A CAPACITY MODEL FOR RESEARCH BASED 
GOVERNMENT MANUFACTURING SYSTEMS 

 
 
 
 
 
 
 

A Dissertation Presented for the 
Doctor of Philosophy  

Degree 
The University of Tennessee, Knoxville 

 
 
 
 
 
 
 
 
 
 
 
 

Tomcy Thomas 
December 2019 

 
 



 

ii 
 

Copyright © 2019 by Tomcy Thomas.  

All rights reserved. 

 



iii 
 

DEDICATION 
 

To my LORD and Savior Christ Jesus for blessing me more than I could 

ever ask, think or imagine. To my mother who conceived me, bore me and gave 

birth to me. To all those people who were a positive influence in my life. To my 

wife Susen, who is always supportive of me especially during the difficult phases 

over the last few years. To all others, who gave valuable lessons in life.  



 

iv 
 

ACKNOWLEDGEMENTS 
 

First of all, I want to express my gratitude to Dr. Rapinder (Rupy) S. 

Sawhney for encouraging me and giving me the opportunity to be in the program 

and also for all the help and guidance provided during the length of the program. 

I would like to thank Carla Arbogast for the support and encouragement during 

the last few years; to Dr. Gary Arbogast for the time spent reviewing and 

providing feedback.  I want to thank the members of my committee: Dr. John E. 

Kobza, Dr. Hamparsum Bozdogan and Dr. Harry Lee Martin for all the academic 

help and support. To Dr. Enrique Macias De Anda and Dr. Ninad Pradhan for 

their patience, encouragement, suggestions and feedback in this research 

project. I express my gratitude to Dr. Roshanak Akram and Dr. Marcel Mutiyama 

for reviewing this dissertation. Also to Dr. Steven Sherman, Dr. Robert Wham, 

Dr. David Depaoli, Steve Owens and others at Oak Ridge National Lab for being 

helpful during the course of the associated case study. Yvette Gooden for 

reviewing and editing this document. All faculty and staff members of the 

department who were always helpful. To all the people I have met and worked 

with while at the university. To all friends who were always helpful. 

  



 

v 
 

ABSTRACT 

 
Manufacturing systems take longer than necessary to be designed and 

implemented, hence the greater developmental cost. A class of manufacturing 

systems exist which would benefit from the concepts of reverse engineering, to 

reduce lead times for establishing critical manufacturing capabilities essential to 

national safety and security. There is a need to reverse engineer these 

manufacturing systems as no current system and/or body of knowledge exists. 

Manufacturing systems vary in their ability to deliver products in an efficient and 

reliable manner and hence the variability in national readiness. Presently the 

design of manufacturing systems for some critical operations ranges from an 

educated trial and error process to duplicating from documentation and 

professional expertise. The literature search highlights the non-existence of a 

current systematic operational reverse engineering model that could be the 

standard for designing manufacturing systems. 

One of the main constraints in the manufacturing is that the time for 

production is limited and denoted by time available (TA). The time to finish (TF) is 

the time needed to complete the manufacturing operations in a facility so that the 

entire quantity demanded is produced, from start to end, in the production line. If 

the TF is less than the TA there is sufficient capacity to meet the demand. 

Literature search indicates that no study has been conducted to compute the TF. 

Further, it also indicates that no study has been carried out focusing on the 



 

vi 
 

impact of variations and disruptions at the design stage, even though these 

topics are covered in analysis of existing operational systems. 

The algorithms and mathematical model were developed. The model will 

compute the exact TF taking into account variation, disruption and flow issues. 

The equation for TF was developed. The model to be designed is validated using 

information from a government manufacturing system. 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 

 
 
 
 
 
 
 
 



 

vii 
 

TABLE OF CONTENTS 

 
1 Introduction .................................................................................................... 1 

1.1 Background and Motivation .................................................................... 1 
1.2 Problem Statement ................................................................................. 3 
1.3 Research Context ................................................................................... 5 

1.4 Boundaries, Scope and Limitations ...................................................... 10 
1.4.1 Product and Customer Characterization ........................................ 10 
1.4.2 Supply Chain Infrastructure Characterization ................................ 10 
1.4.3 Facility Infrastructure Characterization .......................................... 10 
1.4.4 Manufacturing Infrastructure Characterization ............................... 11 

1.4.5 Additional Scope ........................................................................... 12 
1.4.6 Assumptions and Study Limitation ................................................ 13 

1.5 General Approach ................................................................................ 14 
1.5.1 Base of logic .................................................................................. 15 
1.5.2 Tools Used .................................................................................... 19 

1.6 Conceptual Framework......................................................................... 20 

1.7 Contributions ........................................................................................ 24 
1.7.1 Impact of the Model for the Government ....................................... 24 

1.7.2 Theoretical and Methodological Contributions ............................... 24 
1.8 Outline .................................................................................................. 25 

2 Literature Review ......................................................................................... 27 

2.1 Capacity Determination ........................................................................ 27 
2.2 Scaling up from Research to Production .............................................. 30 

2.3 Manufacturing System Design (MSD) .................................................. 32 

2.4 Operational Excellence, Variations, Disruptions and Flow.................... 36 

2.5 CT, LT and TAKT time .......................................................................... 41 
2.6 Effects of Variation, Disruption on CT ................................................... 43 
2.7 Summary of Literature Review ............................................................. 50 

3 Methodology ................................................................................................ 51 
3.1 Roadmap and Framework .................................................................... 53 

3.2 Mathematical Model for TF ................................................................... 55 
3.2.1 List of Variables and Constraints ................................................... 56 
3.2.2 Model Equations ............................................................................ 58 

3.3 Phase 1 – Capacity Based on TF in Ideal Conditions ........................... 71 
3.4 Phase 2 – Strategy to Enhance Capacity based on TF ........................ 73 

3.4.1 Process Characteristics and System Classification ....................... 75 
3.5 Phase 3 – Variation .............................................................................. 77 

3.6 Phase 4 - Disruptions ........................................................................... 80 
3.7 Phase 5 - Flow Design.......................................................................... 86 
3.8 Phase 6 – Output of the Model – System Design ................................. 87 
3.9 Summary of Phases 3 - 6 ..................................................................... 88 
3.10 TF, Schedule and the Time of all Manufacturers in the Chain .............. 89 

3.10.1 Comparison of all the Combinations for Feasibility ........................ 90 



 

viii 
 

3.10.2 Schedule ....................................................................................... 90 

3.10.3 Computing the Total Time for all Manufacturers ............................ 90 
3.10.4 Manufacturers in the Chain ........................................................... 91 

4 Algorithms .................................................................................................... 92 
4.1 Phase 2 Algorithms for Strategy ........................................................... 92 
4.2 Phase 3 Algorithms for Variation Design .............................................. 95 
4.3 Phase 4 Algorithms for Disruption Design .......................................... 103 
4.4 Phase 5 Algorithms for Flow Design ................................................... 107 

5 Validation and Results ............................................................................... 111 
5.1 Business Case Study for Validation of Model ..................................... 112 
5.2 Validation Roadmap ........................................................................... 114 
5.3 Results ................................................................................................ 116 

5.3.1 Design for Manufacturer 4 in the Chain ....................................... 116 

5.3.2 All Manufacturers in the Chain .................................................... 129 
5.4 Recommendations .............................................................................. 131 

6 Conclusions and Future Work.................................................................... 135 

6.1 Conclusions ........................................................................................ 135 
6.2 Future Work ........................................................................................ 135 
6.3 Generalization of the Model to all Manufacturing ................................ 137 

List of References ............................................................................................. 138 
Appendices ....................................................................................................... 146 

Appendix A .................................................................................................... 147 
Phase 1 – Numerical Example of TF computation ..................................... 147 
Phase 2 – Flowchart and Numerical Example of TF computation ............. 147 

Phase 3 – Flowcharts and Numerical Example ......................................... 148 
Phase 4 – Flowcharts and Numerical Example ......................................... 153 

Phase 5 – Additional Flow Design Algorithms ........................................... 159 
Review of the Designed System ................................................................ 176 

Appendix B – List of Distributions .................................................................. 179 
Appendix C – Sample MATLAB Codes and GUI ........................................... 182 

System Classification Algorithm (Phase 2) Sample MATLAB Code .......... 182 

Implementation of floating bottleneck algorithm ......................................... 194 
Utilization Algorithm Code ......................................................................... 195 

Variation of 4 CRs Sample MATLAB code ................................................ 197 
Equipment Related Disruption Sample MATLAB code .............................. 200 
Level Batch Size Determination Sample MATLAB code ............................ 207 

Code to generate GUI of Figure C.2 .......................................................... 213 
Appendix D .................................................................................................... 215 

Additional Literature Review – Reverse Engineering ................................. 215 
Vita .................................................................................................................... 219 

 
 



 

ix 
 

LIST OF TABLES 

 
 
Table 1 Scenarios of Model Formulation ............................................................ 61 
Table 2 Systems Classification and the Order [2] ............................................... 76 

Table 3 Mean Absolute Error Comparison Logic Notations ................................ 79 
Table 4 System Classification ............................................................................. 92 
Table 5 Products and Stations ............................................................................ 94 
Table 6 Production Sequence and Corresponding Products .............................. 94 
Table 7 Notations of Design for Variation ........................................................... 97 

Table 8 Notations for the Disruptions Algorithm ................................................ 104 
Table 9 Comparison of Results ......................................................................... 120 
Table 10 MACTEo and CVo for Selected Combinations ................................... 123 

Table 11 Some Selected results ....................................................................... 128 
Table 12 Recommended Combination of Results ............................................. 133 
Table 13 Notations for Flow Selection Algorithm of Figure A.18 ....................... 160 

Table 14 Discussion about Inventory ................................................................ 161 
Table 15 Nomenclature for flow design algorithm ............................................. 162 
Table 16 Additional Information to Develop Operations Management .............. 164 

Table 17 Notations for Figure A.37 ................................................................... 175 
Table 18 Table of Common Discrete Distributions ( [71] [72] and [84]) ............. 179 

Table 19 Table of Common Continuous Distributions ( [71],[72], [84] ) ............. 180 
Table 20 CV of a Few Common Distributions ................................................... 181 
Table 21 Example for a Triangular Distribution ................................................. 181 

   
 



 

x 
 

LIST OF FIGURES 

 
 
Figure 1.1 CT and TF are different........................................................................ 7 
Figure 1.2 TA Continuous versus TA in Multiple Periods ...................................... 8 

Figure 1.3 OE factors tree structure .................................................................... 16 
Figure 1.4 Variation affects CT and TH ............................................................... 18 
Figure 1.5 Disruption affects CT and TH ............................................................. 18 
Figure 1.6 Process Time Increase and its effect on CT, TH and TF ................... 20 
Figure 3.1 Model Framework .............................................................................. 54 

Figure 3.2 Manufacturing Supply Chain .............................................................. 56 
Figure 3.3 Variations General Blocks .................................................................. 78 
Figure 3.4 Mean Absolute Error Comparison Logic ............................................ 79 

Figure 3.5 Tree structures of the critical resources ............................................. 82 
Figure 3.6 Disruptions General Blocks ................................................................ 85 
Figure 3.7 Blocks in the Design for Flow [2] (FAD) ............................................. 87 

Figure 4.1 Phase 2 Algorithm Flowchart ............................................................. 93 
Figure 4.2 Floating Bottlenecks Formulation ....................................................... 94 
Figure 4.3 Batching block diagram.................................................................... 109 

Figure 4.4 Balancing block diagram .................................................................. 109 
Figure 5.1 Results of paired t test in JMP ......................................................... 122 

Figure 5.2 Graph of Time versus Level ............................................................. 126 
Figure 5.3 TF with CVa and CVp of 0.5 each ................................................... 126 
Figure 5.4 TF with no variation and 95% A ....................................................... 127 

Figure 5.5 Illustration of sample algorithm output ............................................. 128 
Figure 5.6 Graph of Total Time versus Level case1 ......................................... 131 

 
Figure A.1 Algorithm for Floating Bottleneck ..................................................... 147 

Figure A.2 Design for Variations (V) part 1 – Process Control .......................... 149 
Figure A.3 Design for Variations (V) part 2 – Operations Type ......................... 149 

Figure A.4 Design for Variations (V) part 3 – Priority Branching ....................... 150 
Figure A.5 Design for Variations (V) part 4 – People and Schedule ................. 150 
Figure A.6 Design for Variations (V) part 5 – Arrival part1 ................................ 151 
Figure A.7 Design for Variations (V) part 6 – Arrival part2 ................................ 151 
Figure A.8 Design for Variations (V) part 7 – Process ...................................... 152 

Figure A.9 Design for Variations (V) part 8 – Placement .................................. 152 
Figure A.10 Disruptions Algorithm Design General - DIS0 represents the priority 

of the 4CRs ................................................................................................ 153 

Figure A.11 Failure, Repair and Planned Disruptions ....................................... 154 
Figure A.12 Equipment / Machine Disruptions Algorithm part1 ......................... 154 
Figure A.13 Machine Disruptions Algorithm – Operations & Maintenance ........ 155 
Figure A.14 Machine Disruptions Algorithm – Scheduling ................................ 155 

Figure A.15 Set Up Changes Disruption Algorithm ........................................... 156 
Figure A.16 Materials Disruptions Algorithm ..................................................... 157 



 

xi 
 

Figure A.17 People Disruptions Algorithm ........................................................ 157 

Figure A.18 Flow Selection ............................................................................... 159 
Figure A.19 Design Type .................................................................................. 160 

Figure A.20 Inventory Decision ......................................................................... 160 
Figure A.21 Capacity Analysis .......................................................................... 161 
Figure A.22 Flow Planning block diagram ......................................................... 168 
Figure A.23 Work Environment Preparation block diagram .............................. 168 
Figure A.24 PSSW block diagram..................................................................... 168 

Figure A.25 PFF block diagram ........................................................................ 168 
Figure A.26 FAD1 Flow Planning ...................................................................... 169 
Figure A.27 FAD2 WEP .................................................................................... 169 
Figure A.28 FAD3 PSSW .................................................................................. 170 
Figure A.29 FAD4 PFF ..................................................................................... 170 

Figure A.30 Batching FAD5 part 1 .................................................................... 171 
Figure A.31 Batching FAD5 part 2 set up reduction .......................................... 171 

Figure A. 32 FAD6 Balancing part 1 ................................................................. 172 

Figure A. 33 Example of a linear production line............................................... 173 
Figure A.34 FAD6 Balancing part 2 .................................................................. 173 
Figure A.35 FAD7 ............................................................................................. 174 

Figure A.36 FAD8 ............................................................................................. 174 
Figure A.37 Flow Type Selection (F) ................................................................. 176 

Figure A.38 Final Stages of Algorithm .............................................................. 177 



 

1 
 

1  

INTRODUCTION 

1.1 Background and Motivation 

There is a class of manufacturing systems that establish critical manufacturing 

capabilities essential to national safety and security. These manufacturing systems 

are small, percentage wise, when considering the total manufacturing in the US, but 

large in diversity and national impact.  An example of how manufacturing of a 

particular product will affect the future innovation and national importance is the 

space missions [1].  There is not much work published about how government 

production capabilities affect national innovations, safety and security. Most products 

coming out of research, even from the national labs in the United States, are 

commercialized by private industries; the government does not control the 

manufacturing except for a very few products which are detrimental to public safety if 

not controlled (an example is nuclear materials and products). The following are 

environments which require significant manufacturing system design: 

1. There are certain products of national interest that are not currently being 

manufactured in the US for a variety of reasons including procurement of the 

product from external sources. An example is when imported products require 

re-establishment of non-existent processes and skills due to political 

circumstances. 

a. Since these products have not been produced for some time, the 

necessary facilities for manufacturing do not exist. 



 

2 
 

b. In some cases, the product has not been produced in such a long time 

that the knowledge necessary to manufacture the product no longer 

exits.  

2. There are new products developed through research efforts of the government 

that have the following requirements: 

a. Products produced in a research environment in which processes need 

to be scaled up for commercialization.  

b. Current products that require significant design and functional 

modification and therefore a redesign of the manufacturing process.  

3. There are manufacturing systems that are obsolete from an operational and 

technological perspective that are currently in need of a major redesign. 

 

There is no current systematic approach that allows an efficient and reliable 

operational manufacturing system to be designed in the least possible time, 

considering the environments described above. Typically, the processes are 

designed based on expertise and experience. Without a systematic study of the 

product and the mechanisms by which it is assembled, much effort will have to be 

expended to understand how to manufacture this product. This trial and error 

approach has its associated cost; hence valuable resources may have to be wasted 

for this process. In addition, the length of time required to complete the production 

process design is excessive. As a result, the effects of trial and error approach are 

the following: (1) products in the environments discussed will take longer than 



 

3 
 

necessary, (2) there is considerable variation in the efficiency and reliability of those 

systems, (3) there is greater cost in developing that manufacturing process, and (4) 

the ability of the industry to introduce products to market will be delayed. This 

research is focused on creating a standardized approach to design operational 

manufacturing systems. The methodology developed in this dissertation could be 

used from a single user perspective as focusing on the scalability of research-based 

production to manufacture the demanded quantity. It could also be used as a 

production planning tool so that the planners will have an exact idea as to how the 

activities are to be arranged on the production floor in advance. This dissertation 

focuses on the perspective of the user to design their manufacturing operations. 

1.2 Problem Statement 

There is significant variation in manufacturing system design with the key 

variability resulting from the history surrounding the manufacturing process, 

personnel available and their level of expertise. The design of manufacturing systems 

critical to the national interest ranges from an educated trial and error process to 

duplicating the process from documentation and professional expertise. The literature 

search, presented in Chapter 2, highlights the non-existence of a current systematic 

operational manufacturing design model that could be the genesis of any level of 

standardization. This lack of standardization impacts national readiness. 

Manufacturing systems take longer than necessary to be designed and implemented 

and hence the greater developmental cost. Manufacturing systems vary in their ability 



 

4 
 

to deliver products in an efficient and reliable manner and therefore impact national 

readiness. 

The focus of this research is to create a conceptual framework and the 

supporting rationale and methodology that allows for one to design an operational 

manufacturing system based on the concepts of operational excellence. The specific 

objective of this research topic is to design a manufacturing system for a product 

whose manufacturing processes are ill-defined and whose production line is non-

existent. Any operational excellence framework must, beyond the fundamentals of 

the physical equipment, consider flow, disruptions and variation in the system [2]. 

These three principles each have a unique focus and approach but are not 

independent of each other. Therefore, the key objectives of this research are: 

1) Create a conceptual framework that integrates the concepts of operational 

excellence into the manufacturing design framework. 

a. Study the flow issues. 

b. Study the effects of variation. 

c. Study the effects of disruptions. 

2) Study the effect of those issues on the operational time (CT and TF - 

details in Section 1.3). 

3) Reduce time and cost in designing production processes. 

 

A suitable manufacturing/production system should be designed and 

developed to manufacture a given quantity per year (demand - example 2000 



 

5 
 

units/year) of a product.  The demand should be met in full and hence the 

throughput of the system should be at least equal to the demand. There are 

many units (manufacturers) involved in the manufacturing chain for the product from 

its initial materials to the final product. Each manufacturer in the chain has constraints 

and limitations. There are many system level constraints to be satisfied for the 

production to be successful. Available facilities are to be retrofitted for this product 

with explicit production requirements. Some of the manufacturers’ internal systems 

do not allow for changes, whereas some are open to be redesigned completely and 

others are only open to some changes. 

Instead of using trial and error approaches to design the manufacturing 

system, this proposed research approach provides a design which will eliminate the 

wastage of resources during the design phase. The literature review indicates that no 

work has been done applying operational excellence during the design stage of a 

manufacturing system. The approach to apply operational excellence to create a 

standardized design of a manufacturing system is outlined in the general approach in 

Section 1.5. The negative impact of variability is to be overcome by having the ability 

to withstand increased variation and by establishing a system where output from 

each manufacturer conforms to the standards of the design. 

1.3 Research Context 

One of the main constraints in manufacturing is that the time for production is 

limited. The following defines/explains the terms related to time, used in this research 

work.  The average time a job takes at a station is the machine/station cycle time 



 

6 
 

(CTstation). The average time from the start of a job at the beginning of the line until it 

comes out at the end of the routing (when there is more than one station in the line) 

is the cycle time of a given line/routing (CTline).  The lead time (LT) of a given routing 

or line is the time allotted for the production of a part on that routing or line is. Cycle 

times are generally random whereas the LT is a constant [3]. The CT vs LT 

combination looks only at the fact of whether or not a unit or batch of product or part 

meets the customer requirement given that LT is developed based on TAKT time. 

Capacity is checked in a new way in this dissertation. In industry the 

production capacity is defined as the maximum quantity of products that can be 

produced in a unit of time in the optimal operating conditions [4]. Capacity 

determination from various sources in literature is given in Section 2.1. The literature 

does not discuss the actual time to produce/manufacture the quantity set as the 

capacity or demand. In this dissertation research, a new term, time to finish (TF) is 

introduced. The TF is the time from start to end in the production line until the entire 

quantity demanded is completed; it is the actual manufacturing time to finish 

manufacturing activities for the entire demanded quantity. The term time available 

(TA) is used to represent the duration for which the facility is specifically allocated for 

the manufacturing of this product only. If the TF< TA, then there is sufficient 

capacity to meet the demand. This TA could be in one time block or in different 

periods (time blocks). The TF vs TA combination denotes whether the entire quantity 

demanded can be delivered within the constraints. The manufacturing operations are 

to be finished within the allocated time and hence unfinished work is not an option. 



 

7 
 

The computation of TF will help in determining whether the suggested operations are 

practically feasible.  

CT and TF are not the same, even when the TA is in one single 

continuous block. The reason being that CT focuses on the completion time of each 

unit whereas TF focuses on the completion time of the whole demand. An example is 

shown in Figure 1.1 assuming that there are three production runs needed to 

complete the demanded quantity. The time taken for each production run is the CT line 

for that run. The TF is the actual time from the start of the production until the 

demand is completed. If there is only one production run required to meet the 

demand during the time frame allocated, then, the CTline and the TF will be the same. 

 

 
Figure 1.1 CT and TF are different  

 



 

8 
 

If the TA is spread across different periods (p) it is denoted as TA1, TA2, ---- 

TAp and the corresponding time to finish are denoted by TF1, TF2, ---- TFp. TF is 

applicable when the time allocated is in one block whereas TF1, TF2, ---- TFp is 

applicable when the facility is allocated for different periods in a calendar year. An 

example with the time allocated in a single continuous period and the situation where 

the time is allocated in three different periods (TA1 to TA3) is shown in Figure 1.2. 

The value of TF (TF1, TF2… TFp if time is allocated in different periods) are to be 

computed. The time left over in any period cannot be used in any other period, but it 

can be used to produce more units if needed. If the total demand is to be met through 

manufacturing operations in different periods, the total demand will be divided into 

sub-demands corresponding to the time for each period allocated. The TF for each 

period is computed with respect to the TA for that period. If much time is left between 

the TFp and the TAp, then the demand for that period will be increased and the TFp 

will be recomputed. This will guard against the time not being properly utilized, and it 

will cover for some other periods later. 

 

 
Figure 1.2 TA Continuous versus TA in Multiple Periods 

 
 



 

9 
 

Manufacturing spread across different periods will have more variation 

compared to the production processes occurring in one single block. The reason for 

this is because the startup operations are to be carried out at the beginning of each 

period, resulting in a considerable amount of time spent for activities which are not 

productive but needed. The shutdown activities are to be carried out at the end of 

each period. In the TF approach there are more startups, setups, and cleanup 

activities if the manufacturing is spread over different periods compared to one single 

continuous allocation of time. Shutdowns are more specific because material cannot 

be left in the manufacturing facility. Also, the set up and cleanup activities of every 

station is to be repeated in every period. As a result, the time required to manufacture 

the same quantity of products will be much more when the time allocated is spread 

over different periods, rather than one continuous single period. In contrast, 

manufacturing carried out in one single continuous period of time will have only one 

start up and shut down activity. 

TF will be computed for all cases and checked with the TA. If the actual 

operating time (TAO) is less than the TA, then the TF will have to be compared with 

TAO (TF < TAO). In factory environments, there are laws with respect to the break 

time which may force the TAO to be less than TA, which is discussed in Section 3.3. 



 

10 
 

1.4 Boundaries, Scope and Limitations 

1.4.1 Product and Customer Characterization 

 The product is in demand for government programs such as space 

research and is highly sensitive and regulated. It is to be produced 

exclusively for the government agency (customer). 

 A steady and stable supply of this product is required for the foreseeable 

future. The customer demand is based on delivering a specific quantity of 

this product annually. 

1.4.2 Supply Chain Infrastructure Characterization 

 The supply chain of this product spans many organizations which are 

widespread location-wise. It consists of the supplier of the main raw 

material, the manufacturers and the consumer of the product who will 

deliver it to the customer for their use. 

 The manufacturing activities reside over multiple facilities. Turnkey 

production facilities do not exist for the whole manufacturing chain. 

1.4.3 Facility Infrastructure Characterization 

 Manufacturing facility (physical infrastructure) exists. These facilities were 

built for some other purposes and are now being used for the 

manufacturing of this product. 

 The physical facilities of at least one manufacturer cannot be changed or 

modified; machines cannot be added.  



 

11 
 

 All the transformation processes for a particular manufacturer in the chain 

have to be finished within a restricted number of days as the facility is not 

available year-round. Special setups can be carried out before the actual 

start of the manufacturing process. The manufacturing facilities are 

capable of operating 24/7. 

1.4.4 Manufacturing Infrastructure Characterization 

 The focus of this dissertation is on the manufacturing section of the supply 

chain. The manufacturing processes are in discrete batches. The 

manufacturing is sequential in nature; a workstation processes the output 

of the previous workstation. The output product from the initial 

manufacturer in the chain is the main input material for the next 

manufacturer. 

 Since the system is imbalanced, it may not be possible to have the same 

capacity in all process areas of a particular manufacturer. The reasons 

include but are not limited to: (a) cost, (b) technological challenges and (c) 

regulations. 

 The facilities of a particular manufacturer involved in this transformation 

chain are dedicated to several products critical to national interest. 

However, dedicated time is given to each product during the year in which 

only that particular product is manufactured. The dissertation focuses on a 

single product. The production facility of this particular manufacturer is 

allocated for this product on a continual basis for the duration allowed. The 



 

12 
 

other products are manufactured in this facility using the remaining time 

(manufacturing of products has to occur because of long term contractual 

obligations). 

 Demand is to be met within a specified time period for a manufacturer 

because of the above-mentioned limitation. 

 Process time is much longer (weeks) for the last two manufacturers in the 

chain. The transportation time between process areas of a manufacturer is 

negligible (close to zero) compared to the actual processing time (It will be 

pulled in when needed). 

1.4.5 Additional Scope  

 The required time to finish (TF) the production quantity, as given by the 

demand, is much more important than the time required completing each 

unit of the product, in the case of the manufacturer(s) whose facility is time 

restricted for this product. 

 Once the demand is met, the production may not continue for that year 

even though the facilities are still available in that year as per the early 

allocation. 

 The key measures are TF and TH (TH is set equal to demand which is 

known), the number of production runs (X3) and the number of batches (X2) 

in each production run. As the focus is on TF, the CT is not one of the key 

measures. 



 

13 
 

 There cannot be any inventory buildup in between the process areas in the 

case of a manufacturer. The transformed material stays in the machines 

until it is sent to the next process area. 

 At the end of the manufacturing process for the period, there should not be 

any material remaining in the process areas (stations) of Mfgr.n. If the TA is 

spread across p periods, the WIPend of each period is to be zero.  

 The dissertation will look into the aspect of operations in ideal conditions, 

as well as practical conditions. 

1.4.6 Assumptions and Study Limitation 

 All the materials are available at the time it is scheduled.  

 The facility layout is available and completed. Design of the physical 

infrastructure is out of scope for this study. 

 This manufacturing system has limitations based on existing facility and 

equipment for some of the processes in the production. Therefore, initial 

infrastructure is assumed to be available. 

 Raw material procurement (for the transformation process for the 

manufacturing) and transportation to the manufacturer is out of scope for 

this dissertation. The final product is stored in a storage facility and the 

customer will pull the required quantity when needed. Customer will make 

arrangements for the transportation of the final product. The shipment of 

the final product to the customer is also out of scope of this study. 



 

14 
 

1.5 General Approach 

The various factors of the manufacturing process are to be designed from a 

systems perspective. Systems are composed of elements, functions and 

interconnections [2]. The elements and functions are already defined in Section 1.4.  

The various elements are product, customer, facilities, supply chain and 

manufacturing. The function is a given quantity of the specific product in a given 

period of time and the manufacturing cycle continues in the next period. The 

interconnections are defined through TF, number of production runs and number of 

batches in each production run in ideal conditions and practical conditions where 

flow, variations and disruptions are taken into account.  The manufacturing system 

studies the interconnections from a throughput (TH) perspective (which is defined by 

TF compared to TA). The focus is on getting the manufacturing done with less TF, 

which in turn will depend on the CT, number of production runs and the number of 

batches in each run.  

An algorithm is developed based on the key areas of operational excellence 

such as flow, variation and disruption [2]. The output of the algorithm is the different 

options for the operational manufacturing system design. In practice as seen in 

industry, systems are designed/developed initially and then improved using various 

continuous improvement tools to mitigate the effects of flow problems, variations and 

disruptions. This dissertation takes into consideration the effects of flow, variations 

and disruptions in the system design itself. 



 

15 
 

Normally manufacturing is viewed in a forward direction of converting 

materials into output by suitable manufacturing processes using other resources. The 

materials from the supplier(s) are converted to the product using suitable 

manufacturing/production techniques before it is sent out to the customer. This 

particular approach looks in the reverse direction starting from the product features 

and customer demand, and then finds the proper mechanism to meet the demand 

taking into consideration the constraints in which the system should function. 

1.5.1 Base of logic 

One important performance measure of any manufacturing system is the 

throughput (TH). “The average output of a production process per unit time is defined 

as the system’s TH. It is the average quantity of good (non-defective) parts/products 

produced per unit time [3].” According to Little’s Law [5], TH is dependent on CT and 

Work-In-Process (WIP); when the system operates under steady state. The important 

factors which affect the CT are flow, variability and disruptions [2]; when CT changes, 

TH and WIP also change. These factors also have an impact on TF.  Figure 1.3 

shows the connection between them and also the various factors affecting them. The 

important parameter in this research is time; hence the application/comparison of 

Little’s Law is valid. Since the focus of this dissertation is on the time factor, the 

structure in Figure 1.3 does not focus on WIP. 



 

16 
 

 

Figure 1.3 OE factors tree structure 

 
 

Variations could happen anywhere in the system; this dissertation focuses on 

the variation in the arrival, processes and placement. Any manufacturing system is 

composed of four critical resources (CRs): (1) Materials (M), (2) Equipment/Machines  

 (E), (3) Personnel (P) and (4) Schedules/Information (S) [6]. Disruptions to any of the 

four critical resources will affect the time (TF and CT). The TF is determined or 

affected by the number of production runs, number of batches in each run, quantity in 

each run, the CTstation, variations, disruptions, processing time at each station and the 

wait times. 

There are three types of wait times in any manufacturing operation: (1) wait 

time of the queue denoted as CTq, (2) transfer wait time if a station has to wait for a 

particular number of batches to be processed in the previous station and (3) 

additional wait time (denoted as WTfactor in the computations) because of the 

availability of subsequent stations. All three types of wait times increase TF. The 

additional wait time (WTfactor) is because of blocking and is very significant in 

operations where the inventory buffer between stations is restricted or even set to 



 

17 
 

zero. When the stations get blocked more, the TF increases which reduces the 

capacity to meet demand.  Layout changes are not considered in this dissertation 

and hence flow design concentrates only on batching and balance. 

In the proposed methodology, the design includes the concepts of operational 

excellence involving flow, variations and disruptions. This dissertation assumes that 

the machines are available, and the physical infrastructure is in place. Concepts of 

lean/smart manufacturing and Toyota Production System (TPS) [7] helps in designing 

a manufacturing system which includes the concepts of Operational Excellence (OE) 

in the development stage itself, with a reduced TF and CT. 

In the mathematical computation for TF in Section 3.2, first, the impact of variations 

and disruptions are considered separately, and then, a single equation is developed 

which captures the effect of both issues together. The flow is incorporated in the 

equations by the number of production runs and the number of batches per run. The 

design starts with the concept of ideal conditions. An ideal condition is where there 

are no disruptions or variations anywhere in the system. When the variation 

increases, the CT and hence the TF, increases which results in the TH getting 

reduced. When variation increases, the processing time goes up. Simulation was 

used to study the effect of increasing processing time as a result of variation. See 

Figure 1.4 as an illustration of variation on TH and CT on a system not designed to 

withstand variation. The same system with the process time kept at the base level, if 

disrupted for any reason, results in TH going down further and CT goes up as shown 

in Figure 1.5. Variation affects the processing time of any station or system. Knowing  



 

18 
 

 
Figure 1.4 Variation affects CT and TH 

 
 
 

 
Figure 1.5 Disruption affects CT and TH 

  
 

 
 

0.250
0.350
0.450
0.550
0.650
0.750
0.850
0.950

0

500

1000

1500

2000

0% 10% 20% 30%

C
T

 

O
u

tp
u

t 

Process Time Increase 

Impact of Variation on TH and CT 

O/P CT Linear (O/P)

0.25

0.35

0.45

0.55

0.65

0.75

0.85

0.95

0

200

400

600

800

1000

1200

1400

1600

70 75 80 85 90 95 100

C
T 

O
u

tp
u

t 

Availability (%) 

Impact of Disruption on TH and CT 

O/P CT Linear (O/P)



 

19 
 

that as a fact, a study was conducted directly to see the impact of process time 

increase on TH, CT and TF. The result of which is shown in Figure 1.6 for a system 

not designed to withstand variation. There is a point beyond which the processes 

cannot continue if the TF is constrained to be below a particular value. 

1.5.2 Tools Used 

An algorithm was developed to compute the TF to successfully manufacture the 

product. The algorithm was implemented in MATLAB to get the results. The 

developed algorithm created the rule based, model driven program to design the 

operational manufacturing system. A mathematical model was developed based on 

the Factory Physics [3] equations. For the validation of the model, a simulation 

model, developed and verified by the subject matter experts for the associated case 

study, was used. Simulation over forecasts or under forecasts results; in most cases, 

for the software used in the study, it over forecasts the results. To overcome the 

problems of over or under forecasting, the model was run multiple times before 

selecting the final version. The algorithm was tested, and the model results were 

compared with the simulation results so that the results match to the extent possible. 

 

 



 

20 
 

 
Figure 1.6 Process Time Increase and its effect on CT, TH and TF 

 
 

1.6 Conceptual Framework 

To efficiently put together a good production system, the concepts of lean 

design, reliable process design, controlling variations and looking at the constraints 

from the theory of constraints perspective are needed. The production process will 

have to be designed not only by looking at the capacity of the equipment, but also the 

product flow, the logistics of how the parts are brought in, the schedule and how the 

processes are managed. The motivation in Section 1.1, the problem statement given 

in Section 1.2 and the scope and limitations in Section 1.4, defined the need and 

purpose of this research and the boundaries within which the research is bound 

respectively. The general approach in Section 1.5 explained why a more robust and 

holistic operation excellence model is needed in the design stage itself. The 



 

21 
 

framework/methodology is based on the following rationale which is explained in 

more detail in Chapter 3. 

A. Model Design Inputs 

1) Product and Product Characteristics 

Define the product and its fundamental framework for manufacturing. For 

each manufacturer, the product characteristics and details, along with the bill 

of materials, dictate the manufacturing processes. The manufacturing process 

is known. The demand is the leading factor for the capacity analysis. A key 

data point is to see if the equivalent of a Bill of Material (BOM) exists that 

could provide an insight into the fundamental manufacturing operations that 

need to be supported. 

2) Technology Options  

A review of manufacturing technology is investigated to understand the 

capabilities and capacity of each alternative. These options are stored in a 

database. The feasible solutions are determined based on the ability to meet 

current and projected demand.  

B. Capacity Determination 

1) Design Based on Ideal Conditions 

Ideal conditions occur only when there are no variations or disruptions in 

the system and are found only in perfect systems. The capacity of the system 

will be tested to see whether the demanded quantity can be produced. If it 

cannot be produced in ideal conditions, then there may be a need to build 



 

22 
 

additional facilities. This dissertation assumes that the existing facility can 

meet the demand in ideal conditions; in addition, facilities cannot be changed 

or added for some manufacturers. 

2) Design Based on Practical Conditions 

Here the design considers the issues related to Flow, Variations and 

Disruptions and design the system to produce the required quantity demand 

(set TH = demand). It is to be tested to see whether the practical capacity 

meets or exceeds the demand. If it cannot be met, then the existing facilities 

may not be enough for the manufacturing of the product; capacity 

improvement by adding more physical infrastructure may be needed. 

The actual operational aspect of the negative effect of variation and disruption 

is that the time required to complete the whole process increases as the variation and 

disruption increases. Since the processes cannot be completed in the same time, 

with variations and/or disruptions, compared to ideal conditions, the throughput 

reduces within the time period considered. The TF and the CT increase as a result of 

increasing variability or disruptions. 

C. Design Changes 

1) Evaluation of Design Options Based on Flow Efficiency 

Flow is studied and determined; the TF is estimated. If there are problems 

associated with flow, the CT may increase, thereby affecting the TF and TH. A 

different set of options is evaluated to ensure that it adheres to the lean 

concepts. The basic fundamental principle is to ensure maximum throughput 



 

23 
 

of the process by making sure that the principle of balanced lines with 

minimum lot/batch sizes are assured. Pull systems are considered as an 

option to balanced single piece flow. 

2) Evaluation of Design Options Based on System Variations 

Effects of variability are studied next; the revised TF is estimated. When 

variability increases, the CT and WIP increases and the TH goes down if the 

system is not properly designed to withstand the effects of variation. TF will 

also be increased if variability increases. The feasible options from the 

previous step are evaluated to examine several different aspects of variation. 

The first dimension of variation is straight forward to understand the impact 

that it would have on the product quality. This obviously depends on the 

decisions made in the flow design as variation has a greater impact on CT in 

push systems as compared to pull systems. 

3) Evaluation of Design Options Based on System Disruptions 

Disruptions tend to slow down the manufacturing processes by increasing 

the CT. The TF is recomputed based on the impact of disruptions. The 

feasibility options from the previous step are evaluated to examine two key 

disruptions; setups and maintenance. Each of these two disruptions is 

evaluated based on frequency of failures and the length of failures. Following 

the principles of disruptions in production processes, it is best to avoid 

frequent failures and long lasting failures. 



 

24 
 

Initially, the design looks into the aspects of a perfect system or ideal 

conditions. A perfect system has perfect flow, perfect balance and a single piece flow 

or a small batch size close to a single piece system; it also requires perfect material 

at the right time [2].  Balance can be achieved by: (1) Schedule, (2) People, (3) 

Shifts, (4) Tools and (5) Equipment [2]. A perfect system does not have any variation 

or disruptions and has full employee engagement. Once the design meets the 

demand with a near perfect system concept, the effect of detractors will be added to 

the design. There cannot be perfect balance; the focus is to achieve the best balance 

possible. 

1.7 Contributions 

1.7.1 Impact of the Model for the Government 

1. Reduction of the time to design operational manufacturing systems 

2. Reduction of the cost of designing operational manufacturing system by 

avoiding more trial and error methods 

3. Improvement of system effectiveness 

1.7.2 Theoretical and Methodological Contributions 

1.  A conceptual framework and methodology to design operational 

manufacturing systems is developed. 

2. The modification of Little’s law using TF instead of CT is presented. 

3. The logic provided in this dissertation does not exist in literature and the 

gap is identified in Chapter 2. 



 

25 
 

4. A new logic to check the balance of the line and to identify the cause of 

imbalance, using Mean Absolute Error Cycle Time Overall (MAECTo) and 

the Coefficient of Variation Overall (CVo) (Section 3.5) has been 

presented.  

5. A rule based or model driven (not data driven) algorithm has been 

developed on this concept/logic, which allows users to customize their 

design. 

6. The quantitative model developed to compute the TF. A mathematical 

relationship between the TF and its variables is developed. 

7. The developed equations give an accurate estimate about the time needed 

to manufacture the product to meet the demand. 

8. The developed algorithms take into account the effects of detractors (flow 

issues, variations and disruptions) at the design stage itself so that the 

system will be able to withstand the negative impact of the detractors. 

9. By verifying that the TF is less than TA, the production quantity is 

guaranteed to meet the demand. The presented algorithms eliminate the 

need for iterative development. 

10. The model has been validated in a government manufacturing environment 

for which this model is developed. 

1.8 Outline 

The literature review given in Chapter 2 establishes that there is gap in the 

research done by others and the research detailed in this dissertation. Section 2.6 



 

26 
 

explains how variations, disruptions and flow issues affect the actual process time. A 

detailed approach and methodology by which decisions are to be made for the 

successful manufacturing design of the product in the required quantity is given in 

Chapter 3. It gives insight about the resources needed for the successful 

manufacturing of the product. This design starts from the customer side and works 

backward to the supplier side. The proposed approach is explained in Sections 3.1 to 

3.9. The development of the mathematical model for TF is in Section 3.2. There are 

six phases in the methodology. Phase 1 relates capacity and demand based on TF 

and shows the modification to Little’s law by using TF as a method to check capacity 

instead of throughput. In phase 2, the strategy to enhance capacity based on TF is 

defined. This phase also establishes the time buffer and the threshold. It classifies 

the system and also identifies the bottleneck (floating bottlenecks if more than one) 

and utilization. Phases 3, 4, and 5 discuss variation, disruption and flow issues 

respectively. Output of the model is in phase 6. In Chapter 4, all the algorithms are 

presented.  The validation of the model with a case study and the results are 

presented in Chapter 5. Conclusions, recommendations and future work are given in 

Chapter 6. 

 

 

 

 

 



 

27 
 

2  

LITERATURE REVIEW 

A literature review was performed based on the combination of different key 

words, such as capacity, capacity determination, scaling up production, operational 

excellence, manufacturing system design (MSD), cycle time, variations, disruptions,  

lean, lean manufacturing, reverse manufacturing, and reverse or re-engineering or 

reverse MSD. The following gives key findings from the most important current 

studies. 

2.1 Capacity Determination 

There are many books and articles in literature about estimating or 

determining the capacity of facility/plants at the design stage itself or determining or 

analyzing the production capacity of existing plants. Some of those are shown below. 

However, none of the books or articles estimates the capacity in practical conditions 

without doing a continuous improvement project, which may be costly. This 

dissertation plans to introduce the practical conditions in the design stage itself. The 

production capacity is not achieved when the demand falls and also depends on 

whether or not the product is a made to stock item. The following are the key findings 

from the literature about this topic. 

The maximum quantity of products which are of the appropriate quality and 

assortment that can be produced by an enterprise in a unit of time with the full use of 

the basic production assets in the optimal operating conditions is its production 

capacity [4]. A bi-objective optimization problem was solved by developing a robust 



 

28 
 

production capacity planning model with two layers. The layers were connected to 

the objectives which were: (1) to find the maximum WIP fluctuation under a given 

vehicle quantity and (2) to determine the vehicle quantities to minimize the WIP 

fluctuation, as well as the probability of the average WIP exceeding the upper bound.  

The method for this model was based on the monotonicity of the objective functions 

[8]. An improvement plan in an existing automotive plant increased the production 

capacity by  accelerating the cycle time target [9]. For a single specific machine that 

can produce multiple products in make-to-order manufacturing plants, a simple 

deterministic model to determine the capacity and its level of utilization was 

developed. Processing time, set-up time, product defective rate, and maintenance 

downtime were the variables integrated in the model [10]. A model that predicts 

throughput and material flow requirements with a focus on designing flexible capacity 

under different scenarios was developed by analyzing the capacity of the plant [11]. 

An important decision regarding the selection of the optimal quantity and 

portfolio of product-dedicated and flexible capacities are to be made by firms when 

planning for a new manufacturing system that can produce several products over a 

planning horizon. The unfavorable effects of demand uncertainties may be alleviated 

by flexible systems; however, compared to dedicated systems, they require higher 

investment costs. Numerical studies were performed to provide insights on how these 

decisions are affected by the factors, such as the investment costs, product 

revenues, demand forecast scenarios and volatilities over the planning period.  The 

optimal capacity selection problem was formulated for this study [12]. To explore 



 

29 
 

optimal internal pricing and capacity planning, for a service facility with finite buffer 

capacity, an economic model was developed. The jobs that arrive when the system is 

full, will be rejected because of the limited buffer capacity. The system administrator 

was given two separate prices for accepted and rejected users at any desired 

demand level by setting a sufficient condition. This desired demand level becomes 

the unique equilibrium of the system. For the marginal capacity pricing to be optimal, 

another necessary and sufficient condition was set [13]. Many basic system design 

decisions, such as selecting a manufacturing technology for each product type 

(process selection), determining maximum production levels of each product type 

(capacity planning), and locating production resources and routing of products to 

required resources (facility layout), are required in planning a manufacturing system. 

The importance of integrating these decisions was examined and the advantages of 

the integrated approach were illustrated. The structure of the suggested integrated 

model showed how the overall problem was decomposed, as well as the interactions 

between the decision problems [14]. Capacity Oriented Analysis and Design of 

Production Systems [15] is a book which has many chapters on the topic of capacity. 

An equation to calculate the production capacity specific to a product (acid-resistant 

wares) is given in ideal conditions [16]. 

Capacity is defined under three categories: design, effectiveness and actual 

capacity. Design or maximum capacity is the output that an operation can produce 

continuously, at the maximum rate without stopping. Effective or available capacity 

considers how the operation will run on a long-term basis, to include all planned 



 

30 
 

stoppages. Actual capacity also includes unplanned stoppages. Actual output is 

effective capacity minus unplanned losses.  Therefore, the operation which is working 

its assets efficiently is minimizing unplanned losses [17]. 

 Data on capacity utilization in the U.S. economy is gathered and published by 

the Federal Reserve. Capacity utilization tends to fluctuate with business cycles; 

firms adjusts production volumes in response to changing demand. The Fed has 

published capacity utilization figures since the 1960s, spanning a number of 

economic cycles. In the late 1960s and early 1970s, all-time-high capacity utilization 

levels approaching 90% were achieved. The deepest declines in capacity utilization 

occurred in 1982 and 2009, when it fell to 70.9% and 66.7%, respectively [18]. See 

[19], [20] and [21] for more information. 

2.2 Scaling up from Research to Production 

While the ultimate goal is to go directly from process optimization to full scale 

plant, the pilot plant is generally a necessary step. Reasons for this critical step 

include: understanding the potential waste streams, examination of macro-processes, 

process interactions, process variations, process controls, development of standard 

operating procedures, and others. The information developed at the pilot plant scale 

allows for a better understanding of the overall process, including side processes. 

Therefore, this step helps to build the information base so that the technology can be 

permitted and safely implemented [22]. This paper focused on the specific needs of 

the operating plant to allow a new technology to be implemented. 



 

31 
 

The pathway of temperature increase during reaction, as well as adjustment of 

operating condition conducted for laboratory experimental data in order to produce a 

good quality of paste-glue was monitored while scaling up production from a 1,000ml 

reactor to a 500L pilot-scale reactor and a 1,500L near commercial scale reactor. 

Critical parameters for a good product quality, such as viscosity and ceiling 

temperature of the reaction, which are very crucial in order to give optimum operating 

condition as well as some scaling up parameters, have been found. The synthesis 

method of paste-glue production was selected and found the range of the parameters 

in order to produce a very good quality of paste-glue in pilot scale and near 

commercial scale [23]. In [24] the author explains how close R & D interaction was 

needed to design the production facility. 

The review in [25] presents the challenges of up-scaling lentivirus production 

and processing approaches, novel systems for overcoming these issues, and the 

quality assessments recommended for producing a clinical grade lentiviral gene 

therapy product. In a chapter of the book [26], the authors have discussed different 

production hosts, process development, fermentation process, scale up, challenges 

in the scale up of biopharmaceuticals production, purification of biopharmaceuticals, 

and recent developments on scale up of biopharmaceuticals production. When 

transferring film manufacturing from lab-scale to continuous mode, film compositions, 

processing conditions and suitable characterization methods have to be carefully 

selected and adopted [27]. In [28], the authors deal with the technology transfer from 



 

32 
 

a small-scale inkjet printing system to a pilot scale process by incorporating the same 

print head assembly into a continuous ODF production process. 

In [29], a scale-up analysis of a dual cell photo reactor based on a kinetic 

radiation model and mass balance of reactants is presented. A kinetic model that 

includes phenomenological based parameters is developed to evaluate the reaction 

rate under operational conditions of a photo-reactor. The analysis is performed for six 

different scale-up ratios with three different constraints for each case. The analysis is 

followed by an exergoeconomic study in which two case scenarios of a hydrogen 

production plant, with and without oxygen production for three different production 

capacities, are considered. In [30], the authors explain how the reaction conditions 

were transposed from small reactor capacity to a large capacity reactor. The relevant 

parameters, which affect the yield and reaction time, are studied. Also [31] and [32] 

explain how research to production of molecular beam epitaxy was carried out. An 

additional source is [33] where the author explains how they overcame scale-up 

limitations of ultrasonic processing. 

All of the articles/books mentioned above take a step by step approach to 

reach the full production. The presented approach in this dissertation goes from 

research to full production of the product quantity demanded directly, thus eliminating 

the trial and error steps in between. 

2.3 Manufacturing System Design (MSD) 

An ideal MSD is one in which the design satisfies a given set of constraints by 

the selection of functional requirements. It is time variant; selection of specific sets of 



 

33 
 

functional requirements and constraints change the design. A subset of an entire 

manufacturing enterprise, as well as that of a production system, is a manufacturing 

system. Manufacturing enterprises consist of elements and the design of 

manufacturing systems are regarded to be complex. The various elements of 

manufacturing enterprise are machines, tools, material, people and information. The 

functional requirements (FRs) placed on the manufacturing system predicates the 

specific combination of a manufacturing system's elements [34]. There are four 

domains in the design world as per the axiomatic design approach. The domains are:  

the customer domain (customer attributes), the functional domain (functional 

requirements - FRs), the physical domain (design parameters - DPs) and the process 

domain (process variables - PVs). The construct of mappings among these domains 

is the design. Design requirements lead to conceptual design, followed by 

configuration design which will provide the detailed design. The design requirements 

are broadly classified as functional requirements and constraints. Determination of 

manufacturing operations, selection or initial design of machines that provide the 

required operations, determination of the type of manufacturing systems and 

identification of possible material handling systems, are included in the conceptual 

design. The conceptual design is refined by the configuration design where the 

machines are arranged into a system (layout design). All design parameters are 

refined at the detailed design stage and the final design is evaluated for 

implementation [35]. 



 

34 
 

A physical system is required to manufacture a product whether it is in volume 

manufacturing or in a job shop. The inputs for the manufacturing system design are 

the task models (product/part/assembly/planning). MSD is comprised of workstations, 

layout planning, and throughput strategy. Product design is translated into 

manufacturing requirement by process design which leads to time and cost of 

manufacture. Both part planning and assembly planning require task analysis and 

hence their underlying unity. The interactions involved between the customer’s 

requirements and the product’s functional attributes makes product planning 

complex. The workstations provide concurrency of tasks, in physical systems, which 

is a requirement for the volume manufacture. The quantity (volume) of the product to 

be manufactured, shift time and total time of overall tasks determines the number of 

workstations. In order to meet the demands of equipment and transfer systems, 

layout arrangements of workstations are required. A significant part of the overall 

system design is the layout in high volume manufacturing. Large capital, that could 

be more effectively utilized, is tied up when buffer levels are significantly high. Lean 

manufacturing reduces the buffer levels to a minimum or eliminates them completely. 

It is an extension of just-in-time manufacturing which resulted from the need for 

continuous flow of production [36]. 

The high-level structure of the manufacturing system configuration is a 

collection of interacting components. The reconfigurable manufacturing systems are 

the most flexible and productive because they are based on standard/template sub-

systems and elements [37]. A manufacturing system, which satisfies the strategic 



 

35 
 

objectives of a company, needs to be designed according to the following four 

precepts: (1) Separate objectives clearly from the means of achievement, (2) Relate 

low-level activities and decisions to high-level goals and requirements, (3) 

Understand the interrelationships among the different elements of a system design, 

and (4) Effectively communicate this information across the organization. In TPS [3], 

the objectives and means are not clearly distinguished, its focus is on the physical 

tools (the means), the systems solution of which is predefined. The decisions about 

manufacturing system design are taken by relating high level design decisions to 

important system characteristics such as operational costs. How lower-level design 

decisions, such as equipment design and operator work content, affect system 

performance is not communicated. The low-level decisions to high-level system 

objectives are traced by the frameworks developed but they do not state the means 

to achieve the given objectives. Moreover, a strong design link between strategic 

objectives and the operational means to achieve them is not provided by these 

frameworks [38]. 

A closer integration of design, layout, process, and manufacturing within and 

across companies are needed (forced) because of the changes in manufacturing 

processes and the introduction of new materials. The paper examined extra structural 

features apparently added for Design for Manufacturability (DFM) purposes by a few 

manufacturers with similar products by comparing the old and the new products [39]. 

The same idea was presented in [40]. Phases where plans are implemented into 

reality bring forth problems in production management for the first time.  Some 



 

36 
 

examples are: when production commences, prototypes enter manufacturing and 

deliveries are expected. A study was conducted to see how the management can 

anticipate probable near future pitfalls by applying advanced visualization techniques 

to the existing information available with the companies. The problems identified in 

the analysis helped the companies to react in advance [41]. 

The focus of these research studies has been on the product/part or on 

improving the process and Manufacturing System Design (MSD) from a traditional 

perspective (physical design). None of the studies have researched applying the 

concept of reverse engineering to designing a manufacturing system. This 

dissertation will try to fill the gap in the application of reverse engineering to the 

design of a manufacturing system. This research focuses on the operational MSD 

rather than the physical MSD. 

2.4 Operational Excellence, Variations, Disruptions and Flow 

Operational Excellence (OE) is a philosophy of the workplace where ongoing 

improvement in an organization is undertaken by problem solving, teamwork, and 

leadership. The improvement is made by focusing on the customers' needs, keeping 

the employees positive and empowered, and continually improving the current 

activities in the workplace [42]. Some of the core principles of OE are: embrace 

scientific thinking, focus on the system process and think systematically [43]. Some 

of the methodologies used in OE are lean manufacturing, Six Sigma (identify and 

eliminate variation) and kaizen [43]. 



 

37 
 

Variability exists in all production systems and can have an enormous impact 

on performance. Variability is anything that causes the system to depart from regular, 

predictable behavior. Variability causes performance degradation by inflating one or 

more of three buffers (stock, time, capacity). The two primitive elements that make up 

any production system are stocks and flows. A flow represents materials or resources 

moving through the transformation process and is essential. Flows refer to the 

transfer of jobs or parts from one station to another. A stock represents material or 

other resources waiting for transformation. Inventory buffers are kept in stocks while 

the other two buffers (time and capacity) are related to flows. Demand and 

transformation are two essential parts of a production system and are themselves a 

type of flow; demand is an inflow whereas transformation is an outflow. If demand 

and transformation are not perfectly aligned, there will be one or more buffers. The 

usual cause of misalignment between demand and transformation is variability. 

Variation is a measure to determine how the system conforms to the standards. The 

most prevalent sources of variability, which affects the effective process time in 

manufacturing environments, are: (1) Natural variability, which includes minor 

fluctuations in process time due to differences in operators, machines, and material, 

(2) Random outages, (3) Setups, (4) Operator availability and (5) Recycling [3]. 

The effects of variability in the overall production line can be characterized by 

process time variability and arrival variability. The variability, in the worst case, is 

completely predictable and results from bad control; while the variability in the 

practical worst case is due to unpredictable randomness. Controllable variations 



 

38 
 

occur as a result of decisions, whereas random variation is a consequence of events 

beyond control.  Variance (σ2) and standard deviation (σ) are measures of absolute 

variability [3]. All parametric distributions will have variance and mean. 

Lean Manufacturing often talks about reducing wastes by eliminating Non-

Value-added tasks. Lean Manufacturing is an integrated socio-technical system 

which eliminates waste by using a systematic method [44]. In the current practice of 

lean implementation, most of the lean tool's focus is on time and material through 

techniques which rarely captures variability and tries to eliminate different sources 

[45]. Lower Throughput, congestion, high WIP levels, and longer lead times are a few 

of the examples of the effect due to variation. Variability is the enemy of 

manufacturing and the source of many of its problems [3], [45]. Design performance 

fluctuations could be caused by variations in the manufacturing process. Variations, if 

not accounted for, can cause a design to fail to meet performance and/or correctness 

criteria [46]. To identify the source of variation and to reduce it, a new technique 

similar to Value Stream Mapping (VSM) called Variability source Mapping (VSMII), 

has been developed. VSMII captures variation in terms of time and flow [45]. 

Production variability is less for many machines in series than in a single machine 

system in a production system with the same production volume and reliability 

characteristics as a longer line [47]. 

The variability of cycle times in semiconductor manufacturing lines is reduced 

by diminishing the magnitude of overtaking through appropriate sequencing rules 

[48]. By using sequencing rules like first in first out (FIFO), earliest due date, critical 



 

39 
 

ratio and closest to completion by step overtaking can be reduced, but the reduction 

does not always lead to reduction in variation in cycle times. The variance of cycle 

time is also caused by the lots repeatedly returning at different stages of their 

production to the same service stations for further processing, consequently creating 

considerable competition for machines [49]. This leads to variation in cycle time at 

the workstations; to reduce this variation in cycle time the authors have proposed 

scheduling policies called fluctuation smoothing policies. By using these scheduling 

policies, the variance in cycle time can be reduced. Production variability, due to 

random disturbances, cause the observed production rate to be different from its 

average value; the evidence of which is in industries [50]. By using the method to 

estimate the problems causing the variability in a multistage manufacturing system, 

the relation between the output variance, the machine reliability parameters and the 

buffer sizes are obtained [50]. 

There are a few studies conducted about manufacturing cycle time (CT), lead 

time (LT) and also about queues in series. An approximation approach, based on 

observed properties of the behavior of tandem queues to find the queue times with 

variability in the line, is developed [51]. The waiting time (and hence manufacturing 

lead time) distribution and the mean performance measures are derived using the 

factorization principle [52]. A mathematical model focused on the manufacturing lead 

time and the utilization efficiency is derived [53]. An analytical model, which provides 

insights into the connection between the parameters (such as process time, arrival 



 

40 
 

rate and placement of the inspection station) and performance (throughput, 

manufacturing lead time) of a manufacturing system is presented [54]. 

Disruptions can cause significant impact on the performance of a production 

system and can also lead to delays in delivery dates, impacting customers and 

widening delays in delivery dates, which also impacts customers and wider business 

functions. Some examples of disruptions are quality problems, resource breakdowns, 

material unavailability, order changes and rush orders [55]. In a production line that 

deals with interruptions due to lack of resources and product quality, it is necessary 

to analyze the steps for balance between Lean and resilience [56]. One of the major 

causes of disruption in production line is due to equipment. Absence of proper 

maintenance is one of the main reasons for disruptions and unavailability in the 

production equipment. Maintenance should be considered as a key variable in the 

construction of operations and infrastructure strategies and their varied impact on the 

production line should be considered [57]. Disruption is a state during the execution 

of the current operation, where the deviation from plan is sufficiently large, thus the 

plan has to be changed substantially. Just-in-time approach to production, aiming at 

increasing productivity and decreasing the cost of production, gives rise to an 

increased demand for robustness in plans and calls for enhanced tools to handle 

disrupted situations [58]. 

For the production to reach the necessary quantity, required quality, in the 

necessary time and with the most reasonable cost, the ways in which such an 

organization uses the production schedule is to be understood and thus organize re-



 

41 
 

scheduling, scheduling and workflow while considering the disruptions in the 

schedules [59]. When disruption occurs, resistance against any change and 

rescheduling from the previous program may be shown by the internal system factors 

(e.g. operators)  [60]. In airline management, the operators at the operations control 

center carry out the disruption management process in three steps: (1) They 

formulate the problem qualifying it in terms of resolution time, passengers impacted, 

delay propagation through the network and others. (2) Different options to resolve the 

situation are listed and ranked. (3) The most suitable solution is implemented [61]. 

When disruption is caused by an employee, it is usually due to heavy workload, labor 

shortages, lack of information and personal preparation which cause extensive 

interruptions in the workflow and delays in the schedule [62]. System Dynamics is a 

simulation modeling technique that was specially designed to model and explore 

feedback. System dynamics has been used for the analysis of cost or delivery 

overruns on large projects. The system dynamics model consists of three main work 

functions: design engineering, methods industrial engineering, and manufacturing 

[63]. 

2.5 CT, LT and TAKT time 

CT is the actual time to do the processing; LT is set by the management 

whereas TAKT time is set by the customer. Value Stream Mapping is a lean tool 

used to greatly reduce cycle time, as well as lead time. An application of VSM in an 

OEM is provided by the authors [64]. VSM, as a lean tool, has been used to reduce 

the cycle time by identifying and eliminating wastes in a facility with similar or 



 

42 
 

identical product routing. VSM, along with Methods-Time measurement (MTM), are 

used to reduce lead time and increase the productivity in an assembly and 

production-logistic processes. A practical example is used to highlight the redesign of 

assembly workplaces and the redesign of production logistic processes to reduce the 

inventory/ lead times and increase the productivity by standardization of process [65]. 

Efficient Scheduling policies are also used to reduce mean and variance of cycle time 

in a semi-conductor manufacturing plant. Use of new class scheduling policies, called 

fluctuation smoothing policies, helped achieve the best mean cycle time and 

deviation of cycle time in all the configurations of the plants that were tested [49]. 

Cycle time reduction has also been studied in a semi-conductor wafer fabrication 

facility in which the method developed by authors managed to reduce the cycle time, 

increase the capacity and reduce the WIP [66]. Agility is the power to cope with the 

variability and uncertainty in the market or virtual corporation. Two main factors 

affecting the supply chain are waste Total Cycle Time (TCT) and waste information 

flow [67]. 

The information enriched supply chain can reduce the lead times for 

information and material flow and the total cycle time will reduce if the supply chain is 

more agile. The factors that are controllable by the company and affect the Total 

Cycle Time (TCT) are purchasing cycle time, design and manufacturing cycle time, 

inbound transportation cycle time and outbound transportation cycle time. Various 

aspects related to the above mentioned factors are discussed and theories are 

developed to reduce the TCT [68] . Short manufacturing cycle times are required for 



 

43 
 

a firm to meet short lead times, without excessive inventories [69]. A good way to 

reduce the cycle time without increasing the company’s expenses is by including an 

inventory of spare components. This reduces the time for maintenance and repairing, 

thereby reducing the cycle time. The proper amount of spare parts inventory, such 

that the inventory costs are justified, is discussed by the author based on the 

following five factors: (1) mean time to failure for both single and multiple critical 

components, (2) critical component replenishment lead times, (3) workstation arrival 

rates and variances, (4) critical component annual holding costs, and (5) hourly 

revenue increases for cycle time reductions [69]. 

2.6 Effects of Variation, Disruption on CT 

A discussion about the Factory Physics [3] approach to study the effect of 

variation on CT is given here; increasing variation increases CT, placement station 

with variations and propagation of variation impact CT. The formulations in the theory 

are used in the development of the mathematical model for TF in Section 3.2 of this 

dissertation. Variation and disruptions are closely connected in the literature through 

the availability of resources (such as equipment, people and material). Line 

performance is measured by SL as shown in Equation 2.1 [3]. 

𝑆𝐿 = 𝑃{𝐶𝑇 ≤  𝐿𝑇} 2.1 

Capacity is an upper limit on the TH of a production process. TH can be 

increased by increasing the utilization of the bottleneck or its rate. Bottleneck is 

defined as the busiest station (highest utilization), not necessarily the slowest station. 

Utilization of the bottleneck can be increased by buffering it with WIP [3]. It can also 



 

44 
 

be increased by reducing the variations and disruptions in the bottleneck; as well as 

by enhancing capacity. TH of a line is given by Equation 2.2 [3].  

𝑇𝐻 = 𝑏𝑜𝑡𝑡𝑙𝑒𝑛𝑒𝑐𝑘 𝑢𝑡𝑖𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛 · 𝑏𝑜𝑡𝑡𝑙𝑒𝑛𝑒𝑐𝑘 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑒 2.2 

Utilization (𝑢) (Equation 2.3) is the ratio of the arrival rate (ra) to the effective 

production rate (re). The effective production rate is defined as the maximum average 

rate at which the workstation process parts considering the effects of failures, setups 

and all other detractors that are relevant [3]. 

The critical WIP (W0) of the line [3] (Equation 2.4) is the WIP level for which a 

line with given values of rb (bottleneck rate) and T0 (raw process time), achieves 

maximum throughput with minimum cycle time, when there is no variability. T0 of the 

line is the sum of the long-term average process times of each workstation in the line. 

It is the average time a single job takes in the empty line [3]. 

𝑊0 = 𝑟𝑏 · 𝑇0 2.4 

Stable system requires the input to the system not exceed its capacity. The 

capacity of a line must be at least as large as the arrival rate to the system. When a 

production system has variability, then, a sequence of events will cause the system 

bottleneck to starve (run out of WIP) regardless of the WIP level. A steady state 

system avoids this. In steady state, all plants release work at an average rate which 

is strictly less than the average capacity. If there is no limit as to how much WIP can 

be in the system, both CT and WIP go to infinity as utilization approaches one.  If a 

station increases utilization without making any other changes, average WIP and CT 

 𝑢 =  
𝑟𝑎

𝑟𝑒

 2.3 



 

45 
 

will increase in a highly nonlinear fashion [3]. If V is the variability, U the utilization 

and T the time respectively, then the mean time spent in queue (CTq) is given by 

Kingman’s Equation [70], [3] in 2.5 . 

𝐶𝑇𝑞 = 𝑉 · 𝑈 · 𝑇 2.5 

The coefficient of variation (CV) is the ratio of standard deviation (SD (σ)) to 

the mean (µ) as shown in Equation 2.6 [3]. The CV is also denoted as σ/t if the 

random variable considered is time (here t denotes the average of time). 

 The processes are classified as low variability (LV), medium variability (MV) 

or high variability (HV) depending upon the value of the CV. LV processes will have 

their CV< 0.75, whereas MV process will have 0.75 ≤ CV ≤ 1.33 and process will be 

HV when their CV > 1.33 [3]. The probability density function (pdf) of most LV 

processes are bell shaped (normal distribution) [3].   

One important measure of variability is in the process time.  Effective process 

time of a job at a workstation is the total time seen by a job at a station. If A is the 

availability of the machine, t0 is the natural process time, m= number of machines in 

parallel at that station and r0 is the natural capacity (rate), then mean effective 

process time (te) [3] is given by Equation 2.7 whereas the capacity or rate of 

workstation re [3] is given by Equation 2.8. The CV of effective process time (ce) [3] is 

given by Equation 2.9  where mr is the meantime to repair, c0 is the natural CV of the 

𝐶𝑉 =
𝜎

µ
 2.6 



 

46 
 

process and cr is the CV of repair. The three terms, in Equation 2.9 denote natural 

variability, random outages and variability of repair time respectively [3]. 

𝑡𝑒 =
𝑡0

𝐴
 

2.7 

𝑟𝑒 =  
𝑚

𝑡𝑒
= 𝐴 · (

𝑚

𝑡0
) = 𝐴 · 𝑟0 2.8 

𝑐𝑒
2 =  𝑐0

2 + 𝐴(1 − 𝐴) ·
𝑚𝑟

𝑡0
 +  𝑐𝑟

2 · 𝐴(1 − 𝐴) ·
𝑚𝑟

𝑡0
 2.9 

The cycle time of the queue (CTq) for a single machine station is given by 

equation 2.10 where ca = CV of inter arrival time. The CTj is the sum of CTq and te 

(equation 2.11) which leads to Equation 2.12 . The CTline is the summation of the CTj 

of all stations in the line [3], [54] as shown in Equation 2.13. Any overlap in time 

between stations is to be deducted. 

𝐶𝑇𝑞 =
𝑐𝑎

2+𝑐𝑒
2

2
·

 𝑢

1− 𝑢
· 𝑡𝑒        where 𝑢 =  

𝑟𝑎

𝑟𝑒
 2.10 

𝐶𝑇𝑗 = 𝐶𝑇𝑞𝑗  +  𝑡𝑒𝑗 2.11 

𝐶𝑇𝑗 =  
 (𝑐𝑎𝑗

2 + 𝑐𝑒𝑗
2 )

2
·

𝑢𝑗

1 −  𝑢𝑗
· 𝑡𝑒𝑗 +  𝑡𝑒𝑗 2.12 

𝐶𝑇𝑙𝑖𝑛𝑒 =  ∑ 𝐶𝑇𝑗

𝑚

𝑗=1

 2.13 

In Equation 2.12 the average queue and CT grows to infinity as utilization 

approaches 100 percent.  Queues never become infinite in the real world because of 

limitations of space, time or operating policy. Whenever any of the limits are 

reached, the arrival process is stopped. This procedure is called blocking. By 

employing blocking, the stream of work from the previous station to the station where 

the limit is reached, is cut off [3]. 



 

47 
 

The G/G/1 queuing model (CT equations has CTq related to the queue) is 

more appropriate for manufacturing systems as noted by the authors of Factory 

Physics [3]. In G/G/1 queue, the system with a single server, the inter-arrival times 

and service times have a general distribution. When workstations are fed by 

upstream stations whose process times are not exponential, the inter-arrival times 

also are not likely to be exponential. Process times are seldom exponential [3]. The 

variance and mean of the normal, triangular and uniform distributions, which are 

common in manufacturing operations, are provided in Table 20 Appendix B. Also 

presented in Appendix B ([71] and [72]) are some of the common distributions. An 

example of the calculations of mean and CV for a triangular distribution are given in 

Table 21 Appendix C. 

When the variability at one station affects the behavior of other stations in a 

line, it is referred to as flow variability. If an upstream workstation has highly variable 

process times, the flows it feeds to downstream workstations will also be highly 

variable. The variability in flow is characterized by arrivals and departures. Variability 

in departures from a station is the result of both variability in arrivals to the station 

and variability in the process times. The relative contribution of these two factors 

depends on the utilization of the workstation. The actual process time typically 

represents only a fraction of the total CT. The majority of the remaining time is spent 

waiting for various resources/activities [3]. This flow variability is also referred to as 

propagation of variation. 



 

48 
 

If CV of arrival at a particular process is denoted by ca, the CV of the process 

is denoted by ce, the CV of departure from that process is denoted by cd and the 

number in the subscript denotes the location, then from [3] Equations 2.14 to 2.16 

are obtained. This can be generalized as shown in Equation 2.17. 

𝑐𝑑1
2 =  𝑢1

2 𝑐𝑒1
2 + (1 − 𝑢1

2) 𝑐𝑎1
2  2.14 

𝑐𝑑2
2 =  𝑢2

2 𝑐𝑒2
2 + (1 − 𝑢2

2) 𝑐𝑎2
2  2.15 

𝑐𝑎2 =  𝑐𝑑1 2.16 

𝑐𝑎𝑗
2 =  𝑐𝑑𝑗−1

2 =  𝑢𝑗−1
2  𝑐𝑒𝑗−1

2 + (1 − 𝑢𝑗−1
2 ) 𝑐𝑎𝑗−1

2  2.17 

This shows that the effect of variation propagates through the system. 

Variation at a station will affect the next immediate station and subsequently the 

whole system. If the arrivals at the first station can be tightly controlled (ca1 = 0) then 

the departure variation from the first station will be the result only of the variation in 

the process itself and the utilization of that station.   

Queue time is impacted by the utilization of the machines, the process time, 

the coefficient of variation of arrival time, as well as process time. This queue time in 

turn affects the CT. Also, the availability of resources determines the effective 

process time. The coefficient of variation of arrival time at a station is dependent on 

the coefficient of variation of departure time of the previous station. Variability and 

disruptions affect the CT and hence the TH. The effect of variability can be reduced 

by cutting down the disruption in the process and by reducing queue time. WIP and 

CT can be reduced by anything that enables jobs to move from one workstation to 

the next, with less waiting [3].  



 

49 
 

If a station has to wait for a number of batches to be processed at the previous 

station, then there is a Transfer Wait Time (TWT) which is based on the batch 

processing equation of Factory Physics [3]. If the number of batches at a station is 

denoted by X2j and the utilization of the station is denoted by uj, then the TWT is 

given by Equation 2.18. 

𝑇𝑊𝑇 =
𝑋2𝑗 − 1

2 ∗ 𝑢𝑗
· 𝑡𝑗 

2.18 

The TWT will be exactly equal to the real wait time for batching, only when the 

utilization of the station is 50%; if the utilization is more than 50%, then the TWT will 

be less than the actual batch wait time and vice versa. If there is no batching (X2j =1), 

the TWT will be zero. 

When there is batch processing in any of the station(s) and the subsequent 

stations process multiple batches (X2j) from the previous station at the same time 

(one process for multiple batches), then the CT of the station with batch processing 

will be (applying Equation 2.18) as given in Equation 2.19. 

𝐶𝑇𝑗 =  
(𝑐𝑎𝑗

2 + 𝑐𝑒𝑗
2 )

2
·

𝑢𝑗

1 −  𝑢𝑗
· 𝑡𝑗   +  

𝑋2𝑗 − 1

2 𝑢𝑗
· 𝑡𝑗 +  𝑡𝑗  

2.19 

TH is dependent on cycle time (CT) and Work-In-Process (WIP) as per Little’s law [5] 

(Equation 2.20). A system with short CT and low WIP is preferable. 

𝑇𝐻 =  
𝑊𝐼𝑃

𝐶𝑇
 2.20 

Reduction of variation will result in a reduction of CT. The theory and the 

equations discussed above are used and modified in the computational development 

for TF in Section 3.2. The appropriate type of probability distribution in the 



 

50 
 

manufacturing operations will be represented by its CV (ratio of the standard 

deviation to the mean) in the time to finish Equations 3.17 and 3.19. 

2.7 Summary of Literature Review 

The literature review concludes that the effect of variation and disruption 

issues on the timely completion of processes are not considered when systems are 

designed but are considered as continuous improvement projects once the 

processes are running. Manufacturing systems are designed primarily with focus on 

machines and their physical layout, not on the operational aspects. Production 

reaches the estimated demand quantity (capacity) in stages through step by step 

scaling up activities; there is no systematic approach to scale up directly from the 

research to the quantity demanded. Capacity of a plant is not computed based on the 

timely completion of the processes and comparing with the time allocated. The 

concept of TF vs TA combination is not studied and as such has not been used; if the 

processes get completed within the time available, then there is sufficient capacity to 

meet the demand. The various studies about the effect of detractors (such as 

variations, disruption and flow related issues) focus on CT vs LT combination. The 

effect of variation and disruption on CT is available in the literature. Additional 

literature review (Appendix D) was done to check whether re-engineering or reverse 

engineering has been used in designing operational manufacturing systems. It was 

used in product/process design or for making improvements but not for designing 

manufacturing systems. This dissertation has not applied it either but it could be 

applied for future expansion work of this dissertation. 



 

51 
 

3  

METHODOLOGY 

In this chapter, the framework to study and design operational manufacturing 

systems is discussed in detail. The development of the mathematical model which 

computes the time to finish (TF) is given. The methodology considers the effect of 

variation and disruption issues at the design stage itself focusing on the sources 

which are the “four critical resources” [6] explained in Section 1.5.1.  

The focus in the literature is on the cycle time (CT) and lead time (LT) as 

shown in Section 1.3 and Chapter 2. From a CT perspective, the key manufacturing 

metrics are CT, throughput (TH), Capacity and Service Level (SL). The theory about 

CT and its formulation was discussed in Section 2.6 and the TF concept is based on 

this theory.  This model concentrates on the TF concept, which is a novelty and a 

main contribution of this dissertation. In government manufacturing, since the 

facilities are often shared between different products, time allocation for the 

production of a particular product is mostly very rigid; also because of security issues 

the materials are not allowed to be left in the manufacturing areas of the facility. 

Hence quantification of TF is useful in checking whether the production processes 

can be completed within the time allocated (TA). The computation of TF will also help 

in determining whether the suggested operations are practically feasible. Moreover, 

time allocation might not be continuous; it could be in different periods. The detailed 

definition of CT, LT, TA and TF are given in Section 1.3. 

The key manufacturing metrics from the TF perspective are TF, TH, Capacity 

and SL. The formulation of TF developed for this dissertation is discussed in Section 



 

52 
 

3.2. The T