DOKK Library

A Simulation Study for Waiting Line Systems by C Programming

Authors Ozer Ozdemir

License CC-BY-4.0

Plaintext
       INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTERS IN SIMULATION
       DOI: 10.46300/9102.2020.14.16                                                                                         Volume 14, 2020




          A Simulation Study for Waiting Line Systems
                      by C Programming
                                                               Ozer O]GHPLU


                                                                                     Many problems which are not understand directly and are
    Abstract— Simulation is a method that using especially                       not possible to produce experimentally are started to solve by
waiting line systems. In this study, a simulation study has made                 using method of simulation over computers after 1960. This
for waiting line systems by C programming. The particular queue                  approach is used for solving problem as a quite common
will be a place where customers are waiting for service behind                   scientific approach [2].
each other. For the purpose of this study, a simulation in C                          Simulation programs which are written for considering
language is created in order to simulate the aforementioned queue
                                                                                 ranking of conditions from invention which was inserted queue
system. With this simulation, it is aimed to increase customers’
satisfaction by decreasing service cost, and to increase the quality             are developed in 1980s [1].
of service with minimum waiting time. Considering alternative                         Simulation studies which can be useful for these kinds of
solutions, the optimal system is obtained to ensure economical                   systems provide improving system, solving problems in the
stability between the interests of the customers and the                         system and increasing satisfaction of manager, employee and
management.                                                                      customer in a certain economical stability.
                                                                                      This study is prepared for improving queue systems which
  Keywords—C programming language, Cash desks system,                            are caused problems between people and producing a study for
Simulation, Waiting line system.                                                 finding alternative solutions for problems.
                                                                                      A market’s cash desk system which has only one queue is
                         I. INTRODUCTION                                         considered by using C programming in this study. Creating the
                                                                                 most suitable system is purposed aspect of management and
S    imulation is a method that provides acting an original
     system by means of a mathematical model which is
     essential for using computer [3]. Simulation is used in
                                                                                 customer. All good and bad ways of the present system are
                                                                                 denoted after simulation study and suggestions are given for
various spaces. One of the most useful spaces is queue system.                   improving bad ways of the present system. In short, this study
This system can be called waiting line system. Today this                        is prepared for improving queue systems which are caused big
system occurs in many states such as ticket queue, medicine                      problems between people and finding alternative solutions for
queue, salary queue, bank queue or shopping queue and                            problems. Simulations of some distributions which are used in
alternative solutions which are related with problem of queue                    this study will be denoted before simulation application.
aspect of people satisfaction are investigated.                                     There are several papers about simulation studies such as
      History of Simulation starts with China war’s games                        given in [8-17].
which are called “Weich” and comes from 5000 years ago. It
continues until 1780s [1]. To considering simulation as a                              II. SIMULATION OF UNIFORM RANDOM VARIABLE
method of experimentation and aim of scientific runs across
1944. It goes on a study about atom bomb which is prepared
                                                                                    The probability density function of random variable        X
by approach of simulation of Von Neumann and Slam during
Second World War. This study includes simulation of                              which is distributed uniformly in    a, b is given as
probability problems which are related with random neutron
matching diffusion in fissile materiel.
                                                                                               1          for a  x  b
      A simulation method which is called “Monte Carlo”                                       
                                                                                     f ( x)   b  a                       .
becomes popular by studies of Fermi, Von Neumann and Slam
                                                                                               0       for x  a or x  b
and it is used for solving theoretical problems by using
computers by many mathematicians, physicists and
statisticians [2]. Simulation improves as a management tool by
coming business computers and using them by combining                               The cumulative distribution function is:
them in 1950s. Specialized languages of computers are
developed in 1960s for considering extended size problems
effectively [1].

  Ozer OZDEMIR is with the Department of Statistics, Eskisehir Technical
University Eskisehir 26470 TURKEY (corresponding author to provide
phone: 902223350580-804668; e-mail: ozerozdemir@eskisehir.edu.tr).

       ISSN: 1998-0159                                                     115
           INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTERS IN SIMULATION
           DOI: 10.46300/9102.2020.14.16                                                                                Volume 14, 2020


                0             for x  a,                                     E t  
                                                                                         1
               x  a                                                                    
      F ( x)               for a  x  b,
               b  a          for x  b,
                1                                                            So its probability density function is:

      The expected value is                                                    f (t )  et        for t  0,

                  ab                                                                                             t from f (t ) .
      E  x 
                                                                              We can find random sample of
                      ,
                   2
                                                                              Distribution function is:
      and variance is
                                                                                         t

                 (b  a)2                                                     F (t )   e  x dx  1  e t for t  0,
      V ( x)                                                                            0
                    12
                                                                              F (t )  R  t  F 1 ( R)
      Suppose that U is an uniform variable in   0,1 .                                             1
                                                                              1  e t  R  t   ln(1  R) .
                                                                                                           
      F  x  U        for a  x  b,                                        t 
                                                                                             1
                                                                                                 ln R for R[0,1],
                                                                                             
      Solution of this equation is:
                                                                            is uniform number.
                                                                            For example, when   4 customers come for service in
      X  a   b  a U
                                                                         an hour and R  0.9 , time is:

      properly [4-7].
                                                                                   1
                                                                              t   ln 1  0.9   0.577 hour
                                                                                   4
      III. SIMULATION OF EXPONENTIAL RANDOM VARIABLE                             34.5 minutes

      If    F  x   1  e x then F 1 U  can be found by               until another customer comes. R should choose random.
                                                                         Table of random numbers is used for this choice [4-7].
solution of equation      1  e x  u according to variable x .
                                                                                                 IV. RESULTS AND DISCUSSION
      X   log 1  U                                                  Two cash desks system which are called “Express Cash Desk-
                                                                         Maximum Five Products” are going to consider in one of the
                                                                         most famous market in Turkey. Customers take service by
      Hence  Y  F 1 U    log 1  U  , whose expected             using one queue from these cash desks and when customers
value is 1, is distributed exponential for U which is uniform            come for service, they take service from free cash desk. If two
variable in  0,1 .
                                                                         cash desks are free then customers choose first cash desk.
                                                                         Mean of time which is between arrivals of customers for these
                                                                         cash desks will be defined by assuming that it corresponds
      1  U   0,1                                                   with exponential distribution. Also it will be defined by
                                                                         assuming that time of cashiers’ working corresponds with
                                                                         uniform distribution. Customers are serviced according to a
      is uniform [4-7].                                                  principle that is “First service is given to customer who comes
                                                                         firstly”. Purpose of simulation is finding performance
                                                                         measures which can be listed as
      Short Example For Application
                                                                         1.      Mean usage time of two cash desks;
    t shows the time that is between arrivals of customers for a         2.      Mean number of customers who queue(Mean length of
barbershop. t indicates exponential distribution and its mean                           queue);
is:                                                                      3.      Mean waiting time of a customer who queues.




           ISSN: 1998-0159                                         116
       INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTERS IN SIMULATION
       DOI: 10.46300/9102.2020.14.16                                                                                  Volume 14, 2020


     Mean of time which is between arrivals of customers for                Mean of time which is between arrivals of customers is
two cash desks and time of cashiers’ working will be defined            exponential distribution and its mean is 43 seconds according
by considering system for 30 minutes for finding these                  to data. Time of service is 40-82 seconds for first cashier and
performance measures. For that reason, this cash desk system            56-110 seconds for second cashier.
is considered for 30 minutes and a table which is below is                   p denotes time which is between arrivals of customers,
prepared (Between 17:00 and 17:30 o'clock).
                                                                        q1 denotes time of service for first cashier and q2 denotes
     Cash desk system considers customers who come for
service in 30 minutes. 42 customers come for service in this            time of service for second cashier for uniform random number
time. Mean of time which is between arrivals of customers for            0  R  1 .
two cash desks is found (M.A.C.);

M.A.C.     17 : 29 : 40  17 : 00 :17  /  42  1                         These values will be calculated with:
           43 seconds.
                                                                              p  43ln  R  seconds for 0  R  1,
     Mean and standard deviation of time of cashiers’ working           by using equation
are calculated.
                                                                                     1
      For first cashier;                                                      p        ln R
                                                                                     
      1  61 Seconds
                                                                        and
      1  21 Seconds
                                                                              q1  40  42R seconds for 0  R  1,
      For second cashier;
                                                                              q2  56  54R seconds for 0  R  1,
      2  83 Seconds
       2  27 Seconds                                                  by using equation

     Application can be changed as below after these reasons.                 q  a  b  a  R .
     Two cash desks system which are called “Express Cash
Desk-Maximum Five Products” are going to consider in one of
                                                                              Performance measures are achieved by given data,
the most famous market in Turkey. Customers take service by
                                                                        formulates and a program which is written by using “Microsoft
using one queue from these cash desks and when customers
                                                                        Visual Studio C++ 6.0” (Number of customers is given 1000,
come for service, they take service from free cash desk. If two
                                                                        if this number is increased, we can obtain better results).
cash desks are free then customers choose first cash desk.
                                                                              Mean usage time of first cash desk is asked for first
Mean of time which is between arrivals of customers for these
                                                                        performance measure. For answering this question, time that
cash desks will be defined by assuming that it corresponds
                                                                        first cashier did not give service subtracts from time that first
with exponential distribution whose mean is 43 seconds.
                                                                        cashier totally gave service and this result is divided by this
Customers are serviced according to a principle that is “First
                                                                        total service time. There is similar operation for second cash
service is given to customer who comes firstly”. Purpose of
                                                                        desk.
simulation is finding performance measures which are
                                                                              Mean number of customers who queue or mean length of
identified as:
                                                                        queue is asked for second performance measure. For
                                                                        answering this question, total waiting time is divided by
1.        Mean usage time of two cash desks;
                                                                        simulation time.
2.        Mean number of customers who queue(Mean length
                                                                              Mean waiting time of a customer who queues is asked for
          of queue);
                                                                        third performance measure. For answering this question, total
3.        Mean waiting time of a customer who queues.
                                                                        waiting time is divided by number of customers.
                                                                              When the program which is convenient for problem and
      First cashier gives service between   40  61  21  40          data is run, output of the program is obtained.
and   82  61  21  82  seconds by uniform distribution.                  Last customer comes for service at 17 hours (1000
                                                                        customers approximately come for service in 17 hours).
      Second        cashier     gives       service     between         Maximum length of queue is 2 customers for service. This
56 83  27  56  and 110 83  27  110  seconds by                  result is good aspect of management and cashiers who work
uniform distribution.                                                   these cash desks. Because there is no accumulation for queue
                                                                        and there is no customers’ dissatisfaction so, cashiers work


       ISSN: 1998-0159                                            117
       INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTERS IN SIMULATION
       DOI: 10.46300/9102.2020.14.16                                                                                                  Volume 14, 2020


well. We can see from first performance measure that second                        [9] B. Khoshnevis, Discrete Systems Simulation, Mc Graw Hill Inc., 1994.
                                                                                   [10] R. E. Shannon, System Simulation: The Art and Science, Prentice-Hall
cashier works more than first cashier but this situation is                              Inc., 1975.
observed because second cashier works slower than first.                           [11] W. C. Griffin, Queueing; Basic Teory and Aplications, Grid Inc., 1978.
Mean length of queue is very small number as 0.08. Very short                      [12] G. S. Fishman, Discrete-Event Simulation-Modeling, Programming and
queue is occurred obtaining from statistics of maximum length                            Analysis, Springer-Verlag New York Inc., 2001.
                                                                                   [13] P. Bratley, B. L. Fox, L. E. Schrage, A Guide to Simulation, Springer-
of queue. This result is very important aspect of management.                            Verlag New York Inc., 1987.
Because management decides to increase or decrease numbers                         [14] J. A. Payne, Introduction to Simulation, Mc Graw-Hill Inc, 1982.
of cash desks according to this result. If it decides to decrease                  [15] F. L. Severance, System Modeling and Simulation-An Introduction, John
                                                                                         Wiley and Sons Ltd., 2001.
numbers of cash desks, it provides benefit aspect of                               [16] L. Kosten, Stochastic Theory of Service Systems, Pergoman Pres.
economical stability but its result can be seen a simulation                             Oxford, 1973.
study again. Results can be obtained according to length of                        [17] T. K. Shridharbhai, Probability and statistics with reliability, queuing,
                                                                                         and computer science applications, Prentice-Hall, 1982.
queue after decreasing numbers of cash desks. A good result is
seen from statistics of mean waiting time of a customer who
queues as mean length of queue. 5.4 seconds is very short for
                                                                                   Ozer Ozdemir was born in Turkey in 1982. He received his B.Sc., M.Sc. and
waiting for a service. Consequently, mean waiting time of a
                                                                                   Ph.D. degrees in statistics in the Department of Statistics at Anadolu
customer who queues is appreciated by customers and                                University, Turkey, respectively in 2005, in 2008 and in 2013. He has worked
manager.                                                                           as a Research Assistant from 2006-2008, as a Lecturer from 2008-2014 and
                                                                                   as an Assistant Professor from 2014-2018 in the Department of Statistics at
                                                                                   Anadolu University, Turkey. He has worked as an Assistant Professor from
                          V. CONCLUSIONS                                           2018 in the Department of Statistics at Eskisehir Technical University,
A market’s cash desks system with one queue was considered                         Turkey.
                                                                                   He has published over 70 international conference papers and journals in his
in this study and results were obtained. 3 performance                             research areas. His research interests include Applied Statistics, Simulation,
measures concern with system are obtained after this study.                        Artificial Neural Networks, Fuzzy Logic, Fuzzy Modeling, Time Series,
Firstly, mean usage times of both two cash desks are very                          Computer Programming, Statistical Software and Computer Applications in
close each other and these are found approximately %90.                            Statistics.
Number of cash desks can be like this according to this result.
If number of cash desks is increased, it may get economical                           Creative Commons Attribution License 4.0
burden aspect of management. Secondly, mean length of                                 (Attribution 4.0 International, CC BY 4.0)
queue of system which has one queue and two cash desks were
simulated. Very short length of queue was obtained as 0.08                            This article is published under the terms of the Creative
customers. Finally, a result which likes second performance                           Commons Attribution License 4.0
measure was found. Mean waiting time of a customer who                                https://creativecommons.org/licenses/by/4.0/deed.en_US
queues was found 5.4 seconds.
     Both customers and manager can be appreciated from
these results. These results are indicator that management
works well and there is not any problem concern with long
queue.
   In further studies, various queue systems as bank queue,
medicine queue, ticket queue or salary queue can be similarly
considered like this study. Systems with many queues can be
considered instead of one queue and systems with three or
more cash desks can be considered instead of system with two
cash desks.

                             REFERENCES
[1] A. Behiye, Simulasyon Nedir? [What is Simulation?], Retrieved
      September                    10,             2007,             from
      http://bilgibirikimi.tripod.com/simulasyon.htm, 2007.
[2] K. Ozdamar, Bilgisayar ile Benzetim Yöntemleri (Benzetime Giriş),
      Anadolu U. Fen-Edebiyat F., 1988.
[3] H. Sarıaslan, Sıra Bekleme Sistemlerinde Simülasyon(Benzetim) Tekniği,
      A.Ü.S.B.F. Matb., 1986.
[4] A. H. Taha, Operational Research, Prentice-Hall, 2002.
[5] J. Banks, J. S. Carson II, B. L. Nelson, D. M. Nicol, Discrete-event
      system simulation, Prentice-Hall, 2001.
[6] S. M. Ross, Simulation, Elsevier Inc. 2006.
[7] A. M. Law, W. D. Kelton, Simulation Modeling and Analysis, Mc Graw-
      Hill International Editions, 1991.
[8] W. L. Winston, Operations Research Applications and Algorithms, PWS
      - KENT Publishing Company, 1991.



       ISSN: 1998-0159                                                       118