Counter
Based on the results of the distribution test, it
can be seen that the rate of arrival is Poisson
distribution, while for the time the service is
not exponentially distributed (General
distribution) or in conditions like this it is said
to be normal. At this counter only has one
server to service each taxpayer who comes,
then for the queue discipline used is the
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

Table 2. Payment counters queue capacity level is unlimited. According

No. Information Value to the results above, it can be concluded:
1 Steady state 0.07705
The amount of
0.1432 Steady-state size
2 mandatory tax payer in the
tax payer Based on the results of calculations, the
system
The amount of
lowest steady-state size on Wednesday is p =
0.0662 0.628 in the registration and determination
3 mandatory tax payer enque
tax payer wall, for the payment counter the lowest p-
ue ( Lq )
Time Wait for value on Monday with a value of p = 0.07705,
10,5372
4 mandatory tax in a for the lowest counters at Monday with a
seconds
system ( WS ) value of p = 0.0862. Whereas for the highest
Time Wait for p-value at the registration window and
4.8713
5 mandatory tax payer in the determination occurs on Monday with a value
seconds
queue ( Wq ) of p = 0.79147, for the payment counter the
highest p-value on Thursday with a value of p
Submission Counter (Third Phase) = 0.7255, for the highest p-value delivery
Based on the results of tests conducted on counter on Thursday with a value p = 0.7255.
Monday, December 3, 2018, at the delivery Based on these values it can be concluded
counter, the value of λ = 0.01359, the that the queuing system applied in the
variance is 414,027, and for the value of µ = UP3AD Office of Central Java Province in
0.1578. Then after the values of λ and µ are Pemalang Regency is still steady state, which
obtained, we can find the measure of the means the queue system is still effective. The
effectiveness of the performance of the effectiveness of the queuing system at the
queuing system at the delivery counter using UP3AD Office in Central Java Province was
the M / G / 1 equation, namely. caused by various kinds of facilities that
made it easier for taxpayers to pay their
Table 3. Submission Tables
vehicle tax, such as the SAMSAT DRIVE
No. Information Value
THRU facility, SAMSAT KELILING, and the
1 Steady state 0.0862
latest service facilities SAMSAT NIGHT DAY.
2 The amount of 0.13 tax
mandatory tax payer in payer Number of taxpayers expected in the system
the system (LS)
3 The amount of 0.0438 tax Based on the results of calculations that
mandatory tax paye payer have been done, at the registration counter
enqueue ( Lq ) and the determination of the lowest Ls value
4 Time Wait for 9.5659 is obtained on Wednesday with a value of
mandatory tax payer in a seconds 0.9556 or only one person queuing in the
system ( WS ) system. The lowest ticket value is Monday at
5 Time Wait for 3.223
0.1432 or only One person is queuing in the
mandatory tax payer in a seconds
queue ( Wq ) system, for the delivery counter the lowest Ls
value is Monday with a value of 0.13 or only
one person is queuing in the system. While
Discussion
for the highest Ls value occurs at the
The queuing system model implemented by
registration counter and a determination
the Regional Assets Revenue and
occurs on Monday with a value of 1,369 or
Empowerment Service Unit (UP3AD) in
there are two taxpayers queuing in the
Pemalang Regency is a multi-phase model
system. For the payment counter the highest
consisting of 3 stages of service, namely the
value of Ls occurs on Thursday with a value
registration phase, the payment phase, and
of 1.1693, or two taxpayers are queuing up in
also the delivery phase. For each counter at
the system, for the highest number of Ls
the UP3AD Office in Central Java Province in
delivery counters to occur on Thursday with a
Pemalang District using the M / G / 1; FCFS /
value of 1.1724, or two taxpayers are
∞ / ∞ queuing model with Poisson distribution
queuing in the system. So the number of
coming. Service time is not exponentially
taxpayers queuing in crowded conditions is 6
distributed (General distribution), each
and in a quiet condition is 3.
window has only one clerk to service each
taxpayer, queue discipline used is FCFS, the
first to come will be served first, and the
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

Number Of Taxpayers Expected In The 27.529 seconds, for the payment counter the
Queue (LQ) lowest Wq value is Monday with a value of
Based on the results of calculations that have 4.7183 seconds, for the lowest countdown
been done, from the results of calculations window on Monday with a value of 3.233
that have been done, at the registration seconds. Whereas for the highest Wq value
counter and determination of the lowest Lq at the registration counter and determination
value obtained on Wednesday with a value of occurred on Tuesday with a value of 52,100
0.3276 or only one person queuing in the seconds, for the payment counter the highest
system. For the payment counter, the lowest Wq value occurred on Thursday with a value
Lq value is on Wednesday with a value of of 39,982 seconds, for the delivery window
0.2981 or only one person queuing in the the highest Wq value occurred on Thursday
system, for the delivery window the lowest Lq with a value of 40,261 seconds.
value on Monday is 0.0438 or only one
person is queuing in the system. Whereas for CONCLUSION
the highest value occurs at the registration Based on the results of analysis and
counter and determination occurs on calculations about the queuing system at the
Tuesday with a value of 0.6200 or there are 2 UP3AD Office of Central Java Province in
taxpayers queuing in the system, for the Pemalang Regency the following
payment counter the highest LQ value occurs conclusions can be drawn. The queuing
on Thursday with a value of 0.4438, or there model applied by the UP3AD Office of
is 1 person obligatory tax queuing in the Central Java Province in Pemalang Regency
system, for the highest LQ value delivery is included in the multi-phase queue model,
counters to occur on Thursday with a value of where the queuing model is organised into
0.4469 or there is 1 taxpayer waiting in the three stages, as follows. Registration and
system. So the number of taxpayers queuing Determination Counters apply the M / G / 1;
in crowded conditions is 4 and in a quiet FCFS / ∞ / ∞ model. Payment counters apply
condition is 3. the M / G / 1; FCFS / ∞ / ∞ model.Handing
counter using model of M / G / 1; FCFS / ∞ /
Taxpayer waiting time in the system ∞. Services at the UP3AD Office in Central
(WS) Java Province in Pemalang Regency have
Based on the results of calculations that have been effectively seen from the steady-state
been done, From the results of calculations value at each counter or each stage of
that have been done, at the registration service. The average number of taxpayers
counter and the lowest value of Ws is contained in the queuing system at UP3AD
obtained on Wednesday with a value of Office in Central Java Province in Pemalang
80,303 seconds. For the payment counter, Regency in crowded conditions was six
the lowest Ws value on Monday is 10,5372 people, and three taxpayers were quiet. The
seconds, for the counter delivery of the average number of taxpayers contained in
lowest Ws value on Monday with a value of the queue at UP3AD Office in Central Java
9.5659 seconds. While for the highest Ws Province in Pemalang Regency in crowded
value at the registration counter and conditions was four people, and three
determination occurred on Tuesday with a taxpayers were in a quiet condition. The
value of 112.706 seconds, for the payment average time spent by taxpayers in the
counter the highest Ws value occurred on queuing system at UP3AD Office in Central
Thursday with a value of 107.775 seconds, Java Province in Pemalang Regency in
for the delivery window the highest Ws value crowded conditions was 326.102 seconds
occurred on Thursday with a value of and 100.4061 in quiet conditions. The
105.621 seconds. So the time needed in the average time spent by taxpayers in the queue
system when in crowded conditions is equal at UP3AD Office in Central Java Province in
to 462.5081 seconds, equivalent to 8 minutes Pemalang Regency in crowded conditions
needed to complete the process of paying a was 132,343 seconds and 35,4803 seconds
one-year vehicle tax. in quiet conditions.
Based on the results of research at the
Taxpayer waiting time in queue (Wq) Revenue Service and Regional Asset
Based on the results of calculations that have Empowerment Unit (UP3AD) Central Java
been done, at the registration counter and the Province in Pemalang District, the results
determination of the lowest Wq value is show that: (1) Registration and Determination
obtained on Wednesday with a value of Counters with a value of p = 0.79147 indicate
Journal of Research in Management, Vol.2, No.1, 2019, pp.1 - 7

that the value has been steady state (p <1 ) REFERENCES

which means that the service provided is Alfiani, A.R, Sudarno, S. (2013). Analisis antrian
sufficient enough and this needs to be rawat inap berdasarkan spesialisasi penyakit di
maintained. (2) Payment counters with a rsup dr. Kariadi semarang. Jurnal Gaussian.
value of p = 0.7255 indicate that the value is Vol 2. No. 4. pp 269-278.
steady state (p <1), which means that the
Putranto, M. A. (2014). Analisis masalah sistem
service provided is sufficiently active if and antrian model multi phase pada kantor samsat
this needs to be maintained. (3) Submission yogyakarta. Skripsi Tidak Dipublikasikan.
counters with a value of p = 0.7255 indicate Yogyakarta. Universitas Negeri Yogyakarta.
that the value has been steady state (p <1)
which means that the service provided is Taha, H. A. (2007). Operation research: An
th
quite useful and this needs to be maintained. introduction (8 Ed). Asoke K. Ghosh, Prentice
From the three counters above, it can be Hall Of India , Delhi.
concluded that for the overall service to the
Sinalungga, S. (2008). Pengantar teknik industri.
Office of UP3AD Central Java Province in
Yogyakarta: Graha Ilmu.
Pemalang Regency, especially in the one-
year motorcycle tax approval service, it is Tarliah, T. & A. Dimyati. (1999). Manajemen
quite useful and this must be maintained so Operasi: Model-model pengambilan
that taxpayers are satisfied. keputusan. Bandung: PT. Sinar Baru
This research was conducted at UP3AD Algesindo.
Office in Central Java Province in Pemalang
Regency, namely by analysing three Supranto, J. (1987). Matematika untuk ekonomi
variables in the form of taxpayer arrival rates, dan bisnis buku 1. Lembaga Penerbit, Fakultas
Ekonomi Universitas Indonesia, Jakarta.
service time needed to handle each taxpayer,
and the performance of the queuing system Siagian, P. (1987). Penelitian operasional: Teori
itself. For this reason, it is important to dan praktek. UI Press.
examine the problem of this multi-phase
system, not only the three variables must be Heizer, J & Barry, R. (2005). Operations
examined, but also other variables such as management. Jakarta: Salemba Empat.
the rate of arrival from stage to stage to be
more clear. Besides, it is essential to Gross, D. & Harris, C. M. (2008). Fundamental of
implement optimisation programs at other queueing theory: Fourth edition. New Gersey:
John Willey & Sons, Inc.
companies or institutions that apply the multi-
phase queue system. Moreover, it is crucial Djarwanto P.S & Subagyo, S. (1996). Statistic
to use several optimisation programs that Inductive. Jakarta: Beff.
race against the target, for example, the
efficiency of operational costs and time A. Taha, Hamdy. (1996). Riset operasi. Jakarta:
customers are wasted due to queuing, or Binarupa Aksara.
other problems related to the queue.
Mulyono, S. (2002). Riset operasi. Lembaga
Penerbit Fakultas Ekonomi Universitas
Indonesia: Jakarta.

Bronson, R & Wospaknk, H. J. (1998). Operation

research: Teori dan soal-soal. Jakarta:
Erlangga