TH-SOF-2043-Performance Evaluation Model For MMALs Using Multi-System Variables Through Discrete Event Simulation

Performance evaluation model for MMALs using multi-system
variables through Discrete Event Simulation
Author
Rizwan Zafar
0000359151
Supervisor
Dr. Aamer Ahmed Baqai
DEPARTMENT OF MECHANICAL ENGINEERING
COLLEGE OF ELECTRICAL & MECHANICAL ENGINEERING
NATIONAL UNIVERSITY OF SCIENCES & TECHNOLOGY
ISLAMABAD
AUG, 2022
Performance evaluation model for MMALs using multi-system variables
through Discrete Event Simulation
Author
Rizwan Zafar
0000359151
A thesis submitted in partial fulfillment of the requirements for the degree of
MS Mechanical Engineering
Thesis Supervisor
Thesis Supervisor’s Signature: _______________________________________
DEPARTMENT OF MECHANICAL ENGINEERING
COLLEGE OF ELECTRICAL & MECHANICAL ENGINEERING
NATIONAL UNIVERSITY OF SCIENCES & TECHNOLOGY
ISLAMABAD
AUG, 2022
Declaration
I certify that this research work titled Performance evaluation model for MMALs using
multi-system variables through Discrete Event Simulation is my own work. The work has not been
presented elsewhere for assessment. The material that has been used from other sources it has been
properly acknowledged / referred.
Signature of Student
Rizwan Zafar
2020-NUST-MS-ME-0000359151
Language Correctness Certificate
This thesis has been read by an English expert and is free of typing, syntax, semantic,
grammatical and spelling mistakes. Thesis is also according to the format given by the university.
Rizwan Zafar
Registration Number
00000359151
Signature of Supervisor

Plagiarism Certificate (Turnitin Report)
This thesis has been checked for Plagiarism. Turnitin report endorsed by Supervisor is
attached.
Rizwan Zafar
Registration Number
00000359151
Signature of Supervisor

Copyright Statement
• Copyright in text of this thesis rests with the student author. Copies (by any process) either
in full, or of extracts, may be made only in accordance with instructions given by the author
and lodged in the Library of NUST College of E&ME. Details may be obtained by the
Librarian. This page must form part of any such copies made. Further copies (by any process)
may not be made without the permission (in writing) of the author.
• The ownership of any intellectual property rights which may be described in this thesis is
vested in NUST College of E&ME, subject to any prior agreement to the contrary, and may
not be made available for use by third parties without the written permission of the College
of E&ME, which will prescribe the terms and conditions of any such agreement.
• Further information on the conditions under which disclosures and exploitation may take
place is available from the Library of NUST College of E&ME, Rawalpindi.
Certificate of completeness
It is hereby certified that the dissertation submitted by NS Rizwan Zafar, Regn No. 00000359151
Titled: Performance evaluation model for MMALs using multi-system variables through
Discrete Event Simulation, has been checked/reviewed and its contents are complete in all
respects.
Supervisor’s Name: Dr. Aamer Ahmed Baqai Signature: ____________
Date: ________________
Acknowledgments
First and foremost, I am thankful to my Creator Allah Subhana-Watala to have guided me
throughout this work at every step and for every new thought which You setup in my mind to
improve it. Indeed I could have done nothing without Your priceless help and guidance.
Whosoever helped me throughout the course of my thesis, whether my parents or any other
individual was Your will, so indeed none be worthy of praise but You.
I am profusely thankful to my beloved parents who raised me when I was not capable of
walking and continued to support me throughout in every department of my life.
I would also like to express special thanks to Supervisor Dr. Aamer Ahmed Baqai for being
such an inspiring supervisor. He was always approachable and gave great advice during the ups
and downs of my work. This thesis would not have been completed without his excellent and
encouraging support.
I would also like to thank my Co-Supervisor Dr. Uzair Khaleeq-uz-Zaman, and members
of guidance and evaluation committee, Dr. Sajidullah Butt and Dr. Yasir Ahmed for being always
available to answer questions related to my research work and providing me necessary guidance
at the right time. Even though he had no responsibility towards me, he always responded promptly.
I am thankful to Dr. Imran Akhter for providing me the support of High-Performance Computing
Center (HP2C) at College of E&ME for completing my simulations of the research work.
I extend my thanks to Dr. Shahid Ikramullah Butt, head of department at the School of
Mechanical and Manufacturing Engineering, National University of Sciences & Technology for
sparing time to help me in my queries related to theoretical aspects of research work. I would like
to thank him for his support. I also owe my gratitude to my colleagues who enabled me to tread
through the learning curve and develop reasonable expertise within a very short time frame.
My special gratitude to my wife Sadaf, who supported and encouraged me during the entire
span of my MS studies. I would not have been able to dedicate entire attention to my studies
without her active presence and appreciation. My son and daughter remained the source of
’renewable energy’ when ever i needed it.
Finally, I extend my deepest gratitude to Allah for providing me the opportunity and the
ability to undertake this research
Rizwan Zafar
Dissertation submitted to the Faculty of the Department of Mechanical Engineering,
College of Electrical & Mechanical Engineering,
National University of Sciences & Technology (NUST)
in partial fulfillment of the requirements for the degree of
Master of Science
in
Mechanical Engineering
Dr. Aamer Ahmed Baqai, Supervisor
Dr. Uzair Khaleeq-uz-Zaman
Dr. Sajidullah Butt
Dr. Yasir Ahmed
28 August, 2018
College of Electrical & Mechanical Engineering,
National University of Sciences & Technology (NUST), Islamabad

Rizwan Zafar
ABSTRACT
Nowadays, in current global competitive market era, all manufacturing organizations are
endeavoring to integrate innovative process modeling and simulation-based performance methods
to improve production capacity and efficiency. Several system changes are required for existing
manufacturing operations to improve profit and maintain position in global business sectors.
Increase in product demand require the transformation of traditional assembly lines to mixed
model assembly lines in modern production systems. Manufacturing industries face many
problems like long manufacturing lead times, bottlenecks, low-capacity utilizations and low
productivity which impact the overall efficiency. leads to decline in process efficiency of the
manufacturing industries. Various techniques are used to enhance the process efficiency in a mixed
model assembly line e.g., mathematical modeling, heuristic approach, search algorithm approach
and simulation modeling method etc. Among such techniques, simulation modeling is a well-
accepted approach for improving process efficiency. Discrete event simulation (DES) is a
modeling method that reduces bottlenecks in the manufacturing process and ensures peak
performance of the system. This research work uses DES to analyze and recommend viable options
to enhance the production of a high technology parts manufacturing industry. The simulation
model is developed by using ARENA ® Simulation software for a case study. Performance metrics
are assessed for existing and proposed reconfigurable assembly lines with increased productivity
and overall efficiency. The simulation model helps to find the essential performance indicators of
the production assembly lines for the case study. The results are then compared with the goal of
optimizing the workflow in existing assembly lines and increasing the production capacity.
Keywords: Discrete Event Simulation, Simulation Modeling, Workflow Optimization.

Dedicated to my amazing Father (Late), who always encouraged me in
my determination to find and realize my potential, my exceptional and
adored siblings whose tremendous support and cooperation led me to
this wonderful accomplishment.
Table of Contents
Declaration ........................................................................................................................................................... iii
Language Correctness Certificate ...................................................................................................................... iv
Plagiarism Certificate (Turnitin Report)............................................................................................................ v
Copyright Statement ........................................................................................................................................... vi
Certificate of completeness ................................................................................................................................ vii
Acknowledgments .............................................................................................................................................. viii
ABSTRACT .......................................................................................................................................................... x
List of Figures ...................................................................................................................................................... iii
List of Tables ......................................................................................................................................................... v
Chapter 1: Introduction ....................................................................................................................................... 1
1.1 Motivation ............................................................................................................................................. 1
1.2 Background ........................................................................................................................................... 1
1.3 Problem Statement ................................................................................................................................ 3
1.4 Objectives of the Study ......................................................................................................................... 3
1.4.1 General Objective ......................................................................................................................... 3
1.4.2 Specific Objective ........................................................................................................................ 3
1.5 Significance of the study ....................................................................................................................... 3
1.6 Scope of the Study ................................................................................................................................ 3
1.7 Tools for DES ....................................................................................................................................... 4
1.8 Organization of the thesis...................................................................................................................... 4
Chapter 2: Literature Review .............................................................................................................................. 6
2.1 Mixed-model assembly lines (MMALs) ............................................................................................... 6
2.2 Modeling & simulation ......................................................................................................................... 7
2.3 DES based case studies ......................................................................................................................... 7
Chapter 3: Research Methodology .................................................................................................................... 10
Chapter 4: Problem Description ....................................................................................................................... 13
4.1 Data Acquisition ................................................................................................................................. 13
4.2 Product Description............................................................................................................................. 13
4.2.1 Case Company – Product Alpha................................................................................................. 13
4.2.2 Case Company – Product Delta .................................................................................................. 16
Chapter 5: Input Analysis .................................................................................................................................. 19
5.1 Discrete Probability Distribution Types .............................................................................................. 19
5.2 Case Company – Product Alpha ......................................................................................................... 20
5.3 Case Company – Product Delta .......................................................................................................... 28
Chapter 6: Process Mapping ............................................................................................................................. 35
6.1 Case Company – Product Alpha ......................................................................................................... 35
6.1.1 Present Workflow ....................................................................................................................... 35
6.1.2 Setting-up of boundary limits ..................................................................................................... 37
i
6.2 Case Company – Product Delta .......................................................................................................... 37
6.2.1 Present Workflow ....................................................................................................................... 37
6.2.2 Setting-up of boundary limits ..................................................................................................... 40
6.3 Pre-Simulation Process ....................................................................................................................... 40
6.4 Base Simulation Model ....................................................................................................................... 42
Chapter 7: Results & Discussion ....................................................................................................................... 44
7.1 Model Verification & Validation ........................................................................................................ 44
7.2 Number of replications ........................................................................................................................ 44
7.3 Results Analysis .................................................................................................................................. 45
7.3.1 Preliminary results of Product Alpha ......................................................................................... 45
7.3.2 Preliminary results of Product Delta .......................................................................................... 46
7.4 Scenarios Development ....................................................................................................................... 47
7.4.1 Scenarios for product Alpha ....................................................................................................... 48
7.4.2 Scenarios for product Delta ........................................................................................................ 54
7.5 Selection of Optimized Option ............................................................................................................ 59
7.5.1 Optimization of Product Alpha .................................................................................................. 59
7.5.2 Optimization of Product Delta.................................................................................................... 60
Conclusion ........................................................................................................................................................... 62
Recommendations............................................................................................................................................... 64
Future Work ....................................................................................................................................................... 65
REFERENCES ................................................................................................................................................... 66
Appendices .......................................................................................................................................................... 69
Appendix A: Category overview of the existing model for product Alpha ..................................................... 69
Appendix B: Category overview of the optimized model for product Alpha .................................................. 78
Appendix C: Category overview of the existing model for product Delta ....................................................... 87
Appendix D: Category overview of the optimized model for product Delta ................................................... 94
ii
List of Figures
Figure 1. Assembly line types ................................................................................................................................ 6
Figure 2. Methodology block diagram ................................................................................................................. 12
Figure 3. Existing layout of Case study industry - Product alpha ......................................................................... 14
Figure 4. Precedence diagram of Product Alpha .................................................................................................. 14
Figure 5. Existing layout of Case study industry - Product Delta ......................................................................... 17
Figure 6. Precedence diagram of Product Delta ................................................................................................... 17
Figure 7. Layout of ARENA Input Analyzer........................................................................................................ 19
Figure 8. Best fit distribution for workstation-1 ................................................................................................... 21
Figure 9. Best fit distribution for workstation-2 ................................................................................................... 21
Figure 10. Best fit distribution for workstation-3 ................................................................................................. 21
Figure 17. Best fit distribution for workstation-10 ............................................................................................... 24
Figure 27. Best fit distribution for process-1 ........................................................................................................ 28
Figure 36. Best fit distribution for process-10 ...................................................................................................... 32
iii
Figure 41. Process operations on work centers ..................................................................................................... 36
Figure 42. Process mapping of first five workstations of Product Delta .............................................................. 38
Figure 43. Process mapping of first five workstations of Product Delta .............................................................. 39
Figure 44. Configuration setup for simulation model ........................................................................................... 41
Figure 45. Process module settings of the simulation software ............................................................................ 42
Figure 46. Layout of base simulation model for product Alpha developed in ARENA Simulation Software ..... 43
Figure 47. Layout of base simulation model for product Delta developed in ARENA Simulation Software ...... 43
Figure 48. Layout of ARENA Process Analyzer (PAN) ...................................................................................... 48
Figure 49. Layout of simulation model for scenario-1 of product Alpha ............................................................. 49
Figure 50. Layout of simulation model for scenario-2 of product Alpha ............................................................. 50
Figure 51. Layout of simulation model for scenario-3 of product Delta .............................................................. 51
Figure 52. Balanced & optimized model of product Alpha .................................................................................. 52
Figure 53. Comparison chart of scenarios for product Alpha ............................................................................... 54
Figure 54. Layout of simulation model for scenario-1 for product Delta ............................................................. 55
Figure 55. Layout of simulation model for scenario-2 for product Delta ............................................................. 56
Figure 56. Layout of simulation model for scenario-3 of product Delta .............................................................. 57
Figure 57. Comparison chart of scenarios for product Delta ................................................................................ 58
Figure 58. Optimized solution for product Alpha ................................................................................................. 60
Figure 59. Suggested optimized solution for product Delta ................................................................................. 61
iv
List of Tables
Table 1. Precedence chart for assembly line of Product Alpha ............................................................................ 15
Table 2. Process operation details for Product Alpha ........................................................................................... 15
Table 3. Precedence chart for assembly line of Product Delta ............................................................................. 18
Table 4. Process operation details for Product Delta ............................................................................................ 18
Table 5. Summary of selected distributions and expressions for each workstation of product Alpha .................. 28
Table 6. Summary of selected distributions and expressions for each workstation of product Delta ................... 34
Table 7. Summary of the results for product Alpha .............................................................................................. 45
Table 8. Resource utilization of each workstation of product Alpha .................................................................... 45
Table 9. Summary of the results for product Alpha .............................................................................................. 46
Table 10. Resource utilization of each workstation of product Alpha .................................................................. 47
Table 11. Summary of the results scenario-1 for product Alpha .......................................................................... 49
Table 14. Resource utilization of balanced and optimized model for product Alpha ........................................... 52
Table 15. Summary of the results scenario-1 for product Delta ........................................................................... 55
v
vi
Chapter 1: Introduction
1.1 Motivation
Manufacturing organizations all over the world have made several changes to gain and sustain
competitive advantages in global markets. To withstand the challenges, they have adopted
various methods and techniques to optimize performance. Researchers now have the potential
to study and simulate the performance of production systems thanks to the recent development
of data-rich manufacturing settings.
The use of simulation modelling and analysis as a method for enhancing or examining process
performance is growing in popularity. It is an economical approach of assessing how well
resource allocation and various operating policies operate.
Simulation modelling has been used by the automotive, electronics, and other general
production industries to evaluate and enhance their manufacturing facilities. Numerous
manufacturing processes and assembly lines have been modelled and optimized using discrete
event simulation (DES). Since it can explicitly characterize the variation within industrial
systems using probability distributions, it is especially well suited for modelling manufacturing
systems. In order to accomplish production goals, it is thus capable of providing crucial
operational answers relating to throughput, resource allocation, utilization, and supply and
demand.
1.2 Background
In current manufacturing era, industries face various challenges like frequent changes in
product variants, fluctuation of customer demands, shorter product life cycles, and
uncertainties. Since modern manufacturing systems are complex, intricate, and require huge
fundings throughout their life cycle (from design to re-engineering), it is critical to make
decisions that will improve the overall manufacturing system's efficiency [1]. The
organizations are exploring new strategies to improve their performance to achieve the
objectives of increased production rate and improve the efficiency of the system [2]. An
assembly line design and optimization are a key retrofit method that is used to boost both the
efficiency and the efficacy of a manufacturing system, as it consists of a sequence of
workstations doing a repeated set of operations [2].
Manufacturing industries like electronics, automotive and other general production industries,
have recognized ‘simulation modelling’ as a method to evaluate and enhance their
1
manufacturing facilities [5]. To meet the challenges of product and market demands, there is a
need to develop manufacturing systems that are both cost effective as well as responsive to
changes [3]. Although a simple analysis of the system is possible but as the system's complexity
increases, it becomes vital to duplicate its behavior using computer simulation [4]. It can
explicitly characterize variance within production systems using probability distributions.
Discrete Event Simulation (DES) is a modeling and optimization technique for complex
manufacturing processes and assembly lines. DES can explicitly characterize variance within
industrial systems using probability distributions, it is highly suited for modeling
manufacturing systems. Further, it can provide answers to critical operational questions such
as production capacity, utilization, resource allocation, supply and demand, and throughput of
the system. As a result of this vital information, industries have been able to make major and
beneficial changes to their facilities in order to improve operations and reach production targets
easily [5].
Throughput is one of the most critical parameters to assess the production for design,
management and control of the complete manufacturing system [2]. Similarly, a bottleneck is
a parameter that drastically limits the system's total performance [6]. It can be assumed that the
work-in-process (WIP) inter-departure time variance (ITV) is minimal while considering the
features of a bottleneck station [7]. Generally, enhancing the performance on bottleneck may
result in higher overall system throughput as compared to non-bottleneck resources. Therefore,
analysis of the bottleneck has always gained a leading importance in manufacturing operations
and considerable research has been devoted to this area [6].
The manufacturing issues outlined in the preceding paragraphs necessitate the use of
'Simulation,' which is the generic term for the repeated operations of a real-world process or
system over time [8]. The simulation models can be classified as either physical or
mathematical. The physical model characterizes the physical aspects of a system, while the
mathematical model uses mathematical equations and symbolic notations to signify a system
such as Monte Carlo simulations, linear programming and queuing theory. The simulation
models are further segregated as being static or dynamic. Static simulation models represent a
system wherein time is a stationary variable while the dynamic models represent a system that
changes over time. In addition to above, a simulation model is deterministic if it contains no
random variables while stochastic models include one or more random variables as an input.
Finally, in a discrete model, the state variables change only at a discrete set of points with
respect to time while there is a continuous change in state variables in a continuous model [9].
2
1.3 Problem Statement
To develop and apply a comprehensive simulation model for bottleneck reduction of mixed
model assembly lines using discrete event simulation. The developed model will be capable of
identifying and eliminating queues and bottlenecks in the assembly line and application of
discrete event simulation technique will eliminate those queues and bottlenecks in the assembly
line. Ultimately leading to enhance the production capacity of the industry, reducing cycle time
and increasing the overall resource utilization. Feedback system will help to make certain
required amendments in the model to achieve the desired objective.
1.4 Objectives of the Study

1.4.1 General Objective
The general objective of this research work is to analyse, evaluate and model the assembly line
of a high-tech parts manufacturing industry with several system variables.
1.4.2 Specific Objective
a. To analyze the existing workflow of the assembly lines
b. To identify the bottlenecks and queue lengths
c. To optimize the existing workflow by reducing bottlenecks and queue lengths
d. To enhance production capacity
1.5 Significance of the study
Modeling and simulation in the field of manufacturing system and performance analysis plays
a key role in any industry, in order to compete and be productive in today’s competitive market.
This modeling and analysis can be applicable in high technology parts manufacturing industries
where there is a dire need of production capacity enhancement and improvement in overall
efficiency throughout production line. Therefore, any parts manufacturing industry, especially
high technology and precision parts manufacturing industry can benefit from this research by
adopting the core concepts and making some minor improvements to the rest of the points.
1.6 Scope of the Study
The scope of this thesis is to develop a simulation model for the mixed model assembly line of
the high-tech parts manufacturing process for several performance parameters. The simulation
model offers a low-cost tool for analyzing the system and to allocate bottlenecks in the
3
production process and recommend possible improvement for the Company. The thesis is
limited to Modeling and analysing the production line based on the actual collected data from
a well-known and reputable industry.
1.7 Tools for DES
Models developed through DES are frequently utilized to provide specific operations-related
answers. DES software from a commercial source is employed by most of the authors to
simulate these systems [5]. Latest trends in the field of modeling & simulation as introduced
various new software tools for building the simulation models. The most widely used
simulation modelling platforms include Arena, Simio, Simul8, Extensim, Flexim, NetLogo and
Repast. Arena simulation software by Rockwell Automation Systems is most used for discrete
event simulation and automation. Arena has dominated the discrete event simulation software
market for 30 years [21]. It has been used by many industries due to its applications in the
following [5]:
a. To model physical systems using DES.

b. To validate a scheduling method.
c. To assess various control strategies.
d. To reconfigure the assembly lines.
In addition to the applications mentioned above, various advantages of Arena as discussed by

[9] are listed here.
a. Applications in construction field.

b. Practices required for constructing a simulation model.
c. Accessibility and ease to understand.
d. An operative means used in academics.
Based on the advantages discussed here, Arena Simulation software has been selected as
simulation tool to address the problems of the industries under study in this research work.
1.8 Organization of the thesis
This thesis is organized into six different parts which include: introduction, literature review,
case company, simulation model development, and proposed model, and conclusion and
recommendation. In the introduction part, the objectives, problem statement, methodology etc.
will be discussed. In part two different literatures discussed related to the topic is organized. In
the third part background of the company, basic production process was discussed. In the fourth
4
part, data collection and simulation model development is done and an improved system can
be proposed. In the fifth part, the proposed models were developed and at the end part,
conclusion and recommendation will be made and finally, appendices are appended for several
data and simulation results.
Consequently, the thesis is organized into six parts. Section II presents the literature review for
mixed model assembly lines along with DES based case studies. Section III describes the
problem at hand along with process mapping depiction in Section IV. Methodology is
explained in Section V with assumptions listed in Section VI. Pre-simulation process and the
results are explained in Section VII and VIII, respectively. Finally, the conclusions drawn ends
the article.
5
Chapter 2: Literature Review
2.1 Mixed-model assembly lines (MMALs)
Mixed-model assembly lines (MMALs) are crucial in any industry because they allow
manufacturing of many products on the same line at the same time, resulting in an efficient
production flow. MMALs have a lot of advantages, but also have issues and complexities when
it comes to designing and implementing them. The mixed-model assembly line sequencing
(MMALs) problem is one of the most important operational difficulties in manufacturing
environment [10]. Using flexible machinery and polyvalent personnel to solve MMAL
difficulties can result in significant cost and time savings [2]. The layout of types of assembly
lines as depicted by [22] is shown in the figure below.
Figure 1. Assembly line types

Balancing and scheduling are the two key issues in any production line. In production
management, scheduling is a short-term decision-making question while balancing is the
medium and long-term decision-making concern [11]. It involves deciding the sequence in
6
which products arrive at the stations. Work overload, wait time and idle time at each
workstation can be affected by different model sequences of mixed-model products, resulting
in shorter or longer completion times [10].
2.2 Modeling & simulation
Layout, control techniques, equipment selection, buffer sizing, material handling, dispatching
tactics, and other parts of the manufacturing facility can be designed and modeled using
simulation software [7]. Simulation models can be used to analyze the prospective
modifications in an existing system without disrupting them and assist to evaluate the design
of a new system without having to create a real-world system. Also, such models help to
identify and manage bottlenecks in production systems [7].
Due to statistical assumptions for machine performance, the analytical methods are limited to
steady-state and long-term bottleneck discovery. Under a lot of limits and assumptions,
analytical methods can address the bottleneck discovery problem in simplified models [6]. The
complicated system of production line, that must be flexible and commonly include a product
mix, makes the job of the production planners more difficult; such production systems are well
suited to DES modelling [5].
The ARENA ® commercial software has been utilized in a following application [5]:
a. Variety of assembly applications, including modelling a four-closed-loop network
b. Evaluating multiple control policies
c. Validating an aggregated event scheduling approach
d. Confirming an analytic model
e. Creating a mixed model assembly line
f. Modelling an assembly cell.
2.3 DES based case studies
Garoma and Ofgera [2] used ARENA simulation software to create a simulation model for the
production line of a mineral water company where they reduced bottlenecks in the production
line and maximized output. As a result, manufacturing line throughput enhanced from 21.12
% to 54.03 %, and reduced cycle time reduced by 33.33 % [2].
7
A DES study of an emergency department at Moncton's Dr. Georges L. Dumont Hospital
(Canada) was conducted by [12] with a goal to decrease patient wait times that were ultimately
linked to resource availability. While also improving overall service delivery and system
throughput, several options depending on resource scenarios were also developed [12].
ARENA was described as a tool for simulating earthwork operations with the advantage of its
ease and versatility in simulating most types of models in various construction organizations
in a study by [9]. The system's cycle time was lowered by altering the cell selection rule in
research by [13] that focused on few performance metrics to discover bottlenecks using a series
of simulation experiments [13].
The problems generated during reconfiguration and their impact on system performance
measures were addressed in an extensive empirical study by [14] on manufacturing system
reconfiguration, where a reconfigurable assembly line [RAL] simulation model was created to
investigate the line's performance. The RAL simulation model produces excellent outcomes,
with a 10.48 % increase in throughput and a 93.42 % to 98.78 % improvement in worker
utilization [14].
Since absenteeism and turnover entail operator replacement, a study was conducted by [15] to
observe the impact of substituting the experienced absentees by new operators. It was found
that these new operators are more likely to be contenders for the line's bottleneck and lowering
throughput (An example revealed a 66 % reduction in throughput) [15].
The goal of another case study by [16] on a clothing production line by using ARENA
simulation program, was to construct an assembly line that would achieve maximum line
efficiency while maintaining a constant cycle time [16].
According to the findings of a study carried out by [17], simulation software (SIMUL8) was
used to analyze the COVID-19 sample collection method in a hospital drive-in. [17]. In
emergency rooms, DES was used by [18] to model operations and analyze patient flows and
throughput time. It was suggested that patient waiting times might be cut by more than 20%
[18].
Another study by [19] investigated the modelling of assembly line balance by integrating both
manual line balancing approaches and computer simulation to identify the ideal solution in the
Almeda textile sewing line in order to increase production. The output of the simulation model
raised system utilization to 0.69 % with a 58.42 %-line efficiency without incurring additional
costs [19].
8
With the predominant aim of this research in mind, it is imperative from the literature reviewed
that DES has been popularly used to solve MMAL problems. The content to follow in
subsequent sections explains the application of DES on an assembly line from a high-tech
manufacturing industry.
9
Chapter 3: Research Methodology
Different stochastic and analytical methods have been developed over the years to design and
balance assembly lines under different constraints [16]. Since computer simulation models are
effectively used, the technique of identifying bottlenecks is employed in this project while
using a DES software, ARENA ®. Usually, when a machine is inefficient and has long queue,
long-term bottlenecks arise [15].
The methodology applied in this study was divided into five broad phases. The first phase
begins with the conceptualization and understanding of the problem in the industry. The
identification of problem was followed by the preparation of template for the acquisition of
data from the industry. Several visits of the case study industries helped in imputation and
fulfilment of the required data as per the template and the multiple session with the process
owners helped to extract the relevant data. The extraction of relevant data is followed by the
analytical solutions and studying of the operations / processes performed at the workstations.
The next step is to carry out input analysis of the data. The tool used for the input analysis was
ARENA Input Analyzer. Various statistical tests are applied by this tool that include the Chi-
Square test and Kolmogorov-Smirnov test. Minimum value, maximum value, mean value,
number of intervals, degree of freedom and corresponding p-value of nine different distribution
functions are obtained and then the minimum square error among all these nine distribution
functions gives the best fit distribution function.
The second phase begins with the detailed study of the processes and operations of the
workstations. Process mapping of each workstation data is done in the ARENA simulation
software. Best fit distribution obtained in the first phase is fed into the process module of each
workstation. The existing physical system is then configured in the software by using certain
assumptions. Before running the preliminary run of the developed base simulation model,
errors must be checked in the software. If there are no errors, then run the simulation of the
base model.
The next phase of the methodology starts with the verification of the model. The results
obtained by running the base simulation model will then be verified with the existing physical
system. The results must be within the boundary limits as identified and highlighted by the
existing physical system. If the results are in commensuration with the existing physical
system, then the base simulation model is being validated. Validation of the results is followed
by the analysis of the results, which include the identification of the bottlenecks workstations,
10
queues at the workstations and their queues lengths, system throughput, cycle time and resource
utilization.
The fourth phase of the methodology is the optimization phase. The first step in this phase is
the development of different scenarios for the base simulation model. The purpose of making
and applying different scenarios is to achieve the objective of the industry by eliminating
bottlenecks and queues in the existing system. The objective of the case study industry is to
enhance system production, resource utilization, overall efficiency and reduce cycle time.
Comparison of scenarios is then carried out by using the software tool ‘ARENA Process
Analyzer’. This will lead us to identify and select the best or suitable scenario. The next step is
to optimize the simulation model using the ‘Opt Quest’ tool in the ARENA simulation
software. The pre-requisites for optimization are the selection of an objective function, controls
and responses and the constraints. The results obtained from Opt Quest will give the suggested
solution for the optimization of the simulation model.
The final phase of the study is the implementation phase. The first step is to verify that whether
the objective of the study is being achieved or not. If the objective is achieved, then results will
be transferred to the stake holders of the industry for pilot testing and performing quality check
of the suggested simulation model. Cost benefit analysis (CBA) are then carried out by the
concerned department of the case study industry. Final the model will be implemented in the
industry that will have enhanced production, improved resource utilization, better overall
efficiency of the system with reduced cycle time.
11
The applied methodology block diagram is given in Figure below.
Phase-1
Phase-2
Phase-3
Phase-4
Phase-5
Figure 2.Methodology block diagram
12
Chapter 4: Problem Description
4.1 Data Acquisition
The first step in the modelling approach is data collecting, which should be carefully organised
to target the important facts that will reflect the goals of the simulation model. The data is
acquired from a high technology parts manufacturing industry as per the template prepared for
the case analysis. The acquired data encompassed the general information related to the type
of products manufactured / assembled, total production, installed capacity of the existing
manufacturing facility, type of assembly line, number of workstations, theoretical calculations
for resource utilization and cycle time, machine parameters including nominal run speed and
its capacity, reliability data like expected time, and other similar details. Moreover, provision
of a ‘production map’ of the complete product ranging from assemblies to sub-assemblies and
components were also included as part of the received data. The analysis is done for two
products of the case industry.
4.2 Product Description

4.2.1 Case Company – Product Alpha
The case analysis for product Alpha is carried out as a validation study. The assembly line of
the product is operating for mixed model assembly line, and it is a manual assembly line.
Various assembly and manufacturing operations are performed on the assembly line according
to the product requirement and demand. Process time data for each workstation is collected.
Theoretical calculations for the resource utilizations for each workstation, average cycle time,
and other relevant findings are provided by the industry. Some highlights of the given data are
given as under:
a. Total number of workstations = 19

b. Current system production = 14 products / year
c. Cycle Time = 384 hours
d. Line efficiency = 95%
e. Rejection rate = 05%
13
The configuration layout of the assembly line of the product alpha of the case study industry is
shown in Figure-2.
Figure 3. Existing layout of Case study industry - Product alpha

The precedence diagram and precedence chart for the product Alpha is shown as under:
Figure 4. Precedence diagram of Product Alpha
14
Table 1. Precedence chart for assembly line of Product Alpha
Work Element Preceded By Average Task Time (in Hours)

1 - 41.5
2 1 55.5
3 2 33.5
4 3 70.1
5 - 111
6 - 104
7 - 99.4
8 7 25.6
9 8 77
10 4, 5, 6, 9 13
11 10 31.8
12 11 17.3
13 12 16.6
14 13 17.4
15 14 14
16 15 15.7
17 16 13.2
18 17 21.5
19 18 24.2
The details of operations performed at each workstation is shown as under:
Table 2. Process operation details for Product Alpha
S. No Station No. Process

1 A1 Assembling Operation
2 A2 “
3 A3 “
4 A4 “
5 A5 “
6 A6 “
7 A7 “
8 A8 “
9 A9 “
10 A10 “
15
11 A11 “
12 A12 Ring rolling
13 A13 Bending
14 A14 Quenching & Tempering
15 A15 Machining
16 A16 Turning & milling
17 A17 Rotary forging
18 A18 Polishing & surface finishing
19 A19 Inspection & testing
4.2.2 Case Company – Product Delta
Similar case analysis is carried out for another product of the same high technology parts
manufacturing industry for product Delta. The assembly line of this product is also operating
for mixed model assembly line, and it is a manual assembly line. Various assembly and
manufacturing operations are performed on the assembly line according to the product
requirement and market demand. Process time data for each workstation is collected.
Theoretical calculations for the resource utilizations for each workstation, average cycle time,
and other relevant findings are provided by the industry. Some highlights of the given data are
given as under:
f. Total number of workstations = 14

g. Current system production = 110 products / year
h. Cycle Time = 74.3 hours
i. Line efficiency = 95%
j. Rejection rate = 05%
The configuration layout of the assembly line of the product alpha of the case study industry is
shown in Figure-2.
16
Figure 5. Existing layout of Case study industry - Product Delta
The precedence diagram and precedence chart for the product Alpha is shown as under:
Figure 6. Precedence diagram of Product Delta
17
Table 3. Precedence chart for assembly line of Product Delta
Work Element Preceded By Average Task Time (in Hours)

1 - 11.2
2 - 12
3 - 8
4 1,2,3 12
5 4 6.5
6 - 12
7 6 6.1
8 5, 7 9
9 8 7.5
10 9 5
11 10 6
12 11 6.4
13 12 9
14 13 6.7
The details of operations performed at each workstation is shown as under:
Table 4. Process operation details for Product Delta
S. No Process No. Process

1 P1 Assembling Operation
2 P2 “
3 P3 “
4 P4 “
5 P5 “
6 P6 “
7 P7 “
8 P8 Sheet metal forming
9 P9 Laser & arc welding
10 P10 Casting
11 P11 Roll forging
12 P12 Turning & milling
13 P13 Paint & surface curing
14 P14 Inspection & testing
18
Chapter 5: Input Analysis
The data acquisition phase is followed by the analysis of the data for both the products taken
as case studies from high technology parts manufacturing industry. The tool used for input
analysis is the ‘ARENA Input Analyzer’ from ARENA simulation software. Various statistical
tests are applied by this tool that include the Chi-Square test and Kolmogorov-Smirnov test.
Minimum value, maximum value, mean value, number of intervals, degree of freedom and
corresponding p-value of nine different distribution functions are obtained and then the
minimum square error among all these nine distribution functions. This tool will help to
identify the best fit distribution that will be used while preparing simulation model in ARENA
simulation software. The layout of ARENA Input Analyzer is shown in the figure below:
Figure 7. Layout of ARENA Input Analyzer
5.1 Discrete Probability Distribution Types
Discrete probability distributions are used in various types of systems that include
manufacturing systems, defense-related systems, communication systems and transportation
systems etc. The sources randomness in the manufacturing systems include process time,
machine time to failure, machine repair time etc. In defense-related systems, arrival time of
payloads and miss distances are important. In communications systems, it includes interarrival
time of messages, message types, message lengths etc. While in transportations systems, ship
loading time and interarrival time of customers are the concerned sources of randomness. The
types of discrete probability distributions mostly used in manufacturing industries are as under:
19
a. Normal distribution
b. Uniform distribution
c. Triangular distribution
d. Beta distribution
e. Gamma distribution
f. Weibull distribution
g. Erlang distribution
h. Lognormal distribution
i. Exponential distribution
The selection of distribution function is done by carrying out analysis of the statistical data. A
normal random variable is unquestionably the most often used model. When a random
experiment is repeated, the random variable that equals the average (or overall) outcome across
all the replicates tends to follow a normal distribution as the number of repetitions increases.
A discrete random variable with a finite number of potential values and equal probabilities is
the uniform type of discrete probability distribution. The triangular distribution has the
minimum value, the maximum value and the most likely value for the data set. The Gamma
distribution is a two-parameter family of a continuous distribution that is very closely related
to exponential distribution with positive values. The Weibull distribution is frequently used to
simulate how long a physical system will last before failing. The distribution's parameters offer
a significant degree of flexibility for modelling systems where the frequency of failures rises
over time, falls over time, or stays constant over time. Erlang distribution is the extension of
the exponential distribution function with the varying number of events. The lognormal
distribution, as the name suggests, is a natural logarithm of a normal distribution. When the
shape parameters of a system allow the probability function to assume different shapes, then
the distribution of the system will be Beta distribution [23].
5.2 Case Company – Product Alpha
The data for workstation-1 is maintained into the ARENA input analyzer from the root file.
The distribution gained by running the goodness of fit test for nine different distributions for
workstation-1 is ‘Beta’ distribution with least square error of 0.009493. The distribution data
has seven intervals with four degrees of freedom. The distribution function as achieved by input
analysis is displayed as under:
20
Figure 8. Best fit distribution for workstation-1
The statistics for workstation-2 is fed into the ARENA input analyzer from the source file. The
distribution achieved by running the goodness of fit test for nine unique distributions for
workstation-2 is ‘Normal’ distribution having least square error of 0.007697. The distribution
data has five pauses with two degrees of freedom. The distribution function as obtained by
input analysis is shown as under:

The workstation-3 data is fed into the ARENA input analyzer from the source file. The division
got by running the goodness of fit test for nine different distributions for workstation-3 is
‘Gamma’ distribution with least square error of 0.004482. The distribution data takes five
intervals with two degrees of freedom. The distribution function as found by input analysis is
shown as under:

The data intended for workstation-4 is fed into the ARENA input analyzer from the source file.
The distribution that is gained by running the goodness of fit test for nine different distributions
21
for workstation-4 stands ‘Weibull’ distribution with least square error of 0.020880. This
distribution data takes five intervals through two degrees of freedom. The distribution function
as attained by input analysis is presented as under:

The workstation-5 statistics is entered into the ARENA input analyzer from the source file. The
partition which is obtained by running the goodness of fit test aimed at nine different
distributions for workstation-5 is known as ‘Beta’ distribution having least square error of
0.012307. The distribution data takes six intervals with three degree of freedom. The
distribution function is shown as under:

The information for workstation-6 is written into the ARENA input analyzer from the source
folder. The distribution found by operating the goodness of fit test for nine different divisions
for workstation-6 is found ‘Normal’ with least square error of 0.019495. The data has four
intervals having one degree of freedom and the distribution function obtained by input
investigation is shown as under:
22
The figures for workstation-7 is recorded into the ARENA input analyzer from the source. By
running the goodness of fit test for nine different distributions for workstation-7, the gained
distribution is ‘Normal’ distribution having least square error of 0.008399. The distribution
statistics has five intervals using two degrees of freedom. The obtained distribution function is
displayed as under:

The information for workstation-8 is recorded into the ARENA input analyzer from the basic
folder. By proceeding the goodness of fit test for nine different distributions for workstation 8,
the gained distribution is called ‘Beta’ distribution with least square error of 0.006803. The
statistics has six intervals with three degrees of freedom. The distribution results found by input
analysis are presented as under:

The facts taken from the source file for workstation-9 has been entered into the ARENA input
analyzer. The goodness of fit test was used for nine different distributions for workstation-9
which showed ‘Normal’ distribution with least square error of 0.009015. The distribution
statistics has five intervals with two degrees of freedom. The distribution function is displayed
as under:
23
The information for workstation-10 is entered into the ARENA input analyzer from the source.
By means of the goodness of fit test for nine different distributions for workstation-10 we
achieved ‘Beta’ distribution with least square error of 0.014298. The distribution statistics got
six intervals with three degrees of freedom. The distribution obtained by analysis is shown as
under:

The data aimed at workstation-11 is entered into the ARENA input analyzer taken from the
source file. The distribution found by using the goodness of fit test for nine distinct distributions
for workstation-11 is ‘Beta’ distribution having least square error of 0.011425. The distribution
statistics has six intervals with three degrees of freedom. The distribution function is displayed
as under:

The workstation-12 data taken from the source file is fed into the ARENA input analyzer. After
running the goodness of fit test for nine different distributions for workstation-12, the obtained
24
distribution is ‘Beta’ distribution with least square error of 0.047453. The distribution facts
take six intervals with three degrees of freedom. The distribution function as gained by input
examination is shown as under:

The data for workspace-13 is taken into the ARENA input analyzer from the source folder. The
distribution acquired by functioning the goodness of fit test for nine distinct distributions for
workstation-13 is called ‘Normal’ distribution having least square error of 0.009904. The
statistical data has five intervals using two degrees of freedom. The distribution function as
gotten by input analysis is shown below:

The information for workstation-14 is fed into the ARENA input analyzer taken from the
source. The distribution which is gained by running the goodness of fit test for nine different
distributions for workstation-14 is known as ‘Triangular’ distribution having least square error
of 0.013696. The distribution data uses five intervals with three degrees of freedom. The
distribution function as found by input analysis is presented as under:
25
The workstation-15 statistics entered in the ARENA input analyzer taken from the source file.
The distribution attained by running the goodness of fit test for nine dissimilar distributions for
workstation-15 is said to be ‘Lognormal’ distribution with least square error of 0.033498. The
distribution information has four intervals with one degree of freedom. The distribution
function is displayed as under:

The data for workstation-16 is fed into the ARENA input analyzer from the source file. The
distribution obtained by running the goodness of fit test for nine different distributions for
workstation-16 is ‘Triangular’ distribution with least square error of 0.040655. The distribution
data has five intervals with three degrees of freedom. The distribution function as obtained by

The data for workstation-17 is fed into the ARENA input analyzer taken from the source. The
distribution found by using the goodness of fit test for nine different areas for workstation-17
is called ‘Beta’ distribution having least square error of 0.025439. The data has six intervals
with three degrees of freedom. The distribution function as gained by analysis is shown as
under:
26
The workstation-18 statistics is fed into the ARENA analyzer from the source file. The
distribution attained by running the goodness of fit test at nine different distributions for
workstation-18 is known as ‘Weibull’ distribution with least square error of 0.009493. The
distribution information has three intervals with zero degree of freedom. The distribution
purpose is shown as under:

The statistics for workstation-19 is entered into the ARENA input analyzer from the basic file.
By running the goodness of fit test for nine different distributions, the obtained distribution for
workstation-19 is called ‘Normal’ distribution with least square error of 0.010114. The
distribution data takes five intervals with two degrees of freedom. The distribution function
found by input analysis is presented as under:

The summary of distribution expression obtained by the analysis of the input for all the 19
workstations of the assembly line for product alpha is given in the following table:
27
Table 5. Summary of selected distributions and expressions for each workstation of product
Alpha
No. Operations Distribution Expression

1 Assembling Operation Beta 11+2*BETA (1.03, 1.18)
2 “ Normal NORM (41.5, 0.629)
3 “ Gamma 54.1+GAMM (0.26, 5.38)
4 “ Weibull 32+WEIB (1.66, 2.63)
5 “ Beta 46.3 + 3.68 * BETA (1.61, 1.47)
6 “ Normal NORM (111, 0.734)
7 “ Normal NORM (104, 0.693)
8 “ Beta 96 + 3.99 * BETA(1.96, 1.58)
9 “ Normal NORM (25.6, 0.7)
10 Casting Beta 18.1 + 1.9 * BETA (1.37, 1.54)
11 Heating Beta 76 + 3 * BETA (1.31, 1.45)
12 Ring rolling Beta 12 + 2 * BETA (1.47, 0.949)
13 Bending Normal NORM (31.8, 0.579)
14 Quenching & Tempering Triangular TRIA (16.3, 17.3, 19)
15 Machining Lognormal 15.5 + LOGN (1.18, 0.565)
16 Turning & milling Triangular TRIA (17, 17.4, 18)
17 Rotary forging Beta 2.4 + 0.32 * BETA (1.89, 1.51)
18 Polishing & surface finishing Weibull 10 + WEIB (0.569, 5.05)
19 Inspection & testing Normal NORM (15.7, 0.131)
5.3S Case Company – Product Delta
The data for process-1 of the product Delta is entered into the ARENA input analyzer taken
from the source file. The distribution which is gained by running the goodness of fit test for
nine different distributions for process-1 is called ‘Beta’ distribution with least square error of
0.009053. The distribution statistics has six intervals with three degrees of freedom. The
distribution function is shown below:
Figure 27. Best fit distribution for process-1
28
The process-2 statistics is entered into the ARENA input analyzer from the source. The
distribution which is obtained by running the goodness of fit test for nine different distributions
for process-2 is ‘Normal’ distribution using least square error of 0.011798. The distribution
data consumes five intervals with two degrees of freedom. The distribution function as found
by analysis is shown as under:

The facts for process-3 is fed into the ARENA analyzer taken from the source file. By running
the goodness of fit test for nine different distributions, the distribution found for process-3 is
‘Weibull’ distribution having least square error of 0.009347. The data uses four intervals with
one degree of freedom. The distribution function as gained by input analysis is shown as under:

The process-4 data is fed into the ARENA analyzer from the source file. The distribution
acquired by using the goodness of fit test for nine different distributions for process-4 is called
‘Triangular’ distribution with least square error of 0.006111. The data takes five intervals with
three degrees of freedom. The distribution assignment task as achieved by input analysis is
shown as under:
29
The statistics for process-5 is taken into the ARENA input analyzer from the source. The
distribution which is reached by running the goodness of fit test at nine different distributions
for process-5 is known as ‘Normal’ distribution with least square error of 0.013329. The
distribution data takes five intervals with two degrees of freedom. The distribution as obtained
by analysis is shown as under:

The numbers for process-6 is entered into the ARENA input analyzer from the basic file. By
running the goodness of fit test for nine different distributions for process-6 the obtained
distribution is ‘Uniform’ distribution using least square error of 0.027543. The distribution data
takes seven intervals with six degrees of freedom. The distribution function as found by input
analysis is presented as under:
30
The source file is used to feed the data for process-7 into the ARENA input analyzer. The
"Gamma" distribution, which has a least square error of 0.007351, is the distribution that
process-7's goodness of fit test for nine alternative distributions produced. There are five
intervals with two degrees of freedom in the distribution data. Following is an illustration of
the distribution function as determined via input analysis:

The source file is used to provide the data for process-8 into the ARENA input analyzer. When
nine distinct distributions were subjected to the goodness of fit test for process-8, the "Normal"
distribution with a least square error of 0.006503 was the result. Five intervals with two degrees
of freedom are present in the distribution data. The distribution function that input analysis
produced is displayed as follows:

The data for process-9 is read from the source file and fed into the ARENA input analyzer.
Running the goodness of fit test on nine different distributions for process-9 yielded the
'Normal' distribution with a least square error of 0.014408. There are five intervals with two
degrees of freedom in the distribution data. The distribution function derived from input
analysis is shown below:
31
The data for process-10 is fed into the ARENA input analyzer from the source file. The
process-10 is ‘Triangular’ distribution with least square error of 0.018222. The distribution
data has five intervals with three degrees of freedom. The distribution function as obtained by

The ARENA input analyzer receives data for process-11 from the source file. The distribution
obtained by running the goodness of fit test on nine different distributions for process-11 is a
'Triangular' distribution with a least square error of 0.028279. The distribution data consists of
five intervals with three degrees of freedom. The distribution function obtained through input
analysis is shown below:

The data for process-12 is read from the source file and supplied into the ARENA input
analyzer. The 'Normal' distribution with a least square error of 0.074087 was found by
32
conducting the goodness of fit test on nine different distributions for process-12. There are five
intervals with two degrees of freedom in the distribution data. The distribution function derived
from input analysis is presented below:

The data for process-13 is read from the source file and supplied into the ARENA input
analyzer. Running the goodness of fit test on nine different distributions for process-13 yielded
the 'Normal' distribution with a least square error of 0.010645. There are five intervals with
two degrees of freedom in the distribution data. The distribution function derived from input
analysis is presented below:

The data for process-14 is fed into the ARENA input analyzer from the source file. The
process-14 is ‘Normal’ distribution with least square error of 0.033336. The distribution data
has five intervals with two degrees of freedom. The distribution function as obtained by input
analysis is shown as under:
33
The summary of distribution expression obtained by the analysis of the input for all the 19
workstations of the assembly line for product alpha is given in the following table:
Table 6. Summary of selected distributions and expressions for each workstation of product
Delta
No. Operations Distribution Expression

1 Assembling Operation Beta 11 + 2 * BETA (1.05, 1.2)
2 “ Normal NORM (12, 0.413)
3 “ Weibull 6 + WEIB (2.36, 3.72)
4 “ Triangular TRIA (8, 9.01, 10)
5 “ Normal NORM (6.05, 0.403)
6 “ Uniform UNIF (11, 13)
7 “ Gamma 5.47 + GAMM (0.131, 4.2)
8 Sheet metal forming Normal NORM (9.02, 0.414)
9 Laser & arc welding Normal NORM (6.05, 0.435)
10 Casting Triangular TRIA (5, 6, 7)
11 Roll forging Triangular TRIA (5, 6, 7)
12 Turning & milling Normal NORM (6.04, 0.409)
13 Paint & surface curing Normal NORM (9, 0.451)
14 Inspection & testing Normal NORM (6.07, 0.396)
34
Chapter 6: Process Mapping
The input analysis phase is followed by the process mapping of the current physical system of
high technology components manufacturing assembly lines. This step starts with mapping the
current process of the case study industry, then setting up boundary constraints and establishing
the entire physical system with software.
6.1 Case Company – Product Alpha

6.1.1 Present Workflow
The existing assembly line of product Alpha comprises of 19 workstations. Assembly

process operations are performed on the first eleven workstations. Whereas the next eight
workstations include the manufacturing process operations. The process operations from
workstation A1 to A9 are performed under one shed, in four work centers. The activities of
these workstations are grouped basing the techniques of group technology. These work centers
are processing parallel to each other. Activities performed on workstations A1 to A4 are
sequential activities, performed at work center no.1. The set of activities at workstation A1 are
sequential to each other and the average process lead time for workstation A1 is 41.5 hours.
The sub-assembly is then transferred to workstation A2, where the activities are performed
sequential, and the average process lead time for workstation A2 is 55.5 hours. The activities
at workstation A3 & A4 are performed parallel and the average process lead time for these
workstations is 33.5 hours and 70.5 hours respectively. There is only one workstation A5 &
A6 in work center no.2 and work center no.3 respectively. The set of activities performed on
these two workstations are also sequential, and the average process lead time for these stations
are 111 hours and 104 hours. At work center no.4, the workstations A7 to A9 are also sequential
to each other. The activities at workstation A7 and A9 are sequential, and their average process
lead time is 99.4 hours and 77 hours. Whereas the activities performed at workstation A8 are
sequential to each other, and their average process lead time is 25.6 hours. The layout of
activities performed at work center no.1 to 4, as per the model prepared in the simulation
software is shown as under:
35
Work center 1
Work center 2
Work center 3
Work center 4
Figure 41. Process operations on work centers

The assembled sub-products from workstations A4, A5, A6 & A9 will then again go through
assembling process operations at workstation A10 & A11. The set of activities at both the
workstations can be processed parallel to each other in their respective workstation. The
average process lead times for workstations A10 & A11 are 13 hours and 31.8 hours
respectively.
After completion of the assembling process operations, the product will then undergo various
manufacturing process operations at further eight workstations. At workstation A12, ring
rolling manufacturing process is carried out to create a sub-part of the product that will then be
joined with the final product. The average process lead time for the workstation A12 is 17.3
hours. The sub-part obtained from the workstation A12 will then undergo heating and then
bending operation at workstation A13. The average process lead time for this workstation is
16.6 hours.
The sub-product will then undergo quenching and tempering under the workstation A14. These
activities are sequential to each other, and the average process lead time for workstation A14
is 17.4 hours. After going through quenching and tempering operation at workstation A14, the
sub-product will then be moved to workstation A15, where machining of the sub-product will
be done. The average process lead time of workstation A15 is 14 hours.
Turning and milling are the next manufacturing process operations that will be performed on
the sub-product on the workstation A16. Both these manufacturing processes will be done
sequential to each other. The average process lead time for the workstations is 15.7 hours.
36
Turning & milling operations are then followed by rotary forging manufacturing operation
performed at workstation A17. The average process lead time for rotary forging operation is
13.2 hours.
The sub-product is then moved for the final phase of the product manufacturing that begins
with polishing and then providing surface finish at workstation A18, and finally inspecting the
complete product and testing at the last workstation A19. The average process lead times at
each workstation is 21.5 hours and 24.2 hours, respectively.
6.1.2 Setting-up of boundary limits
To proceed with the simulations, the existing system hardware needed to be configured
with the software to obtain correct results. For that matter, some assumptions were considered
which are listed as under:
a. The assembly line configuration is a dedicated and straight with sequential stations
b. No breakdown / failures occur in the system
c. No product is rejected during the complete process
d. The system timing starts from zero seconds
e. The arrival of 2nd unit is after 17 working days
f. Working days per annum are 260
g. Working hours per day are 6.5 hours
h. Operators work in one shift per day
i. All operators work at 100% efficiency
j. Supply chain is primed & dedicated human resource is available
k. Requisite instruments and infrastructure are available.
6.2 Case Company – Product Delta

6.2.1 Present Workflow
The existing assembly line of product Delta comprises of 14 workstations. There is a mix and
match of parallel and sequential operations performed at different workstations of the complete
assembly line. The process operations performed at workstations from P1 to P7 are all assembly
operations, whereas, from workstations P8 to P14, various manufacturing process operations
are performed. The process operations at workstations P, P2 & P3 are parallel to each other.
37
While the set of activities that are performed on these three individual workstations P1, P2 &
P3 are sequential in order. The average process lead times of workstations P1, P2 & P3 are
11.2 hours, 12 hours and 8 hours, respectively. The parts obtained from these three
workstations are then assembled to make one sub-part at workstation P4. The activities
performed at this workstation parallel to each other and the average process lead time of this
workstation is 12 hours. The sub-product is then transferred to the next workstation P5, where
the various sequential activities are performed. The average process lead time of this
workstation is 6.5 hours. The layout of process mapping for first five workstations is shown in
the figure below:
Figure 42. Process mapping of first five workstations of Product Delta

The process operation at workstation P6 starts in parallel to workstations P1, P2 & P3. The set
of activities performed at this workstation are sequential in order, and the average process lead
time for this workstation is 12 hours. The part obtained from workstation P6 is then transferred
to the next workstation P7, where different assembly operations is being performed. The
average process lead time for the workstation P7 is 6.1 hours. The layout of process mapping:
of 7 workstations is shown in the figure below:
38
Figure 43. Process mapping of first five workstations of Product Delta
The assembly process operations are then followed by various manufacturing process
operations at workstations from P8 to P14. At workstation P8, the output from P5 and P7 is
first assembled and then undergoes process operation of sheet metal forming. The average
process lead time for this workstation is 9 hours. As the product moves forward to workstation
P9, it undergoes laser and arc welding process. The average process lead time for this
workstation is 7.5 hours. At workstation P10, casting of sub-product then be carried out. The
average process lead time for this workstation is 5 hours.
The sub-product is then moved to workstation P11, where roll forging of the part will be done.
The average process lead time at workstation P11 is 6 hours. Turning and milling are then
manufacturing process operations performed sequentially at workstation P12. The total average
process lead time of both the process operations is 6.4 hours, respectively. After going through
turning and milling process operations at workstation P12, the product is then moved for paint
and surface coating operations at workstation P13. The average process lead time for this
workstation is 9 hours.
The sub-product is then moved for the final phase of the product manufacturing that begins
with complete inspection of the manufactured product and then carrying out pilot testing before
delivery of the product. The average process lead time at this workstation is 6.7 hours.
39
6.2.2 Setting-up of boundary limits
The software and hardware of the current system have to be setup in order to move
forward with the simulations and produce accurate results. In that regard, the following
hypotheses were considered:
a. Dedicated and straight assembly line with sequential stations
b. No breakdown / failures occur in the system
c. Product rejection rate is not catered
d. System warm-up time is not included
e. The arrival of 2nd unit is after 2 working days
f. Working days per annum are 240
g. Working hours per day are 6 hours
h. Operators work in single shift per day
i. All operators work at 100% efficiency
j. Supply chain is primed & dedicated human resource is available
k. Requisite instruments and infrastructure are available.
l. Line efficiency is 95%
6.3 Pre-Simulation Process
ARENA simulation software has integrated tools, statistical analysis, built in material handling
system constructs, flow chart modelling environment and Add-on ‘OPTQUEST’ (a built-in
feature to optimize any system using the objective function and a set of constraints). These
features are valuable to model and analyze different concepts in a complex manufacturing
system.
ARENA simplifies the unpredictability in the system (which typically is related to arrival and
flow of entities through the system). ARENA also delivers automatic calculation of several
simulation characteristics for each queue such as average waiting time, average queue length,
and resource utilization [15].
To provide a logic of system behavior, a series of blocks are initially established when
developing a model utilizing the pre-defined library of objects. In order to create random events
40
and numbers, numerical attributes are then established for each block and probable patterns are
described. The setup of the animation for entities, resources, conveyers, and environment
components follows. After the model has been cleaned up from the bugs, a simulation
experiment of any length is run. The resulting report for any experiment includes all the
numerical model parameters together with statistical analysis components.
The configuration of the physical system with the software essential before the development
of the model. The warmup time for the case study industry is negligible. The minimum number
of replications selected are 10 and the maximum replication length for each replication is 260
days in a year. There is a single shift of 6.5 hours (excluding breaks) operating in the industry.
The layout of the configuration settings is shown in figure below.
Figure 44. Configuration setup for simulation model

The distribution expressions obtained from the input analysis of each workstation of both the
products are then fed into the process module of the simulation software. The layout of the
process model settings of the simulation software in the figure below.
41
Figure 45. Process module settings of the simulation software
6.4 Base Simulation Model
Process mapping of the complete physical system in the software and configuration of the
physical system with the software is then followed by the checking of the errors in the
simulation software. As the error check is passed and the system has no errors, preliminary run
of the base simulation model will then be carried out. Results obtained are discussed in the next
section, whereas the layout of the base simulation models for both the products with basic
modules using ARENA ® simulation software version 14.0 is shown in the figure below.
42
Figure 46. Layout of base simulation model for product Alpha developed in ARENA
Simulation Software
Figure 47. Layout of base simulation model for product Delta developed in ARENA
Simulation Software
43
Chapter 7: Results & Discussion
7.1 Model Verification & Validation
In the field of simulation, model verification and validation (V&V) is fundamental. It serves
as a foundation for future investigations and scenario planning. Verification is the process of
determining the accuracy with which the conceptual description and specifications are
implemented in the computer programme of the computerised model. Model validation
certifies that the simulation model, within its regions of applicability, accurately represents the
system with the modelling objectives [20].
Any simulation run's reliability primarily based on how accurately the model can represent
reality. A simulation model must be validated and verified to be free of any system failures,
reflect the "as-is" situation, and function as planned. By running the simulation models, one
should be able to derive reliable and agreeable conclusions from the outcomes. Verification is
the method to make sure the model is operating in accordance with the modelling assumptions,
and it provides the answer to the question of how the model was developed correctly. Along
with the model's accuracy, of course. The process of confirming that the model's behaviour
corresponds to the behaviour of the actual system is known as validation. The model is
constructed based on the system's current physical system in order to make it look as real as
possible overall. The actual yeary production 12 products per year for product Alpha and 110
products per year for product Delta. When we compare the throughput of the actual production
rates with the simulation model run results, it is ascertained that the simulation model's
production rate is similar to the industry’s actual production rate per year. The simulation
model is therefore verified. By changing the model's parameters and examining the updated
model's run outcomes, modifications were made to the existing model [2].
7.2 Number of replications
The flow path of the variants is checked using a step-by-step running mode, trace, and
animation approaches. Several simulation trial runs were conducted with a range of input
parameter values to evaluate the model's output. The constructed base simulation model must
accurately reproduce the real system in all aspects with an appropriate level of confidence for
the model to be considered validated. This study uses a terminating discrete event simulation,
therefore the number of replications needed to end the simulation model is required. To begin
with the minimum number of replications, 10 replications are selected to run the preliminary
base simulation model in order to validate the results.
44
7.3 Results Analysis
7.3.1 Preliminary results of Product Alpha
The results obtained by simulating current process flow showed that the total production
capacity would be 14 units per year provided the 2nd unit arrived at the assembly line after 17
days, and the average cycle time is 651 hours. In this case, WIP inventory existed at different
stations and there was a queue at workstation A2 and workstation A4 while performing
different set of assembling activities. The queues lengths as identified by the simulation run for
the workstation A2 & A4 are 195 hours and 140 hours respectively. This further implies that
both these workstations will be choked, and bottleneck will be created here. Hence, this
bottleneck needs to be reduced.
The results of the base simulation model run have identified that there are 10 workstations that
are being under-utilized. Instantaneous utilization and scheduled utilization of all the resource
workstations have been calculated. Moreover, various other system variables like value added
time, non-value-added time, wait time, transfer time, work in process inventory, number in,
number out and number busy can also be analyzed from the results.
Above results for the existing system can further be used to optimize the existing process flow
with an objective to enhance the production capacity of the system. The summary of the results
obtained by using ARENA ® simulation software version 14.0 is given in Table 7 whereas the
scheduled utilization (%) of each source assembly section is given in Table 8.
Table 7. Summary of the results for product Alpha
Category Units
Number Out 14
Minimum cycle time (in hours) 384.66
Queue Time (in hours) A2 195
A4 140
Table 8. Resource utilization of each workstation of product Alpha
Resource Workstation Scheduled Utilization (%)

A1 98.73
A2 97.56
A3 57.42
A4 92.29
A5 99.97
45
A6 94.20
A7 89.03
A8 22.68
A9 65.22
A10 12.46
A11 30.07
A12 15.99
A13 14.76
A14 15.51
A15 12.90
A16 13.93
A17 11.72
A18 18.00
A19 19.93
7.3.2 Preliminary results of Product Delta
According to the outcomes of modelling the existing process flow, 110 units could be produced
annually if the second unit arrived at the assembly line after two days and the average cycle
time is 202 hours. In this instance, WIP inventory was present at several stations, and
workstations P1, P2, P4, and P6 all had queues going while doing various assembling tasks.
For the workstations P1, P2, P4, and P6, the simulation run indicated queue lengths of 55.52
hours, 59.44 hours, 73.50 hours, and 58.58 hours, respectively. This suggests even more that a
bottleneck will form here, choking both of these workstations. Therefore, it is necessary to
eliminate this bottleneck.
Five workstations are being underutilized, according to the results of the base simulation model
run. All of the resource workstations scheduled, and instantaneous utilization rates have been
determined. The data can also be used to assess several other system variables, including value
added time, non-value-added time, wait time, transfer time, work-in-process inventory, number
in, number out, and number busy.
The data for the current system can be further used to optimize the current process flow with
the goal of increasing the system's production capacity. Table 7 provides a summary of the
results from the ARENA ® simulation software version 14.0, and Table 8 lists the scheduled
utilization (%) for each source assembly area.
Table 9. Summary of the results for product Alpha
46
Category Units
Number Out 110
Queue Time (in hours) P1 55.52
P2 59.44
P4 73.50
P6 58.58
Table 10. Resource utilization of each workstation of product Alpha

P1 0.1475
P2 0.1283
P3 0.1338
P4 0.0200
P6 0.0807
P6 0.1208
P7 0.1015
P8 0.1520
P9 0.1704
P10 0.1154
P11 0.2746
P12 0.0804
P13 0.0754
P14 0.1189
7.4 Scenarios Development
The developed model can be used to test out a variety of ideas, plans, and "what-if" situations
that could be too risky, unsafe, uncertain and expensive to use in real life. The model's virtual
system reacts just as the real system would, illustrating how the proposed change will affect
your throughput, cycle time, and profit margins. This reduces the risk of improper
implementations by giving us the assurance to choose and use those solutions that have been
shown to work. The end result will be the maximum operating effectiveness and ideal line
performance. The key issues in any manufacturing industry are processing times, system total
cost, total product output, and utilization of each machine; therefore, various scenarios are
constructed to achieve the desired objective of the industry. The development of the scenarios
is mostly an intuitive process to improve the system performance based on the expert
47
knowledge. Analysis and evaluation of the output is done using the tool available in ARENA
® simulation software is the ARENA Process Analyzer (PAN) tool. The layout of the PAN is
as shown in the figure below.
Figure 48. Layout of ARENA Process Analyzer (PAN)

7.4.1 Scenarios for product Alpha
Three different scenarios for the product Alpha were developed. The prime objective of the
proposed scenarios was to eliminate bottlenecks and queues in the existing system, enhance
the system production capacity and minimize the resource utilization. However, there are
physical limitations that create bounds in the development of the scenarios. In this case study,
the bottleneck stations are at workstation A2 & workstation A4. Developed scenarios are as
under:
a. Scenario-1: Parallel station at workstation A2

b. Scenario-2: Parallel stations at workstation A2 & A4
c. Scenario-3: Hybrid model
7.4.1.1 Scenario-1: Parallel station at workstation A2
As the focus of this case study remains the removal of bottlenecks at workstation A2 & A4. In
this scenario, there will be parallel working at workstation A2. Implementation of this scenario
will reduce the bottleneck at the workstation A2, however, the queue at the workstation A4
48
will increase from 140 hours to 219 hours. The layout of the developed simulation model for
this scenario is shown in the figure below.
Figure 49. Layout of simulation model for scenario-1 of product Alpha

The results obtained from this scenario shows that the system production capacity is enhanced
by 1 product and reached a figure of 15 products per annum. The overall utilization of the
resource workstations is not much affected by implementation of this scenario. Hence, there
still exists the underutilized workstations in the assembly line. The overall summary of the
results obtained from this scenario is shown in the table below.
Table 11. Summary of the results scenario-1 for product Alpha
Category Units
Number Out 15
Queue Time (in hours) A2 -
A4 219
7.4.1.2 Scenario-2: Parallel station at workstation A2 & A4
Scenario-1 depicts that the queues at the bottleneck workstation A2 has been eliminated while
the queue for bottleneck workstation A4 is more increased, as the load is then transferred to
the next bottleneck station. Therefore, parallel working only at workstation A2 will not be
49
enough. This scenario caters for the parallel stations at workstations A2 & A4 at the same time.
Implementation of this scenario will now eliminate the bottlenecks at both the workstations A2
& A4. The queue at the workstation A2 & A4 will now be negligible. The layout of the
developed simulation model for this scenario is shown in the figure below.
Figure 50. Layout of simulation model for scenario-2 of product Alpha

The results obtained from this scenario shows that the system production capacity is more
enhanced by 2 product and reached a figure of 17 products per annum. The overall utilization
of the resource workstations is not much affected by implementation of this scenario. Hence,
there still exists the under-utilized workstations in the assembly line. The overall summary of
the results obtained from this scenario is shown in the table below.
Category Units
Number Out 17
A4 -
7.4.1.3 Scenario-3: Hybrid Model
Scenario-2 has helped to eliminate the bottlenecks at the workstations A2 and A4. However,
the target set for the desired production of 20 products has still not achieved. It leads to develop
another scenario to achieve the set production target. This scenario encompasses the
development of hybrid model which includes the parallel stations at workstations at A2 & A4
and also the reduction in inter-arrival time i.e., the arrival of the raw materials to the initiating
50
workstations for the 2nd product and onwards. By implementation of this scenario, there will
be no bottleneck in the complete system. The layout of the developed simulation model for this
scenario is shown in the figure below.
Figure 51. Layout of simulation model for scenario-3 of product Delta

The results obtained from this scenario shows that the system production capacity is enhanced
to a figure of 19 products per annum. The overall utilization of the resource workstations is not
much affected by implementation of this scenario. Hence, there still exists the under-utilized
workstations in the assembly line. The overall summary of the results obtained from this
scenario is shown in the table below.
Category Units
Number Out 15
A4 -
The next step after the development of the scenarios is to check whether the assembly line is
balanced or not. It is evident from the results obtained by implementation of the scenarios that
the assembly line is still under-utilized. To enhance the overall utilization of the assembly line
of case under study, the technique of assembly line balancing is applied. The advantages
attained by implementation of assembly line balancing technique are as under:
a. Distribute workload evenly
b. Ensure maximum utilization
51
c. Improve overall efficiency
d. Reduce inventory waste
e. Absorb internal and external irregularities
f. Reduce production costs and increase profits
By applying assembly line balancing method, the resources with similar practices were grouped
into one workstation together using the procedure of group technology. The eight under-
utilized workstations were grouped into four workstations. The overall utilization of the
complete system under study was enhanced by applying the assembly line balancing method
by 8.74% and the system production is enhanced to 20 products per annum. The simulation
model and the utilization of each workstation as obtained from the simulation results are shown
in the figure followed by the table below.
Figure 52. Balanced & optimized model of product Alpha

Table 14. Resource utilization of balanced and optimized model for product Alpha

A1 98.73
A2 97.56
A3 57.42
A4 92.29
52
A5 99.97
A6 94.20
A7 89.03
A8 22.68
A9 65.22
A10 12.46
A11 30.07
A12 15.99
A13 14.76
A14 15.51
A15 12.90
A16 13.93
A17 11.72
A18 18.00
A19 19.93
7.4.1.4 Comparison of the scenarios for product Alpha
The scenarios developed for product Alpha are compared using ARENA Process Analyzer
(PAN). The comparison analysis of the scenarios for the product Alpha shoes that the
production capacity of the system has been enhanced from 14 products to 20 products annually
with an increase of 30%, the cycle time of the complete system was reduced from 651 hours to
357 hours with a decrease of 45.16% and the overall resource utilization of the complete system
was increase by 8.74%. The comparison chart of the developed scenarios for product Alpha is
shown in the figure below.
53
Optimization Scenarios
700
600
TIME (HOURS)
500
400
300
200
100
0
Base Model Scenario 1 Scenario 2 Scenario 3 Optimized
A2. Queues 195 2.5 0 0 0
A4. Queues 140 219 0 0 0
Product Out 14 15 17 19 20
Cycle Time 651 650 488 425 357
Utilization 46.21 47.4 49.19 46.5 54.94
Figure 53. Comparison chart of scenarios for product Alpha

7.4.2 Scenarios for product Delta
The scenarios developed for the case under study of product Delta also encompasses three
different scenarios with the key objectives to eliminate bottlenecks and queues in the existing
system, enhance the system production capacity and minimize the resource utilization. Similar
to the product Alpha, the physical constraints are the bounds in developing the scenarios. In
this case study, the bottleneck stations are at workstation P1, P2, P4 & P6. Developed scenarios
are as under:
a. Scenario-1: Parallel station at workstation P1, P2, P4 & P6

b. Scenario-2: Parallel stations at workstation P1, P2, P4, P6, P9 & P10
c. Scenario-3: Hybrid model
54
7.4.2.1 Scenario-1: Parallel station at workstation P1, P2, P4 & P6
The elimination of bottlenecks at different workstations is still the main goal of this case study.
At workstation P1, P2, P4 & P6, parallel work will be taking place in this scenario. The
bottlenecks at these workstations will be lessened if this scenario is implemented, however the
wait time at workstation P9 & P10 would rise and create the new bottlenecks. The picture
below depicts the design of the simulation model created for this situation.
Figure 54. Layout of simulation model for scenario-1 for product Delta
According to the findings from this scenario, the system's production capacity remains
unchanged. Implementing this scenario has minimal impact on the overall use of the resource
workstations. The table below provides a general summary of the outcomes by implementing
this scenario.
Table 15. Summary of the results scenario-1 for product Delta
Category Units
Number Out 110
Queue Time (in hours) P1 -
P2 -
P4 -
P6 -
P9 40.96
P10 40.96
55
7.4.2.2 Scenario-2: Parallel station at workstation P1, P2, P4, P6, P9 & P10
The scenario-1 has eliminated the bottlenecks at workstation P1, P2, P4& P6, however, created
new bottlenecks at workstations P9 & P10. Therefore, scenario-2 will be focusing on the
removal of the bottlenecks created at workstations P9 & P10. The queues at the bottleneck
workstations will bottleneck will become less congested if this plan is implemented. The figure
below displays the simulation model created for this situation.
Figure 55. Layout of simulation model for scenario-2 for product Delta
The outcome of this scenario indicates that the system's production capacity increased by 14
products, to a total of 124 products annually. Implementing this scenario has minimal impact
on the resource workstations' overall consumption. The table below displays a summary of all
the findings from this scenario.
Category Units
Number Out 124
Value added Time (in hours) 74.69
P2 -
P4 -
P6 -
P9 -
P10 -
56
7.4.2.3 Scenario-3: Hybrid Model
The scenarios developed so far for this case study has helped to lessen the congestion at
bottleneck workstations P1, P2, P4, P6. P9 and P10. Despite being set, the objective of
generating 150 products has not yet been achieved. To achieve the production objective, it
motivates the development of a new scenario, keeping the physical limitations of the industry
in mind. The establishment of a hybrid model that includes parallel stations at the same
workstations as done in scenario-2 i.e., workstation P1, P2, P4, P6, P9 & P10, as well as the
reduction of the process in workstations P1, P2, P4 & P6. The process time reduction is done
in consultation with the process owners of the industry under study. If this scenario is put into
practice, there won't be a bottleneck throughout the entire system. The simulation model that
was developed for this scenario is shown in the image below.
Figure 56. Layout of simulation model for scenario-3 of product Delta

The outcome of this scenario indicates that the system's production capacity more enhanced by
26 products, to a total of 150 products annually. The overall resource usage of the resource
workstations is not significantly affected by implementing this scenario. The table below
displays a summary of all the findings from this scenario.
Category Units
Number Out 124
Value added Time (in hours) 72.51
P2 -
57
P4 -
P6 -
P9 -
P10 -
7.4.2.4 Comparison of the scenarios for product Delta
The scenarios developed for product Delta are compared using ARENA Process Analyzer
(PAN). The comparison analysis of the scenarios for the product Alpha shoes that the
production capacity of the system has been enhanced from 110 products to 150 products
annually with an increase of 26.67% and the cycle time of the complete system was reduced
from 202 hours to 97 hours with a decrease of 51.98%. The comparison chart of the developed
scenarios for product Alpha is shown in the figure below.
Optimization Scenarios
250
200
WAITING TIME (HOURS)
150
100
50
0
Base Model Scenario 1 Scenario 2 Scenario 3
P1. Queues (Hours) 55.52 0.936 0.96 2.03
P2. Queues (Hours) 59.44 1 0.85 2.03
P4. Queues (Hours) 73.5 1.5 2.12 4.1
P6. Queues (Hours) 58.58 1.03 0.9 2.49
P9. Queues (Hours) 0 40.96 0 0
P10. Queues (Hours) 0 41.23 0 0
Products Out 110 110 124 150
Cycle Time 202 166 89 97
Figure 57. Comparison chart of scenarios for product Delta
58
7.5 Selection of Optimized Option
The development of the different scenarios for each product of the case company under study
are followed by the selection of the optimized option. For that purpose, the tool ‘Opt Quest’ in
the ARENA ® simulation software version 14.0 is used. To proceed with the optimization of
the system in Opt Quest, we need to first finalize the following information for each case under
study:
e. Objective function
f. Controls
g. Constraints
h. Responses
7.5.1 Optimization of Product Alpha
The performance evaluation of the case under study for product using multi system variables
has the objective to enhance the overall system production up to 18 products per annum. The
physical and cost constraints limit the use of parallel workstation not more than three at each
workstation. The responses and controls used for the optimization process will include all the
workstations in the assembly line. The data required for the optimization of product Alpha
using Opt Quest is as under:
Objective function. Maximize: Total System Production
Subject to 1 ≤ No. of stations ≤ 3
System Production ≥ 18
The suggested optimized solution obtained by running the optimization process for product
Alpha depicts that the maximum system production that can be achieved with these constraints
and limitations does not exceed more than 20 products annually, that we have already achieved
by implementation of hybrid model with line balancing. Therefore, the maximum system
production for product Alpha will remain 20 products annually. The layout of the optimization
of product Alpha using Opt Quest is shown in the figure below.
59
Figure 58. Optimized solution for product Alpha
7.5.2 Optimization of Product Delta
The goal of the performance assessment of the case under study for the product Delta
employing multiple system variables is the same as it is for product Alpha, that is to increase
the system's overall production up to 150 products annually. In this case study, the use of
parallel workstations is only permitted with a maximum of three at each workstation due to
physical and financial restrictions. All the assembly line workstations' reponses and controls
will be used in the optimization process. The following information is needed to optimise
Product Delta using Opt Quest:
Objective function. Maximize: Total System Production
Subject to 1 ≤ No. of stations ≤ 3
System Production ≥ 150
The suggested optimized solution, which was created by simulating the optimization process
for product Delta, shows that the maximum system production that can be achieved under these
restrictions and limitations does not exceed 152 products per year, whereas the maximum
system production as obtained from the developed hybrid model of the scenario-3 was 150
products annually. The suggested solution as obtained from the optimization of assembly line
of product Delta is shown in the figure below.
60
Figure 59. Suggested optimized solution for product Delta
The outcome of the suggested optimized solution for product Delta depicts that there must be
parallel stations at seven workstations i.e., P6, P6a, P7, P8, P9, P10 & P13. However, the
increase of only two products by using parallel stations at seven workstations will be more
costly solution for the industry. Therefore, the industry will stick to the solution obtained from
the hybrid model.
61
Conclusion
Modelling and performance analysis of the assembly lines of any manufacturing industry using
simulation software is getting a lot of attention. Among various commercially available
software for modeling & simulation, ARENA ® simulation software is the most encouraging
and potential tool used for evaluating performance and then taking the necessary actions for
improvements of current inadequacies.
System production throughput and making the most use of the few available facilities are the
main concerns in the highly regulated industry of high technology parts manufacturing
companies. Allocating resources and utilities for competing activities requires methods that are
both efficient and inexpensive. Thus, for bottling companies, applying simulation technique to
provide an optimized solution and enhance the utilization of resources or facilities can result
in significant financial gains. In these circumstances, simulation serves as an excellent tool
since it enables modelling the stochastic behaviour of the system and drawing valid conclusions
about the performance metrics that are of interest. Focus is brought on the system's constituent
parts and their interdependencies during simulation creation.
This study focuses on the development of a simulation model for the assembly lines of a high
technology parts manufacturing industry to identify the bottlenecks in the production lines.
Results and calculations by using the ARENA® Simulation software helped to reconfigure the
existing assembly line keeping in view the constraints of human resource, supply chain and the
equipment. Different options were proposed to enhance the production capacity of the products
taken as a case study from the industry.
The results obtained through modeling & simulation for product Alpha of the case study depicts
that the bottlenecks and queues in the system were eliminated with an increase of 30% in annual
system production, reduction in average cycle time by 45.16% and 8.74% increase in overall
system utilization. Whereas for the product Delta of the case study industry, the bottlenecks
and queues in the system were eliminated with an increase of 26.67% in annual system
production and 51.98% reduction in average cycle time. The DES has provided an optimized
solution for the cases under study to achieve the objective of desired production for the
industry. The analysis done in this research work will help the industry to compete
exceptionally in the global market of high technology and precision parts manufacturing
industries.
62
To conclude, simulation modelling analysis is the most significant and developing subject in
the field of industrial engineering, and this thesis work is hoped to be a valuable plus point to
further strengthen the related research activities. Additionally, it implies that simulation models
may be useful not only in solving optional problems in the industry but also in sparking more
creative thought about alternative issues, models, and applications for simulation models.
63
Recommendations
High levels of automation and integration, intricate relationships between system components,
and high capital expenditures are characteristics of modern production. Even while modelling
and analysis are crucial to ensuring optimal system performance, the integration and
complexity of systems can make it challenging to employ simply analytical tools. As a result,
simulation is still one of the most popular strategies for addressing this demand.
In order to improve manufacturing performance by avoiding or minimising the shortcomings

of the industry, it is necessary to first analyse the overall manufacturing performance of the
factory, then identify the major contributing factors that led to the poor performance, raise
awareness of the need for ongoing performance analysis, and finally implement or use the
proposed methodology to eliminate those factors to enhance the system overall efficiency.
In addition to the study and analysis, the following recommendations are made.
a. Since modelling manufacturing industries using ARENA Simulation Software is a

relatively new field of study, it can be broadened to model other types of manufacturing
systems and complex production types in order to obtain, tailored solutions.
b. Low utilization of the factory's resource (human and equipment/facility) is the other
important component that must be improved because it directly affects performance.
c. The required maintenance expenses for each equipment, machine, or failure are
significant for any industry. The research helped to sort out the issues related to
maintenance of the equipment and machines.
d. All types of manufacturing systems can benefit from the expanding use of simulation
tools, which also allow for the modelling of complex industrial processes to produce
specialized, individualized solutions.
e. Pilot validation and dynamic continuous improvement in the industrial case studies.
64
Future Work
There are some recommendations made for the future work due to shortage or unavailability
of enough time to conduct additional study beyond what has already been done.
a. The research work may be expanded to complex assembly lines of any industry where
specialized processes are involved.
b. Comparison of the results obtained through discrete event simulation and by other similar
techniques can be undertaken as case analysis for any industry.
c. The system variables used in this research work include system throughput, cycle time,
resource utilization, bottlenecks and queues, process time variance and inter-arrival time
variance. Various other combinations of system variables not used in this research may
be used.
d. A novel idea to improve this research work can be taken as the part of the research work.
e. Human learning curve factor may be incorporated.
65
REFERENCES
[1] Arun. B. R. D.S.S.Sudhakar and S. R. Vivek k.Sunnapwar, “Improving the

Performance of Assembly Line: Review with case study,” presented at the International
Conference on Nascent Technologies in the Engineering Field (ICNTE-2015), India, 2015.
[2] D. G. O. Temesgen Garoma Abeya and M. G. Gutu Ofgera, “Manufacturing System

Modeling and Performance Analysis of Mineral Water Production Line using ARENA
Simulation,” Int. J. Eng. Adv. Technol., vol. 9, no. 5, pp. 312–317, Jun. 2020, doi:
10.35940/ijeat.D8033.069520.
[3] K. K. B. Hon and S. Xu, “Impact of Product Life Cycle on Manufacturing Systems
Reconfiguration,” CIRP Ann., vol. 56, no. 1, pp. 455–458, 2007, doi:
10.1016/j.cirp.2007.05.109.
[4] João Tiago Martins Covas, “Simulation of Assembly Lines Balancing in the footwear
industry,” University of Porto, 2014.
[5] Neha Prajapat Ashutosh Tiwari, “A review of assembly optimization applications using
discrete event simulation,” Int. J. Comput. Integr. Manuf., vol. 30, no. 2–3, pp. 215–228, 2017,
doi: 10.1080/0951192X.2016.1145812.
[6] L. L. Qing Chang and Jun Ni, “Data driven bottleneck detection of manufacturing
systems,” Int. J. Prod. Res., vol. 47, no. 18, pp. 5019–5036, 2009, doi:
10.1080/00207540701881860.
[7] C. E. Betterton and S. J. Silver, “Detecting bottlenecks in serial production lines – a

focus on inter-departure time variance,” Int. J. Prod. Res., vol. 50, no. 15, pp. 4158–4174, Aug.
2012, doi: 10.1080/00207543.2011.596847.
[8] O. M. Sébastien Gebus and E. J. Alexandre Soulas, “Production optimization on PCB

assembly line using Discrete-Event Simulation,” Department of Process and Environmental
Engineering, University of Oulu, Finland, Report A 24, 2004.
[9] K. M. S. K. Kilani and M.A. Gomaa, “Analysis of earth-moving systems using discrete-
event simulation,” Alex. Eng. J., vol. 54, pp. 533–540, 2015.
[10] M. E. M. Mahmood janloo, A. B. Behnam Einabadi, and Eva Rother, “A mixed-model

assembly line sequencing problem with parallel stations and walking workers: a case study in
the automotive industry,” Int. J. Prod. Res., 2022, doi: 10.1080/00207543.2021.2022801.
66
[11] H. Sun and L. Xu, “Optimization of Scheduling Problem for Auto Mixed Model
Assembly Line,” in 2009 International Workshop on Intelligent Systems and Applications,
Wuhan, May 2009, pp. 1–3. doi: 10.1109/IWISA.2009.5072993.
[12] C. D. Fatah Chetouane, “Modeling and Improving Emergency Department Systems

using Discrete Event Simulation,” Soc. Model. Simul. Int., vol. 83, no. 4, pp. 311–320, 2007,
doi: 10.1177/0037549707083111.
[13] S. H. Alptuğ Selçuk Büyüksünetçi, “Simulation modeling and analysis of a new mixed
model production lines,” in Proceedings of the 2005 Winter Simulation Conference, Turkey,
pp. 1408–1412.
[14] M. R. A. Paramasivam and K. Surya Prakasa Rao, “Performance Evaluation of

Reconfigurable Mixed Model Assembly Line,” Int. J. Print. Packag. Allied Sci., vol. 4, no. 3,
pp. 1989–2000, 2016.
[15] Yuval Cohen, “Absenteeism as a major cause of bottlenecks in assembly lines,” Int. J.
Prod. Res., vol. 50, no. 21, pp. 6072–6080, 2012.
[16] Selin Hanife ERYÜRÜK, “Clothing assembly line design using simulation and
heuristic line balancing techniques,” Istanbul Technical University, Textile Engineering
Department, Istanbul, Turkey, 4/2012, Sep. 2012.
[17] Martina Kuncova, Katerina Svitkova, Alena Vackova, Milena Vankova, “Discrete
Event Simulation Of The COVID-19 Sample Collection Point Operation,” United Kingdom,
2021, vol. 35.
[18] A. K. Ali Mousavi, “Modeling emergency departments using Discrete Event

Simulation Techniques,” presented at the Winter Simulation Conference, London, 2005.
[19] A. Y. Gebremedhin Gebremicheal and M. H. Teklewold Meraha, “Productivity

Improvement through Line Balancing by Using Simulation Modeling (Case study Almeda
Garment Factory),” J. Optim. Ind. Eng., vol. 13, no. 1, pp. 153–165, Winter & Spring 2020,
doi: 10.22094/JOIE.2019.567816.1565.
[20] Junfeng Wang, Qing Chang, Guoxian Xiao, Nan Wang, Shiqi Li, “Data driven
production modeling and simulation of complex automobile general assembly plant”,
Computers in Industry, vol. 62, pp. 765-775, (2011), doi:10.1016/j.compind.2011.05.004
67
[21] Elena Guseva, Tatyana Varfolomeyeva, Irina Efimova, Irina Movchan, “Discrete event
simulation modelling of patient service management with Arena”, Journal of Physics:
Conference Series, vol 1015, (2018), doi :10.1088/1742-6596/1015/3/032095.
[22] Alghazi, Anas Alsayed, “Balancing and Sequencing of Mixed Model Assembly Lines”,
Clemson University ProQuest Dissertations Publishing, (2017).
[23] Douglas C. Montgomery, George C. Runger, “Applied Statistics and Probability for
Eingineers”, Wiley, 6th edition, USA, ISBN-13 9781118539712.
68
Appendices
Appendix A: Category overview of the existing model for product Alpha
69
70
71
72
73
74
75
76
77
Appendix B: Category overview of the optimized model for product Alpha
78
79
80
81
82
83
84
85
86
Appendix C: Category overview of the existing model for product Delta
87
88
89
90
91
92
93
Appendix D: Category overview of the optimized model for product Delta
94
95
96
97
98
99
100
101
102

TH-SOF-2043-Performance Evaluation Model For MMALs Using Multi-System Variables Through Discrete Event Simulation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TH-SOF-2043-Performance Evaluation Model For MMALs Using Multi-System Variables Through Discrete Event Simulation

Uploaded by

Copyright:

Available Formats

Performance evaluation model for MMALs using multi-system

variables through Discrete Event Simulation

Dr. Aamer Ahmed Baqai

DEPARTMENT OF MECHANICAL ENGINEERING

COLLEGE OF ELECTRICAL & MECHANICAL ENGINEERING

NATIONAL UNIVERSITY OF SCIENCES & TECHNOLOGY

A thesis submitted in partial fulfillment of the requirements for the degree of

Dr. Aamer Ahmed Baqai

Thesis Supervisor’s Signature: _______________________________________

DEPARTMENT OF MECHANICAL ENGINEERING

COLLEGE OF ELECTRICAL & MECHANICAL ENGINEERING

NATIONAL UNIVERSITY OF SCIENCES & TECHNOLOGY

Dr. Aamer Ahmed Baqai

Dr. Aamer Ahmed Baqai

Supervisor’s Name: Dr. Aamer Ahmed Baqai Signature: ____________

Dissertation submitted to the Faculty of the Department of Mechanical Engineering,

College of Electrical & Mechanical Engineering,

National University of Sciences & Technology (NUST)

in partial fulfillment of the requirements for the degree of

Dr. Aamer Ahmed Baqai, Supervisor

Dr. Uzair Khaleeq-uz-Zaman

Dr. Sajidullah Butt

Dr. Yasir Ahmed

College of Electrical & Mechanical Engineering,

National University of Sciences & Technology (NUST), Islamabad

Keywords: Discrete Event Simulation, Simulation Modeling, Workflow Optimization.

1.4 Objectives of the Study

1.4.2 Specific Objective

a. To analyze the existing workflow of the assembly lines

b. To identify the bottlenecks and queue lengths

c. To optimize the existing workflow by reducing bottlenecks and queue lengths

d. To enhance production capacity

1.5 Significance of the study

1.6 Scope of the Study

1.7 Tools for DES

a. To model physical systems using DES.

In addition to the applications mentioned above, various advantages of Arena as discussed by

a. Applications in construction field.

1.8 Organization of the thesis

Figure 1. Assembly line types

2.2 Modeling & simulation

a. Variety of assembly applications, including modelling a four-closed-loop network

b. Evaluating multiple control policies

c. Validating an aggregated event scheduling approach

d. Confirming an analytic model

e. Creating a mixed model assembly line

f. Modelling an assembly cell.

2.3 DES based case studies

Figure 2.Methodology block diagram

4.2 Product Description

a. Total number of workstations = 19

Figure 3. Existing layout of Case study industry - Product alpha

Figure 4. Precedence diagram of Product Alpha

Work Element Preceded By Average Task Time (in Hours)

Table 2. Process operation details for Product Alpha

S. No Station No. Process

f. Total number of workstations = 14

Figure 6. Precedence diagram of Product Delta

Work Element Preceded By Average Task Time (in Hours)

Table 4. Process operation details for Product Delta

S. No Process No. Process

Figure 7. Layout of ARENA Input Analyzer

5.1 Discrete Probability Distribution Types