Product Cost Estimation: Theoretical Development and Industrial Validation

Product and Manufacturing Cost Estimation:
Theoretical Development and Industrial Validation
By
Adnan Niazi
This thesis is submitted to the Department of Mechanical Engineering,

School of Physical Sciences and Engineering, King’s College London,
University of London, for the Degree of Doctor of Philosophy.
1
To
My Mum and Dad

My Wife
and
My Children
2
Abstract
Product Cost Estimation (PCE) deals with predicting the cost of a product before it is
manufactured. Due to market competitiveness, there is a need to predict product costs
early and accurately. However, the available methods of cost estimation compromise
accuracy in an attempt to deliver early results. Conversely, the accuracy can only be
fully achieved once the design and process planning details are available by which time
cost estimation will be too late. The main aim of the work is to develop a methodology
for early and accurate estimation of a product’s cost without relying on design and
process planning details.
The main contributions of the thesis are as follows. First, following a comprehensive
literature review of the available techniques, an extensive hierarchical classification
system is developed. The classification is based on categorizing the techniques into
qualitative and quantitative with further subdivisions down to four levels. The
developed system identifies that qualitative techniques deliver early results and
quantitative techniques are known for accuracy. The review also identifies that overall
accuracy greatly depends on accurate estimation of overheads. Secondly, based on the
feedback from the review and the proposed classification system, a new method of
overhead estimation based on separating time- and material-dependent overheads is
developed with improved accuracy. Thirdly, a comprehensive mathematical model
based on combining the attributes of early and accurate estimation from the qualitative
and quantitative techniques is developed and called a Hybrid Model. The developed
model is optimized through the introduction of the cost deviation indices. Fourthly, the
deviation indices are modelled considering past product cost details, inflation and other
deviations in order to predict future costs early and accurately without requiring product
design and process planning details.
3
Abstract
The developed models are validated by industrial trials on two distinct global locations
adding further towards understanding the implications of geographical locations on
aspects of cost control.
4
Table of Contents
Abstract ........................................................................................................................ 3
Table of Contents.......................................................................................................... 5
List of Figures............................................................................................................. 11
List of Tables.............................................................................................................. 14
List of Notations ......................................................................................................... 16
List of Acronyms ........................................................................................................ 22
Acknowledgements..................................................................................................... 24
C hapt e r 1 Introduction ......................................................................................... 27
1.1 Overview..................................................................................................... 27
1.2 Problems identification................................................................................ 29
1.3 Aims and objectives .................................................................................... 33
1.4 Thesis structure ........................................................................................... 35
1.5 Conclusions................................................................................................. 38
C hapt e r 2 Background and Related Work ............................................................ 39
2.1 Introduction................................................................................................. 39
2.2 Manufacturing and cost control ................................................................... 41
2.3 Cost estimation............................................................................................ 45
2.4 Cost estimation in the early design stages .................................................... 47
2.5 Cost estimation for specific applications...................................................... 48
2.5.1 Cost estimation for a specific segment in a production cycle................ 49
2.5.2 Cost estimation for specific machining and manufacturing processes... 50
5
Table of Contents
2.5.3 Cost forecasting for specific parts and products ................................... 50
2.5.4 Cost estimation for generic systems ..................................................... 51
2.6 Conclusions................................................................................................. 53
C hapt e r 3 PCE Technique Classification System ................................................. 55
3.1 Introduction................................................................................................. 55
3.2 Development of hierarchical classification system (HCS)............................ 57
3.3 Intuitive cost estimation techniques ............................................................. 60
3.3.1 Case-based methodology ..................................................................... 60
3.3.2 Decision support systems (DSS) .......................................................... 63
Rule-based systems ......................................................................................... 65
Fuzzy logic approach....................................................................................... 68
Expert systems ................................................................................................ 69
3.4 Analogical cost estimation techniques ......................................................... 70
3.4.1 Regression analysis models.................................................................. 70
3.4.2 Back-propagation neural network (BPNN) models............................... 70
3.5 Parametric cost estimation techniques.......................................................... 71
3.6 Analytical cost estimation techniques .......................................................... 73
3.6.1 Operation based approach .................................................................... 73
3.6.2 Breakdown approach ........................................................................... 75
3.6.3 Tolerance-based cost models ............................................................... 76
3.6.4 Feature-based cost estimation .............................................................. 78
3.6.5 Activity-based costing (ABC) system .................................................. 79
3.7 Conclusions................................................................................................. 81
C hapt e r 4 MRO and TRO Estimation Methods .................................................... 87
4.1 Introduction................................................................................................. 87
4.2 Cost estimation methodology at the selected company................................. 91
6
Table of Contents
4.2.1 Material cost estimation ....................................................................... 92
4.2.2 Direct labour costs ............................................................................... 93
4.2.3 Overheads estimation........................................................................... 93
4.3 Proposed methodology for overheads estimation ......................................... 95
4.3.1 MRO estimation model........................................................................ 95
4.3.2 TRO estimation model......................................................................... 98
4.4 Model implementation and validation........................................................ 101
4.5 Conclusions............................................................................................... 112
C hapt e r 5 PCE Hybrid Model ............................................................................ 114
5.1 Introduction............................................................................................... 114
5.2 Product cost and modelling approach......................................................... 116
5.3 Direct cost elements .................................................................................. 122
5.3.1 Direct material costs .......................................................................... 123
5.3.2 Direct labour...................................................................................... 126
Labour units .................................................................................................. 128
Labour rate.................................................................................................... 130
5.4 Indirect cost elements ................................................................................ 132
5.4.1 Processing cost .................................................................................. 132
Processing units............................................................................................. 133
Processing rate .............................................................................................. 134
5.4.2 Material dependent cost ..................................................................... 135
5.4.3 Tooling cost....................................................................................... 136
Machine tool rate........................................................................................... 137
Labour tool rate............................................................................................. 138
5.4.4 Building space cost ............................................................................ 139
Building space rate........................................................................................ 140
7
Table of Contents
Manufacturing space ..................................................................................... 141
5.5 Production overheads ................................................................................ 142
5.6 Conclusions............................................................................................... 143
C hapt e r 6 Industrial Implementation and Analysis of the PCE Hybrid Model .... 145
6.1 Introduction............................................................................................... 145
6.2 HMI algorithm and the implementation methodology................................ 148
6.3 PCE at the company .................................................................................. 157
6.3.1 Information and details ...................................................................... 157
6.3.2 Material cost estimation ..................................................................... 160
6.3.3 Labour cost estimation....................................................................... 163
6.3.4 Overhead estimation .......................................................................... 165
6.4 Implementation of the PCE Hybrid Model................................................. 168
6.4.1 Material cost estimation ..................................................................... 169
6.4.2 Labour cost estimation....................................................................... 171
6.4.3 Processing cost estimation ................................................................. 173
6.4.4 MDC estimation ................................................................................ 175
6.4.5 Production overheads estimation........................................................ 177
6.5 Conclusions............................................................................................... 179
C hapt e r 7 Comparisons and Validation Analysis................................................ 181
7.1 Introduction............................................................................................... 181
7.2 Preparations for comparisons..................................................................... 183
7.2.1 Labour and machine cost estimation .................................................. 185
7.2.2 Material cost and factory expenses..................................................... 187
7.3 Comparison analysis for product cost ........................................................ 189
7.4 Comparison analysis for cumulative costs ................................................. 202
7.5 Comparison analysis for product and production overheads....................... 206
8
Table of Contents
7.6 Cost breakdown analysis ........................................................................... 212
7.7 Conclusions............................................................................................... 223
C hapt e r 8 Conclusions ....................................................................................... 225
8.1 Summary................................................................................................... 225
8.2 Contributions............................................................................................. 231
8.2.1 Development of a technique classification system.............................. 231
8.2.2 Development of a decision support model (DSM).............................. 231
8.2.3 Development of time- and material-based overhead estimation

methodology...................................................................................... 232
8.2.4 Development of a PCE methodology for batch production................. 233
8.2.5 Development of cost deviation indices............................................... 234
8.2.6 Development of HMI algorithm and industrial implementation.......... 234
8.2.7 Comparison and validation ................................................................ 235
8.2.8 By – products..................................................................................... 235
8.3 Current Trends and Future Work ............................................................... 236
8.4 Concluding remarks .................................................................................. 239
Publications Arising from the PhD Study.................................................................. 241
References ................................................................................................................ 243
Appendix A Bill of Material (BOM)...................................................................... 261
A.1 Introduction............................................................................................... 261
A.2 BOM for Hammer Drill ............................................................................. 262
Appendix B Deviation Indices............................................................................... 274
B.1 Material cost deviation index..................................................................... 274
B.2 Labour cost deviation index....................................................................... 276
B.3 Processing cost deviation index ................................................................. 277
B.4 MDC deviation index ................................................................................ 277
9
Table of Contents
B.5 Tool cost deviation index........................................................................... 278
B.6 Building cost deviation index .................................................................... 279
B.7 PO deviation index .................................................................................... 279
10
List of Figures
Figure 1-1: Main area of research................................................................................ 29
Figure 2-1: Relationship between productivity and flexibility ..................................... 41
Figure 2-2: Flexibility and productivity of different manufacturing systems................ 44
Figure 3-1: PCE Preliminary Technique Classification................................................ 59
Figure 3-2: Flow Diagram of the Case-Based Approach for Cost Estimation............... 61
Figure 3-3: Decision support system approach to cost estimation ................................ 64
Figure 3-4 Cost Estimation Process Model Based On User Constraints ....................... 66
Figure 3-5: Classification of the PCE Techniques ....................................................... 84
Figure 3-6: Decision Support Model for cost estimation methodology selection.......... 86
Figure 4-1: Break down of Selling Price and Manufacturing costs .............................. 89
Figure 4-2: Cost estimation results for 25, 100, 200, 500 and 1000kVA
transformers............................................................................................ 104
Figure 4-3: Cost element values in DT and PT .......................................................... 106
Figure 4-4: TRO and MRO values breakdown for DT and PT................................... 107
Figure 4-5: Cost trends in (a) DT and (b) PT............................................................. 109
Figure 4-6: Cost trends comparison in DT and PT..................................................... 110
Figure 5-1: Pictorial representation of the mathematical model for product cost
estimation ............................................................................................... 119
Figure 5-2: Development of the Hybrid Model within the framework of the
technique classification system ............................................................... 125
Figure 5-3: Types of work centres............................................................................. 126
Figure 6-1: PCE Hybrid Model implementation phase .............................................. 150
11
List of Figures
Figure 6-2: Material cost estimation implementation process .................................... 151
Figure 6-3: Labour cost estimation implementation process ...................................... 153
Figure 6-4: Processing cost estimation implementation process................................. 154
Figure 6-5: MDC estimation implementation process................................................ 155
Figure 6-6: Manufacturing cost estimation implementation process .......................... 155
Figure 6-7: PO estimation implementation process ................................................... 156
Figure 7-1: The comparison of the actual costs against the estimated costs................ 190
Figure 7-2: Estimation error trends for the three methods (2003 – 2005) ................... 192
Figure 7-3: Percentage cost estimation variations from actual product costs. ............. 194
Figure 7-4: Estimation error trend across the product range....................................... 196
Figure 7-5: Error linearization for the results given by the company’s method .......... 198
Figure 7-6: Error linearization for the results given by the Jung’s method ................. 199
Figure 7-7: Error linearization for the results given by the Hybrid Model.................. 200
Figure 7-8: (a) Cumulative actual costs against total estimated values, (b)
estimation errors for cumulative costs..................................................... 203
Figure 7-9: Optimization for the estimation accuracy achieved by the Hybrid
Model against (a) the company’s method (b) the Jung’s model ............... 205
Figure 7-10: Estimation error trends for overheads using the three methods
(2003 – 2005) ......................................................................................... 208
Figure 7-11: Overheads estimation analysis for the cumulative values ...................... 209
Figure 7-12: Optimization achieved for overhead estimation based on (a) the
Company’s method (b) the Jung’s model ................................................ 211
Figure 7-13: Production cost (actual) break down analysis (2002 – 2005)
presented in values and percentage ......................................................... 213
Figure 7-14: Production cost elements trends ............................................................ 214
Figure 7-15: (a) Production and manufacturing costs (b) elemental costs effect
on production and manufacturing costs ................................................... 216
12
List of Figures
Figure 7-16: Overheads break down analysis (2002 – 2005) presented in values
and percentage........................................................................................ 217
Figure 7-17 Overhead elements trends (2002 – 2005) ............................................... 218
Figure 7-18: Production overheads breakdown (2002 – 2005)................................... 219
Figure 7-19: Production overhead elements trends analysis....................................... 220
Figure 7-20: Processing cost elements trends ............................................................ 221
Figure 7-21: MDC elements trends ........................................................................... 222
Figure A-1 Hammer drill .......................................................................................... 262
Figure A-2 Product structure (Hammer Drill)............................................................ 264
Figure A-3 Dismantled drill with assemblies and sub-assemblies.............................. 265
Figure A-4 Parts and components in product structure .............................................. 266
Figure A-5 Winding (Stator and rotor) and drive assembly ....................................... 266
Figure A-6 Driven assembly (with gear and shock absorber) and Drill/Hammer
switch ..................................................................................................... 267
Figure A-7 Trigger assembly .................................................................................... 267
13
List of Tables
Table 2-1: Summary of the published literature references for various

applications of PCE .................................................................................. 53
Table 3-1: The PCE techniques; key advantages, limitations and list of
discussed references.................................................................................. 85
Table 4-1: MRO for power and distribution transformers (2003-2004)........................ 97
Table 4-2: Budgeted time related overhead rate calculation (2003-2004)..................... 99
Table 4-3: Summary of TRO rates, MRO percentage fractions and overhead
rates for 4 years ...................................................................................... 100
Table 4-4: Cost estimation for 25kVA transformer.................................................... 102
Table 6-1: Industrial output values for the CMD (2002 – 2005) ............................... 158
Table 6-2: Product range at the SPSD ....................................................................... 159
Table 6-3: Yearly production quantities for the product range in the SPSD
(2001 – 2005) and Actual unit product cost for the given product
range (2002 – 2005)................................................................................ 160
Table 6-4: Actual material cost (Cumulative and per unit costs)................................ 161
Table 6-5: Estimated unit material costs for the given product range (2003 –
2005) ...................................................................................................... 162
Table 6-6: Aggregate labour rate calculation............................................................. 164
Table 6-7: Total man-hours and estimated labour costs (2003 – 2005) for the
given product range ................................................................................ 165
Table 6-8: Overhead costs for individual elements (2001 – 2005) ............................. 167
Table 6-9: Estimated product cost values (2003 – 2005)............................................ 168
14
List of Tables
Table 6-10: Material cost deviation index and estimated unit material cost
values (2003 – 2005) .............................................................................. 170
Table 6-11: Labour units calculation ......................................................................... 172
Table 6-12: Labour cost deviation index and estimated labour costs.......................... 173
Table 6-13: Estimation of processing units, deviation indices, rates and costs ........... 174
Table 6-14: MDC deviation index and estimated MDC per unit values ..................... 176
Table 6-15 PO fractions (actual & estimated) and deviation indices .......................... 177
Table 6-16: Estimated per unit values for manufacturing, PO and product costs........ 178
Table 7-1: Lead times, machine running costs, rates (operator and machine)
and labour and machine costs.................................................................. 186
Table 7-2: The total factory expenses, the expenses rate, the estimated per unit
values (factory expenses, product cost) (2003-2005)............................... 188
Table A-1 Hammer Drill (product level) 1.6 Kg........................................................ 268
Table A-2 Cumulative material quantities at the lowest level for the hammer
drill......................................................................................................... 273
15
List of Notations
The following is a list of the main symbols used for modelling in the thesis, together
with their brief descriptions.
a Total no. of broken or worn out labour tools in the ‘nth’ year
b Total no. of broken or worn out machine tools in the ‘nth’ year
d No. of time dependent cost elements
h No. of building space cost elements
i No. of machine centres
j No. of work centres
mk Amount of the kth material in product ‘p’
mn amount of nth direct material
q No. of overhead elements
r No. of work centres routed by the ‘pth’ product
t ipn Work time consumed by the ‘pth’ product at the ‘ith’ machine centre in the ‘nth’ year
t njp Work time consumed by the ‘pth’ product at the ‘jth’ work centre in the ‘nth’ year.
tx time spent by a skilled labour in the ‘jth’ work centre working on product ‘p’
ty time spent by a semi-skilled labour in the ‘jth’ work centre working on product ‘p’
tz time spent by a non-skilled labour in the ‘jth’ work centre working on product ‘p’
w No. of material dependent cost elements
16
List of Notations
x No. of skilled labour working on product ‘p’ in the ‘jth’ work centre
y No. of semi-skilled labour working on product ‘p’ in the ‘jth’ work centre
z No. of non-skilled labour working on product ‘p’ in the ‘jth’ work centre
C dn Total cost in the individual time dependent cost elements (such as utility cost,
maintenance cost, repair cost, depreciation, insurance, tax etc.) for the ‘nth’ year
n −1
Cdmt overall direct material costs in the (n-1)th year
Cdn unit cost of the nth direct material
C nft−1 overall freight & transportation costs in the (n-1)th year
C hn Total cost in the individual building space cost elements (such as plant depreciation,
building insurance, maintenance, repair, tax, utilities etc.) for the ‘nth’ year
Cin −1 overall inspection costs in the (n-1)th year
Cimn −1 overall indirect material costs in the (n-1)th year
C kn Unit cost of the kth material in product ‘p’ in the nth year
n
C md Total material dependent costs in the ‘nth’ year
n +1
C mdp Estimated material dependent cost for the ‘pth’ product in the (n+1)th year
n
C mp Material cost for ‘pth’ product in the nth year
n +1
C mp Material cost for ‘pth’ product in the (n+1)th year
n
C mt Cumulative material cost for ‘p’ products in the nth year.
n +1
C mt Cumulative material cost for ‘p’ products in the (n+1)th year.
C pn +1 Estimated cost for the ‘pth’ product in the (n+1)th year
C Pn −1 overall purchase department costs in the (n-1)th year
C sin −1 overall stores & inventory costs
C tdn Total time dependent costs in the ‘nth’ year
17
List of Notations
n +1
Ctdp Estimated processing cost for the ‘pth’ product in the (n+1)th year
n −1
Ctotal total capacity in the (n-1)th year
C wn Total cost in the individual material dependent cost elements (such a indirect
material, purchasing cost, stores & inventory cost, freight & transportation cost,
material inspection cost, packaging cost, quality cost, etc.) for the ‘nth’ year
n +1
C Bp Estimated building cost for the ‘pth’ product in the (n+1)th year
C Btn Total building space cost in the ‘nth’ year
n +1
C Ep Estimated engineering cost for the ‘pth’ product in the (n+1)th year
n +1
CGp Estimated manufacturing cost for the ‘pth’ product in the (n+1)th year
C Gtn Total manufacturing cost in the ‘nth’ year
Direct labour cost for the ‘pth’ product in the nth year
n
C Lp
n +1
C Lp Estimated direct labour cost for the ‘pth’ product in the (n+1)th year
n
C LT Total labour tool cost for the ‘nth’ year
C Ltn Total labour cost in the ‘nth’ year
C Ltn +1 Total estimated direct labour for the (n+1)th year
n
C MT Total machine tool cost for the ‘nth’ year
CTpn +1 Estimated tooling cost for the ‘pth’ product in the (n+1)th year
CTtn Total tooling cost in the ‘nth’ year
Da Total depreciation of the broken labour tool
Db Total depreciation of the broken machine tool
D Ln Total depreciation of the ‘Lth’ tool in the ‘nth’ year
DMn Total depreciation of the ‘Mth’ tool in the ‘nth’ year
Gtjn Total wages for ‘jth’ work centre in the ‘nth’ year,
18
List of Notations
H qn Total cost in the individual overhead elements (such as computer software cost,
general administration cost, financing expenses, selling expenses etc.) for the ‘nth’
year
L Total no. of useable labour tools in the ‘nth’ year
Lnjp Labour units consumed by the ‘pth’ product at the ‘jth’ work centre in the ‘nth’ year.
M Total no. of useable machine tools in the ‘nth’ year
M ipn Processing units consumed by the ‘pth’ product at the ‘ith’ machine centre in the ‘nth’
year
MLT manufacturing lead time
N pn No. of units of ‘pth’ product produced in the nth year.
N pn +1 No. of units of ‘pth’ product produced in the (n+1)th year.
O nj Occupied space by the ‘jth’ work centre in the ‘nth’ year
O pn +1 Estimated overheads for the ‘pth’ product in the (n+1)th year
Orpn +1 Space occupied by the ‘pth’ product at the ‘rth’ work centre in the (n+1)th year
Otn Total overheads for the ‘nth’ year
OM Material-related overhead for a new product in the nth year
OT time-related overhead for a new product in the nth year
Pa Initial purchase price of the broken labour tool
Pb Initial purchase price of the broken machine tool
R Bn Building space rate for the ‘nth’ year
RBn+1 Estimated building space rate for the (n+1)th year
n
R LA Actual labour rate for the ‘nth’ year;
n
RLT Labour tool rate for the ‘nth’ year
n +1
RLT Estimated labour tool rate for the (n+1)th year
19
List of Notations
n +1
RLE Estimated labour rate for the (n+1)th year
n
RMA Actual processing rate for the ‘nth’ year
n +1
RME Estimated processing rate for the (n+1)th year
n
RMT Machine tool rate for the ‘nth’ year
n +1
RMT Estimated machine tool rate for the (n+1)th year
Sn Total scrap value in the ‘nth’ year
S opn +1 Total space occupied by the ‘pth’ product in the (n+1)th year
S otn Total space occupied by all work centres in the ‘nth’ year
S utn Total unoccupied space on the manufacturing floor in the ‘nth’ year
n +1
S Gp Total manufacturing space for the ‘pth’ product in the (n+1)th year
S Gtn Total manufacturing space in the ‘nth’ year
n −1
TROtotal total time-related overhead in the (n-1)th year
Labour units for ‘pth’ product in the ‘nth’ year

n
U Lp
n +1
U Lp Labour units for the ‘pth’ product in the (n+1)th year
U Ltn Total labour units consumed in the ‘nth’ year.
U Ltn +1 Total labour units consumed in the (n+1)th year
Processing units consumed for ‘pth’ product in the nth year

n
U Mp
n +1
U Mp Processing units for the ‘pth’ product in the (n+1)th year
n
U Mt Total processing units consumed in the ‘nth’ year
α skill index for semi-skilled labour (0.40.8)
β skill index for non-skilled labour (0.250.4)
δ n+1 Building space cost deviation index in the (n+1)th year
20
List of Notations
ε n +1 Labour cost deviation index in the (n+1)th year
φ n+1 Material cost deviation index in the (n+1)th year
ψ n +1 Machine tool cost deviation index in the (n+1)th year
ηi Machine index (1.252.0)
µ n +1 Processing cost deviation index in the (n+1)th year
ρ n+1 MDC deviation index in the (n+1)th year
σ n +1 Labour tool cost deviation index in the (n+1)th year
τ n +1 PO deviation index in the (n+1)th year
21
List of Acronyms
The following is a list of acronyms used in the thesis, together with their brief
descriptions.
ABC Activity based costing
ASSD assembly & services sub-division
BCDI building cost deviation index
BOM bill of materials
BPNN back-propagation neural network
CBR case-based reasoning
CCS cost control system
CIM computer integrated manufacturing
CMD crane manufacturing division
CMR cumulative material requirements
CUI cost uncertainty index
DFC Design for cost
DSM decision support model
DSS decision support system
DT distribution transformers
HCS hierarchical classification system
HMI Hybrid Model Implementation
ICSD installation & commissioning sub-division
22
List of Acronyms
LCDI labour cost deviation index
JIT Just-in-time
LTCDI labour tool cost deviation index
MCDI material cost deviation index
MDC material-dependent costs
MDCDI material-dependent costs deviation index
MLT manufacturing lead time
MRO material related overheads
MRP material requirement planning
MTCDI machine tool cost deviation index
MTO make-to-order
PCE product cost estimation
PCDI Processing cost deviation index
PO production overheads
PODI production overheads deviation index
PT power transformers
QFD Quality Function Deployment
SED ship engineering division
SPSD spares & parts sub-division
TRO time related overheads
23
Acknowledgements
Alhamdulillah, with the thesis finally in my hand, a long and hard struggle seems to
have concluded with nothing less than a sweet reward. My efforts alone in treading this
difficult journey from its very start would take me no where, had it not always been
coupled with the explicit or implicit support of various people all along.
I would like to heartily acknowledge the never-ending support of Prof Jian Dai as not
only my PhD Supervisor but as a kind and compassionate mentor. His valuable
guidance in matters of both personal and professional grooming played a pivotal role. I
would like to extend my special thankfulness to Dr Stavroula Balabani whose continued
assistance in issues of academic and non-academic relevance has been instrumental to
the overall success of the project. I would also like to acknowledge the support of Prof
Lakmal Seneviratne as Head of the Division of Engineering, King’s College London.
The active support and dedication of all these academics made the publication of two
quality journal papers possible. The submission of another three papers in leading
journals was also made easier with their continued support.
It is difficult to find words to pay gratitude to Dr Salim Habib’s generosity through
Pakistan Scholarship without which my PhD simply would not have been achieved. His
unceasing personal interest in my academic progress remained a revitalizing element
24
Acknowledgements
throughout my PhD studies. King’s College’s support through Principal’s Discretionary
Fund during crucial stages of my PhD proved vital. Caroline Usher played an important
part not only in liaising with Dr Habib but in guiding me towards the discretionary fund.
I would like to remember the support of Mr Muhammad Arif Hasan and Mr Saeed Iqbal
in making the industrial visit to an Electrical Engineering Company in South Asia
possible. I would also like to name Mr Peter Faccenda from Manufacturing Advisory
Service UK for making the industrial visit in the UK possible by liaising with the Crane
and Ship Engineering Company. I would have liked to name the companies involved
that were central to the success of the project. I would acknowledge the support
provided by both the companies that agreed to disclose very sensitive data crucial to
their business competitiveness and equally vital to advance my research studies. Due to
the confidentiality accord and understanding the need for their anonymity, I would like
to thank all those namelessly who provided key information necessary for the research
progress and its eventual completion. I would also like to acknowledge the positive
feedback and interest from Professor Frank J Fabozzi of Yale School of Management
(USA) for my first journal publication.
This uphill task simply would not have been possible without the support and care from
a loving family like mine. My parents visited the UK twice from Pakistan during the
course of my PhD to reassure their heartfelt love at times when I really needed it most. I
would not be where I am now without the infinite love and guidance from my parents.
My wife Uzma stood by all my judgements with an unwavering belief in my abilities at
times when any uncertainty on her part would not be unreasonable. My daughters,
Daanya and Iqra remained the inner strength for me. Last but not least, the support from
25
Acknowledgements
my brother, Kamran and my sister, Farhea peaked when they visited me from Pakistan
during some of the very difficult times during my studies.
26
Chapter 1 Introduction
This chapter outlines the scope of the thesis by giving an overview of the research area
under study. Main problems are identified in the field and aims and objectives of the
study are outlined. The chapter also outlines the structure of the thesis and ends with
the conclusions.
1.1 Overview
Chapter 1 is an introductory chapter and is aimed at providing an overview of the
research field; identifying main problems and setting the relevant aims and objectives
for the research study.
The advent of information technology has brought with it an unprecedented era of
globalization. The ever-increasing pressure on firms to stay competitive in such an
environment is constantly forcing them to remain innovative in all aspects of their
business. No wonder, not only products and services are increasingly customized to suit
the end users’ needs but the required business tools are ever more novel. One of the key
business tools is a pricing strategy that combined with market awareness not only
allows an enterprise to remain competitive but thrive in the market.
27
Chapter 1: Introduction
A good pricing system rests on the application of cost engineering. The fundamentals of
the former and the principles of the latter are not mutually exclusive. Cost estimating,
cost control and profitability are the elements of cost engineering [1]. The American
Association of Cost Engineers (AACE) defines cost engineering as “that area of
engineering practice where engineering judgement and experience are utilized in the
application of scientific principles and techniques to the problems of cost estimation,
cost control, and profitability”.
Whereas market awareness can deal with the issues of profitability to some extent,
maximizing profits comes down to better cost control. An effective cost control system
(CCS) aims to reduce the gap between the budgeted and the actual costs. Setting the
realistic budgets in turn is one of the functions of cost estimation. The ultimate
responsibility of maximizing profits is, therefore, often closely linked with the
provisions of accurate and timely cost estimates in order to facilitate key managerial
decisions.
Cost estimating unlike cost accountancy requires sound engineering knowledge in order
to deal with problems involving scientific principles and techniques. Cost estimation
deals with predicting the cost of a product, project or a service. Product cost estimation
(PCE) methodologies aim to predict product costs early and accurately before the actual
production takes place and sometimes even before the design cycle. The scope of the
thesis is PCE with specific focus on manufacturing costs. Figure 1-1 highlights the area
of research followed in the thesis.
28
Area of research Project

Cost
Estimation
Applied Cost Engineering

Service
Cost
Cost Estimation Estimation
Cost Control Product

Cost
Estimation
Profitability
Figure 1-1: Main area of research
1.2 Problems identification
There are a number of issues in the area of PCE that need investigation. Following is the
brief outline of some of the problems that form the basis of the research study:
• A vast number of estimation techniques are available but no classification
system available
• Difficulty in selecting an estimation methodology for a given condition
• Early estimation and accuracy are counter to each other
• Available methodologies mostly estimate only part or component costs not
the entire product cost
• Most of the methodologies that can predict the entire product cost, fall
short of predicting accurately the costs breakdown
29
• One of the breakdown elements is overhead (sometime referred to as
indirect costs) that is difficult to predict accurately for individual products
A vast number of methods and techniques have been developed over a period of time to
facilitate estimators and designers to predict a product’s cost. However, they differ in
applications in terms of compatibility to the needs of a system, delivering optimized
results in given conditions and the level of resource consumption. A great number of
methods, on the other hand, share common grounds. Estimators often find it difficult to
select a methodology to suit the needs of a system under consideration. This is normally
because an in depth analysis of a specific methodology to check its compatibility with
an organizational framework is often time-consuming let alone analyzing more than one
techniques. The selection of a particular methodology is normally based on the
availability of data, level of accuracy and the stage of estimates required. A better
exploitation of the differences and similarities of the techniques could result in helping
estimators to select a specific methodology to suit the needs of a system. The non-
availability of a classification system is a barrier to the notion. Such a system could also
help in developing a decision support tool in order to help estimators of product costs to
make decisions to select an estimation methodology to satisfy a system’s needs and
make the best use of available resources.
The aim of estimating a product’s cost is not just predicting it accurately but as early as
possible in order to facilitate key business policy decisions. Often a methodology
selected for PCE in the early stages of a design cycle, does not furnish accurate enough
results due to the non-availability of the design details. Historical data can overcome the
problems to some extent but the requirement of extensive past results limits the use of
30
relevant techniques. Such techniques also fall short of predicting accurate results for
new designs. Study of a product’s features could be helpful in such circumstances but
not only require skilled estimators but mostly fall short of delivering early results due to
the detailed designs required. The reliance of the existing methods on product design
and process planning details could result in accurate estimation but often leads to
unnecessary delays. On the other hand, the uses of past experience, data or knowledge
do have the potential to deliver early estimates but compromise accuracy. Such
discussion is elaborated with necessary references in detail in chapter 3. Due to the
significance of an early and accurate estimation process, it is important that a correct
technique is employed within a specific set of conditions. However, most of the
available techniques make a compromise between early estimation and accurate results.
Most of the available techniques predict part or component costs instead of an entire
product’s cost using certain design parameters (such as design features, dimensions,
tolerances etc.) and manufacturing operation times. The individual components’ costs
add up to an entire product’s cost along with considering any process planning details
for assembly operations based on standard techniques, historical data or time and
motion studies etc. The process is time and resource consuming. A methodology aiming
to predict an entire product’s cost by not relying on details like product designs or
process planning details could be the answer. However, an alternative to such details
would have to be found.
A more traditional approach can predict the entire product cost based on allocating the
factory-wide resources to individual products. Such a method divides the entire product
costs into direct and indirect costs. Direct costs refer to direct material and labour
31
consumed for the benefit of the product whereas, indirect costs, also referred to as
overheads, are incurred for the shared benefit of all the products. The method obtains
material and labour costs either from past data for the existing products or uses design
and process planning details to calculate for the new products. However, using such
details for a new product design could not allow the method to predict costs early. The
method allows the allocation of the third cost element, overheads, to individual products
by considering their lead times and an aggregate overhead rate. However, again not only
the similar problem arises for any new products but the use of the aggregate rate results
in less accurate estimates.
Another problem with using the traditional method is its breakdown of the entire
product cost into only three elements (i.e. material, labour, overheads) resulting in
overlooking possible areas of optimization for cost control. In a production environment
where a product is the result of a mix of activities, the degree of consumption of these
activities is reflected in a product’s cost. However, in such an environment, finding the
costs incurred on individual breakdown elements and sub-elements is difficult,
especially for overheads. The early and accurate estimation of this element and its sub-
elements is crucial to the overall accuracy of a product’s cost. Predicting such costs
beforehand requires much more than just accountancy laws. The early and accurate
estimation of the individual elements and sub-elements costs could provide effective
cost control opportunities.
32
1.3 Aims and objectives
The set of identified problems in the area of PCE is helpful in establishing an outlook
within which the aims and objectives for the overall research work can be set. The
individual problems identified would help to establish the objectives for the research
work in order to find or attempt to find pragmatic solutions. The overall aim of the
work, however, is to make a genuine and significant contribution to the area of cost
engineering in general and to PCE in particular.
Following is the list of the objectives set for the research study:
1. Development of a technique classification system
2. Development of a decision support model (DSM) to allow estimators and
designers to select an estimation methodology in given conditions
3. Development of an overhead estimation methodology
4. Development of a PCE methodology with cost deviation indices for early
and accurate estimation of a product’s cost with its elements’ costs
5. Industrial implementation of the developed cost estimation technique
6. Comparison analysis for the validation of the developed model
A comprehensive review of the available cost estimation techniques could help to
identify the similarities and the differences between them. A careful analysis of their
advantages and limitations could be useful in grouping them. This in turn would allow
the development of a technique classification system. Such a system would also lead the
way to establish conditions under which designers and estimators make decisions to
select a specific methodology. A DSM can, therefore, be developed too.
33
The development of a PCE methodology for early and accurate estimation of a
product’s cost would require a thorough analysis of the existing methods. One of the
problems of the existing methods is their reliance on product designs and process
planning details. The developed method should aim at eliminating or reducing the need
of such details. This can be achieved by making use of the past production details in
order to predict the estimates for the future production. However, if a precedent is not
available, the use of such technique can be limited. Such a limitation can be overcome
by making use of past design cases resulting in a case-based framework. One of the
objectives of the study would, therefore, be to develop a case-based framework in order
to facilitate effective utilization of past details for early and accurate estimation of
product costs.
The determination of an entire product’s cost instead of only part and component costs
would require evaluating any existing methodologies that can predict such costs. The
areas of improvement in those methods will be identified. One of the areas, for example,
is overhead estimation that leads to inaccurate estimation of an entire product’s cost.
Another objective of the study, therefore, would be to identify problems with the
existing method of overhead estimation and develop an improved methodology. The
effect of the proposed methodology on the overall estimation of a product’s cost will
then be studied. An indication of improved overhead estimation results would pave a
way for modelling an entire product’s cost with its breakdown elements.
Finding pragmatic solutions would require industrial implementation, comparisons and
validation to back up any theoretical advancement in the field. One of the objectives of
the research would, therefore, be to validate the proposed mathematical models from
34
industrial applications. The aim of the study is, therefore, to make theoretical
advancement in the area of PCE and present industrial validation analysis.
1.4 Thesis structure
The overall research work is detailed in the thesis with the aim of disseminating not just
the research findings but knowledge in the area of PCE. A logical sequence is
maintained in presenting the work throughout the thesis starting from general concepts
and literature review to methodology development, implementation, application and
validation. The work comprises eight chapters in all.
Chapter two uncovers the background and related work in the area of cost estimation.
The concepts of costs and costing are outlined. Price and pricing are briefly described.
Cost estimation is defined and discussed with a special focus on manufacturing and cost
control. Manufacturing systems are discussed with a view of selecting a suitable system
to develop a methodology for. The functions of cost estimation are outlined and its
significance in the early stages of a design cycle is highlighted. A comprehensive
literature review in the area of cost estimation with special emphasis on its specific
applications is also presented in this chapter. The chapter identifies a batch
manufacturing environment as a potential area to develop a cost estimation
methodology for.
Chapter three establishes comprehensive theories in the area of PCE. Existing technique
classification systems are briefly discussed before developing a comprehensive
35
classification system for the available techniques in the area. The literature for each
category is comprehensively reviewed along the course with the identification of
strengths and weaknesses for the individual methods. The underlying principles for the
categories are mentioned and the conditions for selecting a method from each one of
them are set out. Establishing such conditions helped to develop a DSM for
methodology selection. The chapter also presents a framework for the case-based
methodology. A useful summary of the reviewed work and the key advantages and
limitations for the proposed categories is then presented in the end. The chapter
identifies that a hybrid approach combining the elements of qualitative and quantitative
techniques could result in early and accurate estimation of a product’s cost.
Chapter four forms part of the methodology development for PCE. It considers a
representative case of the existing method for overhead estimation. Problem
identification leads to the development of a new overhead estimation methodology
based on material– and time– dependent overheads. The effects of the proposed
methodology on the estimation results for the overall product costs are analysed
retrospectively for a four year period. The industrial validation analysis results reveal
the superiority of the proposed methodology. However, the room for further
improvement in the proposed methodology are also identified. The chapter also reveals
cost elements breakdown statistics typical of the South Asian region and help to
understand the implications of geographical locations on the effectiveness of a CCS.
Chapter five is a step forward in the development of a comprehensive methodology for
PCE in a batch type manufacturing environment. The concept of overhead estimation
proposed in the preceding chapter is carried forward to develop comprehensive
36
mathematical models for estimating an entire product’s cost in a batch type
manufacturing environment. The developed model is hybrid in nature based on a cost
breakdown structure and the concepts of modified activity-based costing (ABC) system.
The proposed methodology is based on an effective utilization of past data to predict
future costs facilitated by incorporation of cost deviation indices.
Chapter six is based on the industrial implementation and application of the proposed
Hybrid Model in a batch type manufacturing environment in the UK. The chapter is a
step towards the overall validation of the developed Hybrid Model. In order to facilitate
the industrial implementation of the proposed model, an implementation algorithm is
developed and the already proposed indices from the preceding chapter are modelled.
The developed model is then implemented retrospectively in a crane and ship
engineering company based in the UK. The implementation process generates the cost
estimates for a given product range. The company’s own method of cost estimation is
also described and is used to generate cost results for further comparison analysis.
Chapter seven presents the comparisons and validation analysis of the Hybrid Model.
The developed model is compared against a published model and the company’s own
method as the representative cases from the two domains. A carefully selected
published model is used to furnish cost estimates. Comparison analysis includes
comparing the estimated costs obtained from the three methods against the actual costs.
The developed model is validated based on generating more accurate and more
consistent results as opposed to both the company’s own method and the published
model. The chapter also reveals cost elements breakdown statistics typical for the UK.
37
Chapter eight concludes the research work and the thesis. The overall contents in
general and the research work in particular are summarized. The main achievements of
the work are highlighted and the success of the research study is evaluated. This
involves a comparison of the achievements against the planned aims and objectives. The
work is also evaluated for the originality and its contribution. The areas for possible
improvements in the research work are pointed out with a focus on the existing
limitations. The chapter also discusses the current and future research trends in the field.
Finally, the possible avenues for future research directions in the established research
work are highlighted with a view of providing a platform for any further research in the
area.
1.5 Conclusions
This chapter presented the scope of the research work. An overview of the research area
focussed on the importance of cost estimation for cost control. Some of the problems
were outlined and discussed briefly in order to establish the likely aims and objectives
for the study. Aims and objectives were set with the view of furnishing solutions to the
outlined problems and establishing a significant theoretical development in the area.
Finally, the overall structure of the thesis was outlined with the aim of translating the
objectives in logical order in different chapters.
38
Chapter 2 Background and
Related Work
The aim of Chapter 2 is to develop background information in the area of cost
estimation. The concepts of cost and costing are established. Manufacturing systems
with a focus on CCS are discussed. Cost estimation as a key component of CCS is then
discussed. Various application areas for cost estimation techniques are discussed with a
comprehensive literature review in the area. The rationale for developing a cost
estimation methodology in a batch type manufacturing environment is established. The
chapter ends with conclusions.
2.1 Introduction
Chapter 2 deals with background information in the chosen area of study and provides a
comprehensive literature review in the area with a focus on cost estimation techniques
applications.
Our day to day lives are full of instances of buying and selling goods and services. The
amount paid to buy a commodity reflects the incurred cost from a buyer’s perspective.
The same amount is perceived a selling price from the seller’s view point. The
commodity may be resold with a higher amount making a profit. Cost can therefore, be
defined as:
39
Chapter 2: Background and Related Work
“The amount of money paid or required as a payment to buy a goods or service”
The original commodity may often be altered or modified before being resold and is
said to have been added with a value. The value addition process thus incurs further
costs. The incurred costs could be the result of tangible and/or intangible values. The
process of determination of the final cost of the modified commodity before being
resold is referred to as costing.
Costing determines selling price. Setting prices in turn is called ‘pricing’. As costing
normally refers to considering all the costs already incurred for a commodity, any
unexpected costs at the beginning of the value addition process are discovered and
accounted for. For the same reason, however, its role is normally limited to pricing and
price adjustments. Cost control may, therefore, not be its domain.
Since selling price is one of the major factors of gaining competitive advantage in the
consumer market, cost control can not be ignored. Cost accounting also known as
management accounting could provide solutions to cost control by recording cost values
and providing opportunities to minimize them in future.
Operational matters requiring aspects of cost control within the framework of a project
execution, service delivery or manufacturing set up may be different. The aspects
relating to manufacturing environment come under the scope of the current study.
40
2.2 Manufacturing and cost control
Manufacturing involves converting materials from one form to another by adding value
to them. This often requires a series of specific treatments or processes to be carried out
on different materials. These processes accomplish the desired final product form and
shape specified by a design engineer prior to actual manufacturing. Selection of a
particular process among various alternatives is closely linked with many factors
including the associated costs. Incorporating certain features (from available alternatives
into a product design) that require costly processes is often a result of poor product
design. Therefore, the final product cost reflects the design quality and also determines
selling price. The latter is also governed by consumer demand and is crucial in gaining
competitive advantage. Maximizing profits (while remaining competitive in the market)
requires an effective CCS alongside keeping high quality standards. An effective CCS
monitors various aspects of the final product cost.
P
Increasing
Productivity (P)
product variety
Increasing
production volume
Flexibility (F)
Figure 2-1: Relationship between productivity and flexibility
41
In a manufacturing environment, where high volumes of products are produced without
making many changeovers in production setup, the aspects of cost control are a routine
task. The effectiveness of a CCS, however, is more strongly required within a
manufacturing setup of routine changeovers in production setup. The ability to
changeover to a new production setup economically and quickly to produce new
products in response to market or engineering changes is called Product Flexibility. For
example, the automobile industries are subject to tough competitions and in order to
remain competitive, develop different car models and introduce them at shorter
intervals. On the other hand, Productivity is a measure of the extent to which the
resources of an organization are consumed effectively in transforming inputs to outputs.
The overall productivity of a firm is greatly influenced by flexibility. The productivity
of a manufacturing system remains in conflict with flexibility where increasing
productivity compromises the system’s flexibility. The relationship is elaborated in
Figure 2-1 between flexibility and productivity. An increased flexibility of a
manufacturing system demands higher changeover time thereby reducing its
productivity. Increased productivity also refers to higher production volumes whereas;
an increased flexibility is a reflection of an increased product variety.
Business goals for a manufacturing enterprise are set within the framework of the
organizational corporate planning structure. Strategic planning framework in return
ensures that a manufacturing system be in place in order to achieve those objectives. For
example, the level of flexibility and productivity required to achieve business objectives
would need a specific set up for a manufacturing system. High production volumes and
low flexibility are the attributes for mass production and can be achieved by setting up
transfer lines. Such systems make use of automated assembly lines using industrial
42
robots etc. The investment needed for such a manufacturing set up is high. However,
both labour skills and labour costs are relatively low. Small production quantities with a
high flexibility, on the other hand, can be achieved by conventional job shops.
However, the level of skill required to operate general-purpose machines in such a set
up is high. Since the production rate is low, the unit cost in job shops is usually high.
Conventional flow line, flexible manufacturing system, manufacturing cells, numeric
control systems etc. can be effective in achieving a desired combination of flexibility
and productivity and normally result in batch production. As a result multi-skilled
labour (low, medium, high) is generally employed to handle tasks for varied levels of
flexibility and productivity. Small to medium size batches can be produced on machines
ranging from general-purpose with computer controls to equipment with specifically
designed fixtures and tools. Figure 2-2 elaborates various levels of productivity and
flexibility achievable through different manufacturing systems. The overlap between the
systems is due to the various levels of technical attributes (automation, computer control
etc.) achievable in each system. Due to the impact of the changeovers on product costs,
the need for an effective CCS in a batch manufacturing set up is wider in its scope. Job
shop, although, highly flexible can be partly covered by batch production cost control
strategies. Both mass production and a job shop system are, therefore, left with little
scope for comprehensive CCS.
43
Transfer
Line
Increasing Productivity
Mass Production
Batch Flow Line
Flexible Manufacturing
System
Batch Production
Manufacturing
NC Cell
Production
Job Shop
Job Shop Production
System
Increasing Flexibility
Figure 2-2: Flexibility and productivity of different manufacturing systems
Whether a manufacturing system is based on mass, batch or job shop production, an
effective CCS makes sure at every stage of the execution process that the planned
budget is adhered to. Accountancy laws are easier to apply once the plans are executed
and actual costs are available. The aim is to minimize the difference between the
planned budget and the actual costs. In a highly competitive business environment, any
non-compliance to planned budgets could seriously jeopardize the accomplishment of
an enterprise’s business objectives. Setting realistic budgets in turn is, therefore, also an
important element of a CCS and requires accurate predictive tools. A carefully devised
cost estimation methodology based on sound engineering principles serves as a reliable
tool for predicting accurate costs likely to be incurred during an execution phase.
44
2.3 Cost estimation
Cost estimation is the process of determining a combination of tangible & intangible
values associated with the expected and unexpected cost occurrences linked with a set
of activities prior to their execution. Accurate cost estimation is essential whether the
end product is a project, manufactured goods, or a service. The cost estimated close to
the accounting cost should serve as a valuable decision aid tool at the early stages of an
execution plan.
Different cost estimation approaches have been developed to best suit the needs of end
users or customers. One approach may require the evaluation of uncertainty as an
intangible value associated with the loss of goodwill owing to delayed job completion,
whereas another may necessitate the estimation of expected cost tied with the
development or manufacture of a product. The use of risk assessment and uncertainty
evaluation techniques is quite common in project proposals and bidding processes. PCE
techniques, on the other hand, generally tend to consider the effect of expected cost
occurrences.
Cost estimation affects certain organizational functions. The following is a list of
functions or processes in a manufacturing company where cost estimation is involved.
The list is not exhaustive but provides a good idea of the impact of cost estimation in
the operational and corporate planning structure of an organization. The functions
include:
• deciding whether to produce in-house, purchase, or out-source;
45
• selecting material, manufacturing processes, and routings;
• evaluating different design alternatives;
• setting economic lot sizing for batch production;
• assessing suppliers’ quotations as a trade-off between customer needs
(quality) and available budget (cost);
• quoting prices to customers before actual manufacturing starts;
• allocating enterprise-wide resources (man, machinery, tooling, etc.) and
budgeting to different technical attributes of a design in proportion to their
desirability (extent of customer requirements);
• manufacturing cost control and hence overall control of production cost.
Good cost estimation has a direct bearing on the performance and effectiveness of a
business enterprise as overestimation can result in loss of business and goodwill in the
market whereas underestimation may lead towards financial losses to the enterprise.
Due to this sensitive and crucial role in an organization, cost estimation has been a focal
point for design and operational strategies and a key agenda for managerial policies and
business decisions. As a result, a substantial research effort has been expanded in
exploring design implications, new techniques and methods for producing accurate and
consistent cost estimates not only to generate optimum design solutions but also to
achieve the maximum customer satisfaction in terms of low cost, high quality and in-
time product delivery.
46
2.4 Cost estimation in the early design stages
The key to thrive for a manufacturing enterprise in the 21st century is based on product
quality, competitive cost, fast delivery, and flexibility. On the other hand, factors like
globalisation and mass customisation put an extra pressure to a business enterprise to
survive and remain profitable at the same time. Whereas, an innovative approach and a
new product development process may attempt to deal with issues such as flexibility
and product quality, they may still be time consuming and less cost effective. In
addition, the prospective end user of a would-be product often demands a price quote as
soon as possible, sometimes even unconcerned and oblivious of factors such as the
extent of the customisation, the nature of the data required and the design complexity.
To make matters worse, often a manufacturer ignores the significant factors like design
module availability, manufacturability and the level of accuracy required for processing
time estimation. The overall situation, therefore, could either lead to an underestimation
resulting in a profit loss and a blow to operational targets or a more profound strategic
damage caused by overestimation leading towards the loss of customer goodwill and
market share. All the above highlights the ever-increasing importance of devising
methods to forecast the cost for a new product in the early design and development
phases with accuracy.
Since most of the product costs sustained during later in the production life cycle are
determined during the conceptual design phase [2], the cost estimation in the early
phase of the design cycle is crucial. Many researchers have emphasized the importance
of cost estimation at the early design stages when 70 to 80 percent of a total product
cost is determined [3-6]. Some researchers have developed methodologies with a special
47
emphasis on early cost estimation [7, 8]. A framework for developing a cost database
was suggested by Sheldon et al. [9] and aimed to serve different groups of design for
cost (DFC) system users to determine appropriate cost structures by analysing the
information provided by a cost accounting system. Knowledge representation in such a
way facilitated the generation of cost estimates at an early design stage. A framework to
integrate design costs into Quality Function Deployment (QFD) was used by Bode and
Fung [10]. The approach adopted is a helpful tool for designers at the early stages of
product design for making trade-off decisions between quality and cost prioritising the
attainment of technical attributes based on customer requirements. All the above
mentioned highlights the significance of not only estimating a product cost accurately
but as early as possible.
2.5 Cost estimation for specific applications
A number of cost estimation methods and techniques were developed with reference to
particular applications. The techniques only suit the conditions for which they were
tailored and can only be effective for their application areas. These techniques range
from the evaluation of certain manufacturing and machining processes to the dedicated
techniques designed to suit specific manufacturing systems, from composite material
costing to the cost analysis of parts and assemblies and from dealing with specific
segments in a production cost cycle to covering a product life cycle cost. The section is
aimed at providing a comprehensive review of the available techniques for cost
estimation with an emphasis on their application areas. This not only leads to develop
better understanding of the application areas but could also be helpful in providing a
48
platform for any future exploration within a specific application area of the cost
estimation.
2.5.1 Cost estimation for a specific segment in a production cycle
Different costs are associated with various stages in a production cycle starting from the
ones incurred during the early phases of a design cycle to those linked with the
manufacture of an actual product on the shop floor. Many researchers devised methods
to evaluate the costs associated with a specific segment in a production cycle. For
example, if a methodology is applicable at the QFD stage [10] the other is developed for
the costs associated with the design and development phase of a product [11]. Methods
for process planning cost evaluation and optimization can also be found in [12-14].
Aldrich [15] estimated the cost associated with the bill of materials (BOM) using
MRPII software. Cost calculation in manufacturing and machining can be found in [16].
Costs associated with conventional manufacturing processes [6], non-conventional
manufacturing processes [17] and the machining accuracy [18] can also be found.
Some researchers developed methodologies that could be used for cost estimating in
several stages of a product design cycle. For example, Weustink et al. [19] developed a
framework for product cost control by estimating the various cost elements and storing
the data in a generic way. The methodology allows cost estimation on different
aggregation levels, e.g. feature level, component level, assembly level etc. Koonce et al.
[20] developed a system capable of generating cost estimates in all phases of the design
stage. The system, which was prototyped in JAVA, used simple parametric cost
relations in the early stages of a design, when detailed design was not available to
49
produce cost estimates for a product. When the design was developed, estimates could
be produced using design features, manufacturing features or a process plan.
2.5.2 Cost estimation for specific machining and manufacturing
processes
The selection of suitable manufacturing processes is governed by an accurate estimation
of the costs associated with them among other factors. Methods have been devised to
predict the costs of specific machining and manufacturing processes. These include the
assembly costing techniques [21, 22] and the cost models for die-casting [23]. Further
models were developed for a hole-making process [24], welding [25] and milling and
drilling [26].
2.5.3 Cost forecasting for specific parts and products
Many researchers focussed on the application of the cost estimation techniques to
specific products ranging from standard parts and components to a particular product
group. Schreve et al. [27] presented a cost model using mild steel fabricated parts,
whereas the one for machined components can be found in [28]. Hicks et al. [8]
developed four cost modelling approaches for various classes of engineering
components. These components were defined as standard selected, standard designed
and bespoke designed. French [29] used a function cost modelling methodology to
estimate the costs of mechanical components whereas Ulrich and Eppinger [30] used
previous orders and procurement records to estimate the cost of similar components.
Kendall et al. [31] presented cost information for automotive components using a
50
simulation technique. Gutowski et al. [32] presented a process-oriented cost model to
estimate manufacturing cost of advanced composite aerospace parts. Cost models for
injection-moulded components can also be found in [33-35]. On the other hand, cost
models for specific product groups include those for PCB manufacturing [36-38],
developmental equipment [39], packaging products [40, 41], injection moulding tool
[42] and gear drive manufacturing [6].
Another type of products covered by cost estimation techniques is based on composite
material. For example, cost and consolidation model for commingled yarn based
composites was presented in [43]. Process flow simulation techniques to evaluate
manufacturing costs for composite products were discussed in [44]. Cost analysis of
thermoplastic composites using different techniques can also be found in [45, 46],
whereas Walls and Crawford [47] used historical data to produce cost information for
continuous fibre-reinforced thermoplastic products.
2.5.4 Cost estimation for generic systems
Cost estimation models to suit the needs of a generic system were also developed by
many researchers. For example, cost models for job shop manufacturing systems can be
found in [48, 49]. On the other hand, a mathematical and simulation model to estimate
the manufacturing and product cost in material requirement planning (MRP) and just-in-
time (JIT) systems was proposed in [50]. Similarly, a simulation model to estimate the
cost in a flexible manufacturing environment was proposed in [51], whereas the one for
cellular manufacturing configuration was proposed in [52].
51
Cost estimation techniques were further developed for specific industry sectors [53].
For instance, work has been carried out for airplane manufacturing industries [54],
electronics industries [55], automotive industries [56], aerospace and defence industries
[57], telecommunications [58], and ship building industries [59].
Table 2-1 summarizes the discussed literature references for various applications of
PCE. It is clear that the techniques have been designed to suit the conditions for a given
application area and may not be suitable for any other applications. For example, a
methodology designed to predict machining costs may not be applicable for other areas
of manufacturing due to the differences in the requirements of the parameters for the
two conditions. There is, thus, a need to develop a method that can cover a number of
application areas without compromising accuracy yet furnishing estimates in the early
stages of a design cycle or even before. Such a system could eliminate or minimize the
need for a methodology selection under varying conditions. The techniques designed to
work for a specific generic system usually do not have the limitations for other
application areas. For example, a methodology developed for a job shop system,
although, may not be suitable to work in a batch manufacturing environment but can
provide estimates for manufacturing parts and be applicable in the different phases of a
design cycle. Maximizing the application areas for a methodology could, therefore, be
achieved by originating it from the applications for generic systems. In other words, if
an estimation methodology is designed to suit the needs of a generic system, it could
work well for the other application areas.
52
Table 2-1: Summary of the published literature references for various applications of PCE
S. No. Application area References
Cost estimation for a specific segment in a production

1. [6], [10-20]
cycle
Cost estimation for specific machining and

2. [21-26], [75-79]
manufacturing processes
Parts and components [8], [27-35]

Cost estimation for
3. specific parts and Special products [6], [36-42]
products
Composite material products [43-47]
Specific manufacturing system [48-52]
Cost estimation for
4.
generic systems Specific industrial sector [53-59]
It was already established in Section 2.2 that the scope of a batch manufacturing system
is wider in terms of not just encompassing a range of manufacturing set ups but for a
realistic need of a CCS. The effectiveness of such a CCS was then linked with the
effectiveness of a cost estimation methodology. Now, that the need to develop a cost
estimation methodology for a generic system is evident from the view point of
maximizing the application areas already mentioned, the only manufacturing system left
for exploring the possibilities for a methodology development is a batch type
environment.
2.6 Conclusions
This chapter presented a comprehensive survey of literature in the area of cost
estimation in order to establish a background in the research area and to develop an
53
understanding of the problems. The techniques used for cost estimation in various
application areas were investigated.
The chapter started with defining basic concepts of cost and costing. Manufacturing
systems were discussed with cost control implications. Flexibility and productivity were
discussed and their relationship was detailed. It was noted that various combinations of
production volumes and product varieties resulted in three main categories of
manufacturing environment: mass, batch and job shop production. It was found that
batch production environment demanded an effective CCS and was suggested as a
desirable area for developing an estimation technique. Cost estimation, its functions and
the significance of early estimation were discussed.
A comprehensive literature review with a focus on four main application areas was
presented. Techniques from each application area were thoroughly analysed. They
included techniques for specific segment in a production cycle, techniques for specific
machining and manufacturing processes, techniques for specific parts and products and
finally the techniques for generic systems. It was noted that techniques developed for
generic systems did have the potential to fulfil the demands of the other application
areas. In order to maximize the scope of an estimation methodology, it was deemed
necessary to develop a technique for a generic system. Batch manufacturing
environment was considered as a viable generic system for which a cost estimation
technique could be developed. Finally, the published literature references mentioned
for various applications of cost estimation were summarized for any useful research in
future.
54
Chapter 3 PCE Technique
Classification System
Chapter 3 provides a technique classification system based on the state of the art on
product cost estimation. PCE techniques are classified to form an extensive hierarchical
classification system. The developed classification system is further used to form a decision
support system for methodology selection. A case-based model is also developed for
effective utilization of past product details. Key advantages and limitations of the
techniques are discussed along the course. The chapter concludes with the suggestion of
developing a methodology for cost estimation based on combining concepts from both
qualitative and quantitative techniques for early and accurate estimation of a product’s
cost.
3.1 Introduction
Product cost estimation refers to predicting all the costs associated with manufacturing a
product from raw material to a finished product. Accurate product cost estimates are
important in establishing a good CCS and help in setting competitive price plans. The role
of PCE is, therefore, to predict the overall product cost likely to be incurred throughout the
55
Chapter 3: PCE Technique Classification System
entire production phase. The accuracy of the predicted costs will determine the
effectiveness of the technique employed and, combined with the information access, can
help the entire production enterprise achieve its corporate goals. It, therefore, not only
serves operational matters (such as cost control, resource allocation, price setting etc.) but is
also deemed as a key component for strategic production planning and control system.
The published literature on PCE covers a wide variety of issues ranging from
manufacturing cost estimation of standard mechanical components to cost analysis of
highly customized assembled products, from process cost optimisation techniques to
specific methods for overhead costing, from unique approaches for estimation at the
conceptual design stage to general costing rules designed for use at a later stage in the
design cycle and also from classical costing methods to highly novel cost estimation
techniques. Several textbooks [60-62] can be found on some of the subjects.
Due to the significance of the cost estimation process in the organizational structure of an
enterprise, the influence of an effective estimation methodology can not be underestimated.
On the contrary, the availability of an extensive range of estimation techniques makes the
selection of an appropriate methodology to suit the prevailing conditions a daunting task.
The classification of these techniques based on certain criteria is a step towards overcoming
the issue. A number of researchers have attempted to categorize the PCE techniques using
certain criteria. Zhang et al. [40] categorized some techniques into traditional detailed-
breakdown, simplified breakdown, group-technology based, regression-based and activity-
based cost estimation techniques. Ben-Arieh and Qian [11] classified cost estimation
56
methods into intuitive, analogical, parametric and analytical methods. Shehab and Abdalla
[63] mentioned intuitive, parametric, variant-based and generative cost estimating
approaches without defining them clearly. The same authors [64] later classified cost
modelling approaches at the design stage into knowledge based, feature based, function
based, and operations based approaches. Cavalieri et al. [2] identified three approaches for
cost estimation: analogy-based, parametric and engineering approaches. However, a
comprehensive hierarchical classification of the estimation techniques has not been
exploited.
3.2 Development of hierarchical classification system (HCS)
The availability of a wide range of PCE techniques demands an extensive classification
system in order to facilitate the selection of a suitable methodology for a given condition.
The existing classifications fall short of furnishing a complete framework for methodology
selection. These classifications group together the techniques with similarities in separate
categories but do not consider the differences within the respective groups. Filtering a
methodology is, therefore, constrained. The development of HCS is aimed at providing a
framework for not only classifying the PCE methodologies but facilitating their selection to
suit different problems. The developed system comprising an extensive hierarchical
classification is based on grouping the techniques with similar features into various
categories. The methodologies discussed in different categories are distinct and reflect the
nature of that category. An effort is also made to elaborate each group or category with
57
reference to published work. Meanwhile, mathematical models are presented on occasions
with particular references in order to better understand the nature of a given category. The
hierarchical classification expands due to the identification of the variations within a group.
The PCE techniques are broadly classified into qualitative and quantitative techniques
followed by an extensive sub-classification.
Qualitative cost estimation techniques are primarily based upon a comparison analysis of a
new product with the products that have been manufactured previously in order to identify
the similarities in the new one. The identified similarities help to incorporate the past data
into the new product so that the needs to obtain the cost estimate from scratch are greatly
reduced. In that sense the past design and manufacturing data or previous experience of an
estimator can provide a useful help to generate reliable cost estimates for a new product
which is similar to a past design case. Sometimes, this can be achieved by making use of
the past design and manufacturing knowledge encapsulated in a system based on rules or
decision trees etc. Historical design and manufacturing data for products with known costs
may also be used systematically to obtain cost estimates for new products. For example,
regression analysis models and neural network approaches could provide an efficient way
to predict costs for new products by using historical cost data. In general, qualitative
techniques help obtain rough estimates during the design conceptualisation. These
techniques can further be categorised into Intuitive and Analogical Techniques, which are
discussed in detail in Sections 3.3 and 3.4 respectively.
58
Quantitative techniques, on the other hand, are based on a detailed analysis of a product
design, its features and corresponding manufacturing processes instead of simply relying on
the past data or knowledge of an estimator. Costs are, therefore, either calculated using an
analytical function of certain variables representing different product parameters or as the
sum of elementary units representing different resources consumed during a whole
production cycle of a given product. Although, these techniques are known to provide more
accurate results, their use is normally restricted to the final phases in the design cycle due to
the requirement of a detailed product design. Quantitative techniques can be further
categorised into Parametric and Analytical Techniques, which are discussed in detail in
Sections 3.5 and 3.6 respectively.
These techniques categorised as qualitative and quantitative can be illustrated in a tree
diagram in Figure 3-1 showing the preliminary classification of PCE techniques.
Product Cost Estimation Techniques
Qualitative Techniques Quantitative Techniques
Intuitive Techniques Parametric Techniques
Analogical Techniques Analytical Techniques
Figure 3-1: PCE Preliminary Technique Classification
59
3.3 Intuitive cost estimation techniques
The intuitive cost estimation techniques are based on using the past experience. A domain
expert’s knowledge is systematically used to generate cost estimates for parts and
assemblies. The knowledge may be stored in the form of rules, decision trees and
judgments etc. at a specific location, e.g. a database, to help the end user to improve the
decision making process and prepare cost estimates for new products based on certain input
information. The present study identified three sub-categories under intuitive techniques.
3.3.1 Case-based methodology
This approach also known as case-based reasoning (CBR) attempts to make use of the
information contained in previous design cases by adapting a past design from a database
that closely matches the attributes of a new design. This often requires making necessary
changes to parts and assemblies of previous design cases and incorporating missing details
to it. Figure 3-2 shows a complete framework for the case-based approach with the dotted
lines representing the cost interfaces to the system. The process starts by outlining a new
product’s design specifications followed by retrieving a closest design match from a design
database. The system then attempts to find the changed assemblies and subsequently
changed parts in the assemblies. Changes are incorporated in the design either by retrieving
similar parts and/or similar assemblies from the design database or by designing the new
ones altogether. All the necessary changes are incorporated in a similar way until the new
design conforms to the outlined design specifications.
60
Figure 3-2: Flow Diagram of the Case-Based Approach for Cost Estimation
61
The new design is later stored in the design database. This technique allows the cost
estimation for a new product by combining the past results with those for the newly
designed components and assemblies, thereby greatly reducing the need to design from
scratch. The approach is, therefore, helpful in making good estimates at the conceptual
design stage, since the use of the past cost data to generate new estimates greatly minimizes
the estimation time. However, the methodology is applicable only when similar past
designs are available to incorporate the relevant cost data during cost estimation for new
products.
A typical example can be seen in the methodology presented by Rehman and Guenov [3] in
which an attempt is made to predict design features from incomplete design descriptions
based on past designs and production knowledge. In their work, a system allows the
retrieval of past design cases that match the new problem description. The cost modeller
detects necessary modifications in the retrieved designs and the cost data is updated
accordingly using the adaptation rules stored within the design models. The method, thus,
allows the cost estimation and evaluation for innovative designs. The method applied by Li-
Hua and Yun-Feng [65] evaluated costs for new products by implementing the functions of
CBR. These functions included organizing case bank, indexing case, initializing case,
seeking and searching case and adapting case. The method was useful for rapid costing to
satisfy customers’ demands on pricing. Ficko et al. [66] conceived a CBR system for
predicting total cost of the tool manufacture. The system is based on extracting geometrical
features from CAD-models stored in a database and calculating the similarities with the
problem description of a new product’s features.
62
Although, the developed system is only limited to tools for manufacture of sheet metal
products by stamping, it provides good-quality predictions based on enough similar cases.
Balarman and Vattam [67] analyzed other applications of CBR in the domains of help-
desks, diagnosis, cost estimation and design based on its functions such as representations,
indexing, matching, adaptation and process of problem solving. Their work is an effort
towards building a general-purpose case-based problem solver.
3.3.2 Decision support systems (DSS)
These systems are helpful in evaluating design alternatives. The main purpose of these
systems is to assist estimators in making better judgments and decisions at different levels
of the estimation process by making use of the stored knowledge of experts in the field.
This is illustrated in Figure 3-3. One such system incorporating expert rules has been
developed by Kingsman and De Souza [68] for cost estimation and price setting in versatile
manufacturing companies dealing with make-to-order (MTO) systems. The developed
system not only focused onto different factors that influence the decision making process in
handling a customer enquiry but discussed the rules that cost estimators apply when making
decisions about these factors.
To incorporate experts’ experience, the artificial intelligence (AI) philosophy is used to
represent and utilize a domain expert’s knowledge in a way that is oriented towards
problem solving and serves as a decision aid tool. In the particular context of the PCE, for
example, it may constitute a segment of the system containing information about machining
63
processes, manufacturability analysis and constraints, product characteristics with design
functions and relationships with each other set out in logical statements. It may also
incorporate rules about the actions to be taken or more conventional mathematical
formulae. It can point outside to external programs and databases that can be associated
with it including some that can cope with uncertain or conflicting judgments.
Figure 3-3: Decision support system approach to cost estimation
64
Shehab and Abdalla [69] developed knowledge-based cost models for the PCE in early
design stages whereas Luong and Spedding [24] developed a knowledge-based system by
integrating process planning into cost estimation. Another approach adopted by Gayretli
and Abdalla [4] focussed on developing a prototype system for manufacturing process
optimization. The system assisted designers to create real-time cost estimates and feasible
process plans by retrieving manufacturing form features and parameters from the feature
database.
One of the most common ways to represent DSS is based on storing design, manufacturing
or other constraints as a set of rules. Since many practical situations deal with uncertainty
and non-availability of heuristic data, fuzzy logic techniques are used to some extent to
overcome such problems. Another non-conventional approach makes use of Expert systems
(ES) or Expert Support Systems in the domain of DSS.
Rule-based systems
These systems are based on process time and cost calculation of feasible processes from a
set of available ones for the manufacture of a part based on design and/or manufacturing
constraints. Such a system reflects these constraints in a respective rule class with the
information encapsulated in it by an expert in the area. A rule-based algorithm is an
example of one such approach, which helps to establish design and manufacturing
constraints. This approach is shown diagrammatically in Figure 3-4. Based on a set of user
constraints, manufacturing processes are selected, which are then used to calculate the
65
product cost. The set of constraints may need to be changed to obtain a different set of
manufacturing processes to obtain an acceptable product cost estimate. This methodology is
helpful for cost optimization based on process evaluation criteria. However, obtaining the
optimized results can be very time-consuming especially, when there are a large number of
processes to be evaluated.
Figure 3-4 Cost Estimation Process Model Based On User Constraints
Gayretli and Abdalla [70] developed a rule-based algorithm for the selection and
optimization of feasible processes to estimate process time and cost based on parts features.
66
A detailed description of part features with possible processes and constraints was given.
Process times were calculated using a standard formula as follows:
FormFeatureVolume
Pr ocessTime =
Material Re movalRate (3-1)
The process time is then used to calculate Lot Time, which is based on a form feature
quantity. The total process cost is subsequently calculated as follows:
Total Process Cost = Lot Time × PHC (3-2)
Where, PHC is the Productive Hour Cost given by a cost estimation database [71]. The
total cost is then calculated as follows:
Total Cost = Material Cost + ∑ [(Lot Time × PHC) + Tool Cost + Setup Cost] (3-3)
The proposed system allowed the selection of a combination of feasible processes from the
possible ones and hence the calculation of process time and cost based on the user input
constraints, e.g. maximum allowable cost and process time for a particular feature. A
criterion of feasibility was judged against the level of satisfaction for input constraints. The
process allowed flexibility based on user constraints. Another example of this category can
be found in [3] where manufacturing and assembly rules were used to update cost data in
67
the proposed system, whereas, an object-oriented and rule-based system can also be found
in [64] for product cost modelling and estimation.
Fuzzy logic approach
This approach to cost estimation is particularly helpful in handling uncertainty. Fuzzy rules
such as those for design and production are applied to such problems to get more reliable
estimates. However, estimating the costs of objects with complex features using this
approach is quite tedious and requires further research in the area. Shehab and Abdalla [63]
used fuzzy rules with linguistic expressions and assigned truth-values to them. They used
several steps to develop a fuzzy logic model. These steps were fuzzification of input
variables followed by fuzzy inference based on a set of rules and finally defuzzification of
the inferred fuzzy values. A fuzzy technique consisting of a decision table providing a
means for system rules and indicating the relationships between the input and output
variables of the fuzzy logic system, is used to handle the uncertain knowledge on cost
estimation. The construction of a set of rules from the decision table enables the estimation
of the machining time (Ti) for a given feature, which is multiplied by the unit time cost (Ri)
to get the machining cost (Cm) for that feature, i.e.:
Cm = Ri Ti (3-4)
The developed fuzzy logic based system was capable of estimating the total product cost
apart from enabling the material selection and estimating the assembly cost. The same
68
investigators [64] carried out a similar but more comprehensive study by considering other
essential costs such as non-productive and set up costs.
Expert systems
This approach is based on storing the knowledge in a database and manipulating it on
demand to infer quicker, more consistent and more accurate results based on an attempt to
mimic the human expert thought process with the help of an automated logical reasoning
approach, normally achieved by rule based programming. Within the specific context of
cost estimation, the expert system approach refers to a model and associated procedure
exhibiting a degree of expertise comparable to that of a human expert in generating or to
help in generating reliable cost estimates. Expert systems applied to the PCE have mainly
focussed on formalising the theoretical techniques largely from textbooks etc. rather than
encapsulating the practical knowledge (e.g. the expert conceptual estimator developed by
Musgrove [72]). Further research in the area considering the human contemplating process
of an estimator has a better potential to exploit the typical characteristic. Cost estimation
methods for various applications using expert systems or expert support systems can be
found in [73, 74].
69
3.4 Analogical cost estimation techniques
These techniques employ similarity criteria based on historical cost data for products with
known cost such as regression analysis models or back propagation methods. The following
sub-sections describe these methods in detail.
3.4.1 Regression analysis models
These models make use of the historical cost data to establish a linear relationship between
the product costs for the past design cases and the values of certain selected variables so
that the relationship can be used to forecast the cost of a new product. The regression
analysis approach based on the similarity principle was adopted by Hundal [75] and Poli et
al. [42] to use a basic cost value and consider the effects of variable cost factors by
assuming linear relationships between the final product cost and the cost factors. Lewis [76]
further used existing designs to provide cost estimates for similar new designs whereas Pahl
and Beitz [5] provided more general costing approaches based on similarity.
3.4.2 Back-propagation neural network (BPNN) models
These models use a neural network (NN) that can be trained to store knowledge to infer the
answers to questions that may even not have been seen by them before. This means that
such models are particularly useful in uncertain conditions and are adaptable to deal with
non-linearity issues as well. The back-propagation neural network (BPNN) is the most
70
common of all network types and also suits better the nature of the PCE. The application of
neural network in cost engineering is discussed in [77].
Shtub and Zimerman [78] compared the cost results obtained with the regression model and
the back-propagation neural network model and observed the superiority of the latter in
many ways. In another study [40], a featured-based methodology was proposed using
BPNN to estimate the cost of packaging products. Cost-related features of packaging
products were used in conjunction with historical cost data to obtain a relationship between
cost and cost-related features based on BPNN. The proposed method overcame the
limitations of regression analysis models such as the assumption of non-linear relationships
between product cost and its variables as well as those of traditional breakdown
approaches, e.g. the requirement of detailed cost information like process planning cost.
Zhang and Fuh [41] proposed a similar approach for early cost estimation, whereas Chen
M-Y and Chen D-F [79] proposed a BPNN model for strip-steel coiler. Further, a back-
propagation algorithm [2] was used with momentum and a flat spot elimination term for a
Multilayer Perceptron (MLP) Neural Network, in which neurons are organized in several
layers including an input layer, a number of hidden layers and an output layer.
3.5 Parametric cost estimation techniques
Parametric models are derived by applying the statistical methodologies and by expressing
cost as a function of its constituent variables. These techniques could be effective in those
71
situations where the parameters, sometimes known as cost drivers, could be easily
identified. Parametric models are generally used to quantify the unit cost of a given
product. Cavalieri et al. [2] developed a parametric model for the estimation of unit
manufacturing costs of a new type of brake disk using the weight of the raw disk, unit cost
of raw material and the number of cores as parameters in their model, which is expressed as
follows:
 C TF 
C = FC +  C co N co + rm W (3-5)
 1 − SC 
where,
C = Unit cost of disk brake, FC = Fixed Cost Factor (Coefficient), Cco = Core Cost per Kg
of Cast iron (Coefficient), Nco = Number of cores, Crm = Unit cost of raw material, SC =
Scrap rate (Coefficient), TF = cast iron / steel conversion factor (Coefficient), W = weight
A simple linear regression model using one of the cost drivers would not be effective
because of variances between the data. However, the developed model overcame this
problem by using more parameters. Validation analysis of the model by comparing the
estimated costs with the actual ones of the brake disks demonstrated the superiority of the
proposed parametric model over the linear regression model.
A wide range of parametric models can be found in the literature. For example, Hajare [80]
modelled parametric costing of components using the product specifications. Roberts and
72
Hermosillo [81] used approximate tool paths and process parameters from available factory
resources to estimate time and cost for small surface units. Boothroyd and Reynolds [82]
adopted a parametric costing approach using the volume of typical turned parts as a
parameter to estimate the cost in the early design stages. Unlike the detailed-breakdown
approach, the method adopted by them could be used in the early design stage without the
need of a process plan. Similar work can be found in [83].
3.6 Analytical cost estimation techniques
This approach requires decomposing a product into elementary units, operations and
activities that represent different resources consumed during the production cycle and
express the cost as a summation of all these components. These techniques can be further
classified into different categories, which are discussed in detail below.
3.6.1 Operation based approach
This approach is generally used in the final design stages due to the type of information
required and is one of the earliest attempts to estimate manufacturing costs. The approach
allows the estimation of manufacturing cost as a summation of the costs associated with the
time of performing manufacturing operations, non-productive time, and set-up times.
Several techniques have been developed to select the alternative manufacturing operations
that optimise the machining cost.
73
The cost model proposed by Jung [84] estimated the manufacturing cost by considering
three different times including set up time, operation time and non-operation time.
Formulation was provided. The total cost was given by:
Mfg cost = (Ro + Rm)[( Tsu/Q)Tot + Tno] + material cost + factory expenses. (3-6)
where, Ro = operator’s rate, Rm = machine rate, Tsu = set-up time, Q = batch size, Tot =
operation time, Tno = non- operation time.
The model could not be used to evaluate design alternatives due to its availability only in
the final stages of design cycle. Feng et al. [12] presented a digraph based mathematical
model that uses the geometric features including cylinder, rectangular block, chamfer, flat
surfaces and hole, and developed an algorithm to estimate the minimum cost using an
operation-based approach. The process plans of alternative design solutions with explicit
modelling of the machining time of various features represented the criteria of estimating
manufacturing costs. Gupta et al. [13] developed a similar methodology using
manufacturing features for the evaluation of alternative process plans to estimate the
manufacturing cost of the part. Wei and Egbelu [85] used geometric design data and
developed a method based on a tree representation of alternate processes to estimate the
product manufacturing cost. Although the approach focussed on obtaining the optimum
results, it did not consider direct labour cost. Further, Kiritsis et al. [14] proposed a method
for the cost estimation of the machining of parts based on the description of given features
and associated alternative manufacturing operations. The proposed methodology was based
74
on Petri nets to determine overall costs including machining cost, moving cost, setup cost
and tool change costs. However, getting the optimised results using the proposed
methodology was time consuming.
3.6.2 Breakdown approach
This method estimates the total product cost by summing all the costs incurred during the
production cycle of a product, including material costs and overheads. The method requires
detailed information about the resources consumed to manufacture a product including
purchasing, processing and maintenance details.
The cost model developed by Son [86] included labour costs, machining cost, tool cost, set
up cost, space occupied cost, computer software cost and material cost. The model also
separated the raw material cost and the labour cost into different categories. The proposed
model included insurance, utility, maintenance, repair and property costs. The machining
cost (Cm,), is hence represented in the following equation as
Cm= (utility cost) + (maintenance cost) + (repair cost) + (insurance cost) + (property cost)
= Σ (CuTm + CmtTmt + CrTr + aFk + bFk) (3-7)
75
where, Cu = utility cost per unit time, Tm = machining time, Cmt = maintenance cost per
unit time, Tmt = total maintenance time, Cr = repair cost per unit time, Tr = total repair time,
a = insurance premium, Fk = initial investment, and b = property tax.
Further, equations for other cost elements including labour costs, tool cost, set up cost,
space occupied cost, computer software cost and material cost were also provided. The
requirement of such detailed information restricted the use of the model in the final design
stage. Further, manufacturing costs were considered by Bernet et al. [43] as the sum of
material, labour and overhead costs and Ostwald [60] estimated product cost as the
summation of material cost, manufacturing cost, labour cost and overhead expenses based
on hourly usage of machinery or direct labour. Such traditional cost estimation and cost
accounting techniques were also discussed in detail in [61].
3.6.3 Tolerance-based cost models
The objective of such models is to estimate product cost considering design tolerances of a
product as a function of the product cost.
Singh [87] presented a framework for the concurrent design of product and processes
considering the criteria of minimum cost, maximum quality and minimum manufacturing
lead time. Three models were presented to jointly design the products and processes. They
are unit cost of production model, the quality model and the lead time model. The unit cost
model was expressed as follows:
76
X0 (d, j) = Ki (d, j) [Xi + f(j)] – Ks (d, j)Xs (3-8)
where, Ki and Ks are technology coefficients that can be found from the following
equations:
Ki = 1/ [1- SC (d, j)] (3-9)
Ks = SC (d, j)/ [1-SC (d, j)] (3-10)
SC is the scrap rate given by the following equation:
SC (d, j) = Ø[-d / σ (j)] + 1 – Ø[d / σ (j)] (3-11)
where, j = jth manufacturing process selected for producing a product, Xo (d, j) = the unit
cost with tolerance d, Xi = the unit raw material cost, f(j) = the unit processing cost for jth
process, Xs = the unit salvage value, Ki = technology coefficient (input), Ks = technology
coefficient (scrap), SC (d, j) = scrap rate, σ (j) = standard deviation, Ø (x) = Cumulative
distribution function of probability distribution with the mean equal to 0 and the standard
deviation equal to 1.
The modelling methodology was based on obtaining the optimal tolerances and hence
setting up the acceptance regions for the design variables meeting certain criteria. The
77
objective of the cost model was to select the process and design variables that minimize the
cost function. However, the modelling methodology eliminated the needs for design
changes because it considered various design and manufacturing factors at the early stage
of the design. The cost-tolerance relationships and relevant models can also be found in
[88, 89].
3.6.4 Feature-based cost estimation
The feature-based cost estimation methodology deals with the identification of a product’s
cost related features and the determination of the associated costs. A considerable research
has been carried in order to extract and quantify representative product features that
contribute to the total cost. These features can either be design related such as the type of
material used for a specific product, geometric details, etc. or process oriented, i.e. a
particular process required for manufacturing the product, e.g. machining, casting,
injection-moulding etc. The methodology allows the selection of a particular design or
manufacturing form feature for DFC system users. However, the approach can have
limitations for complex or very small geometric features especially if machining processes
are used to produce these features.
Zhang et al. [40] proposed a feature-based cost estimation system for packaging products
extracting 32 cost related features in both the design and manufacturing domain. These
features were then quantified based on a relative cost influence among the various possible
states of a particular feature. However, no attempt was made to quantify these features
78
objectively. Ou-Yang and Lin [6] looked into the feature-based costing by focussing on the
machining type features and developed a manufacturing cost estimation model based on
feature shapes and precision. With the process planning information and geometrical data,
the machining time of a feature was estimated. One limitation of their proposed framework
was that it only considered conventional machining processes. Further research work in the
area of feature-based cost estimation can be found in [90-92].
3.6.5 Activity-based costing (ABC) system
The ABC system focuses on calculating the costs incurred on performing the activities to
manufacture a product. The method was first discussed by Cooper and Kaplan [93]. They
presented the ABC system as a useful means to distribute the overhead costs in proportion
to the activities performed on a product to manufacture it. Hundal [94] presented similar
methods. The ABC system proved a good alternative to traditional estimation techniques
since it provided more accurate product manufacturing cost estimates [95].
Various information sources for the implementation of the ABC system within a specific
context can be found in [96-101]. The effectiveness of the ABC system was discussed by
Kaplan [102] in providing helpful cost information to product designers for developing
economic designs. The capabilities of the ABC were investigated by Tornberg et al. [103]
with a particular emphasis on providing useful cost information to product designers. The
study focused on modelling product design, purchasing and manufacturing processes with
graphic flow charts in the form of activity chains. With the help of the process models and
79
the activity-based cost calculations, product designers were able to estimate the effects of
different design options on product costs. The work of Tseng and Jiang [104] attempted to
combine the feature-based costing methodology with the ABC approach. Their ABC
analysis model could evaluate different manufacturing costs for multiple feature-based
machining methods. Yang et al. [105] used process planning, scheduling and cost
accounting information to estimate manufacturing and machining cost through an activity-
based method. Other examples of manufacturing and machining cost estimation using the
ABC approach can be found in [106, 107].
Some other researchers used the ABC approach to model the manufacturing costs in a
specific manufacturing set-up. For example, Koltai et al. [108] estimated costs for flexible-
manufacturing systems based on the ABC analysis, whereas Aderoba [48] developed an
ABC model for job shops. The latter was based on the classification of all the activities into
machine-based production, labour-intensive production, technical services and
administrative services. Cost rates for all such activities were provided which were then
used to estimate the cost of a new order. For example, the cost of a machine-based
production activity (C) was given as follows
C = [(M+m)tm + (L+l)tl + btb + utu] (3-12)
where, M = cost rate for machines in the activity per hour, m = periodic rate for
complimentary tools in the activity, L = labour rate for direct worker on machine activity, l
= labour rates for ancillary worker on machine activity, b = building space rate for machine
80
activity, u = utility rate for machine activity, tm = machining time, tl = labour working time,
tb = time spent on building space occupied by activity, tu = time of utility being used.
Expressions were also provided for the cost rates for machines and tools, the labour rates
for direct and ancillary worker, the building space rate, and the utility rate, which are not
described here for brevity of presentation. The proposed method proved useful in
highlighting high cost elements; however, its accuracy depended on how reliable the
activity time estimates for a new product were.
3.7 Conclusions
This chapter extensively reviewed the pertinent literature on manufacturing and product
cost estimation along with a critical evaluation of some of the techniques developed in the
area. An extensive classification scheme is developed. A pictorial representation of the
classification system adopted in the present study is given in Figure 3-5. The horizontal
dotted lines are used to show the different levels in the hierarchical tree diagram proposed.
The techniques were classified into two main groups as qualitative and quantitative, which
were then subdivided into two categories each. The chapter examined all the categories in
detail with references to the published literature. Mathematical models were presented on
some occasions to illustrate certain techniques. In addition, the PCE techniques discussed
81
in the chapter are tabulated together with the key advantages, limitations and corresponding
published literature in Table 3-1.
The study of individual techniques also revealed the key conditions under which they can
be applied. The conditions can be grouped together to form a decision support model
(DSM) for cost estimation methodology selection and is presented in Figure 3-6. The
developed model is a helpful tool for estimators in making decisions about selecting a
suitable estimation methodology. It can be observed that a particular technique linked with
a specific class is more applicable in certain situations. During the early phases of the
design cycle, when limited data is available, qualitative cost estimation techniques are more
appropriate and provide a helpful starting point for a detailed analysis at a later stage. For
example, the proposed case-based methodology systematically makes use of available past
data to generate estimates for a similar new product. One problem linked with such
techniques is the limited availability of past data, which is overcome to some extent by
making use of the past experience or knowledge of the estimator generally encapsulated in
the form of decision rules. Qualitative techniques, therefore, are helpful either in furnishing
rough cost estimates or serve as a decision aid tool for designers or estimators especially
during the early phases of design process. However, when the detailed design becomes
available, quantitative techniques provide more accurate estimates, which are necessary for
factors like design rationalization and determination of profit margins etc. The data
requirements restrict the use of such techniques in the final phases of design and
development process. Techniques such as the ABC systems overcome the problem to some
extent by making use of the pre-determined activity rates to calculate the total amount of
82
activities consumed to manufacture a product rather than requiring any detailed design and
manufacturing information. This, however, requires lead times for individual products in
the early design stages, which may be obtained using methodologies such as the case-based
approach. Therefore, a combination of the two approaches, the qualitative and the
quantitative techniques, could play an important role in developing a cost evaluation system
capable of providing useful cost information on various stages of design and development
phases.
The proposed classification system and the developed decision support model are aimed at
supporting the decision making process of the estimators and designers. The two elements
are formulated to provide guidelines to the users for the selection of an effective estimation
methodology. A rightful selection is eventually an aspect of a CCS.
83
Product Cost Estimation
Techniques
Level 1
Intuitive Techniques Analogical Techniques Parametric Techniques Analytical Techniques

Level 2
Regression Analysis Back-Propagation

Model Neural Network Model
Level 3
Decision Support
Techniques Operation-
Based Approach Tolerance- Activity-Based
Based Cost Cost Estimation
Case-Based Models
Technique Breakdown
Approach Feature-Based
Cost Estimation
Level 4
Rule-Based System Fuzzy Logic System Expert System
Figure 3-5: Classification of the PCE Techniques
84
Table 3-1: The PCE techniques; key advantages, limitations and list of discussed references
Product Cost Estimation Techniques Key Advantages Limitations References

Intuitive Cost Estimation Techniques
Case-Based Systems Innovative design Dependence on past cases
[3], [65-67]
approach
Qualitative Cost Estimation Techniques
Decision Support Systems Rule-based Can provide Time-consuming [3], [64],

Systems optimized results [70]
[4], [24], [68-69]
Fuzzy logic Handles uncertainty, Estimating complex
[63, 64]
systems reliable estimates features costs is tedious
Expert Quicker, more Complex programming

Systems consistent and more required [72-74]
accurate results
Analogical Cost
Regression Analysis Simpler method Limited to resolve [5], [42],

Techniques
Estimation
Model linearity issues [75, 76]
Back Propagation Deal with uncertain Completely data-

[2], [41],
neural network model and non-linear dependant, Higher
[77-79],
problems establishment cost
Utilize cost drivers Ineffective when cost

Parametric Cost Estimation
effectively drivers can not be [2], [80-83]
Techniques
identified
Operation-based cost Alternative process Time-consuming, require

Quantitative Cost Estimation Techniques
models plans can be detailed design and [12-14],

evaluated to get process planning data [84, 85]
Analytical Cost Estimation Techniques
optimized results
Breakdown cost models Easier method Detailed cost information

[43], [60,
required about the
61], [86]
resources consumed
Cost tolerance models Cost effective design Require detailed design

tolerances can be information [87-89]
identified.
Feature-based cost Features with higher Difficult to identify costs

[6], [40],
models costs can be for small and complex
[90-92]
identified features
Activity-based cost Easy and effective Require lead times in the

[48], [93-
models method using unit early design stages
108],
activity costs
85
Input for selected methodology
Start
Selected methodology
Decision Node / Available alternatives

NO YES
Estimates
at Back
1 Apply quantitative techniques conceptual Apply qualitative techniques propagation
design algorithm
stage?
YES
NO
Past
Past
product Apply analogical
experience
Parametric YES data techniques
available?
relation available?
available?
NO YES NO
Apply analytical techniques Apply parametric model Apply intuitive NO

Apply
techniques Resolving
BPNN
linearity?
models
NO Past
Process planning Operation-based
product
details / operation models
Product
1 cases YES
times available available?
parameters.
Breakdown Shape, Size,
Cost elements’ Weight etc. Apply regression
breakdown available models
models
Past NO
Cost as a function Tolerance-based experience YES
of design models stored in AI
tolerances technology? Apply case- Linear
Rule-based based relationship
Individual features Feature-based algorithm, technology using
cost known models Fuzzy logic- historical
based system, YES cost data
Expert system Matching
Unit activity costs ABC Apply DSS algorithm
known models
Stop
Figure 3-6: Decision Support Model for cost estimation methodology selection
86
Chapter 4 MRO and TRO
Estimation Methods
Overhead is a major element of the manufacturing cost. The chapter identifies the need
to implement an improved overhead estimation methodology based on dividing
manufacturing overheads into time- and material-related costs. New overhead
estimation methods called material-related overhead (MRO) and time-related overhead
(TRO) are, therefore, introduced, developed and implemented in an electrical
engineering company for a four-year period. The results based on a four-year
retrospective validation analysis confirm the superiority of the proposed methodology
over the existing one, as overheads are more accurately estimated.
4.1 Introduction
The aim of Chapter 4 is to present material-related overhead (MRO) and time-related
overhead (TRO) estimation methods. This is essential to lay the foundation for
developing a comprehensive methodology for early and accurate estimation of a product
cost. It has already been considered in previous analysis that early and accurate
estimation would require adopting a hybrid approach combining the elements from
qualitative and quantitative techniques. It has also been shown that developing a
methodology for estimating a product’s cost in a batch type manufacturing environment
87
Chapter 4: MRO and TRO Estimation Methods
is in conjunction with a greater need for the system than for either mass production or
job shop system. Further, the development of a methodology to predict an entire
product’s cost rather than part or component costs demand a careful analysis of the
developed classification system. Breakdown approach presents the entire product cost
by summing up all the costs incurred during the production cycle of a product and
could, therefore, hold the key to start the development process. In a broader sense, the
development of a hybrid system for product cost estimation in a batch type
manufacturing environment is, therefore, congruent to the research area.
Developing such a system, however, requires identifying the elements from the
qualitative and quantitative techniques for hybrid approach. One of the identified
elements is from the quantitative techniques and is based on predicting an entire
product’s cost. Breakdown approach requires defining the total cost as a summation of
the constituent elements. Before these elements and the total cost are defined; there is a
need to understand any existing breakdown models and to identify problem areas.
Product cost consists of general administrative costs, engineering costs (costs incurred
during design and development phases including those linked with customer
requirements, designing of the part, its process planning, and prototyping), and
manufacturing costs. The latter is the largest element of the overall product cost and is
mostly determined during the product design phase. Better understanding and control of
this cost element is a major step towards achieving an effective CCS. Figure 4-1 shows
a breakdown of the selling price and manufacturing costs of a typical product [109]. It
can be seen that the manufacturing cost accounts for 40 percent of the selling price. Half
of this cost is incurred on parts and material, whereas direct labour accounts for 12
88
percent. Manufacturing overheads (indirect labour, indirect materials, plant and
machinery depreciation, energy costs, etc.) account for 38 percent of the manufacturing
costs, which is a substantial figure.
Figure 4-1: Break down of Selling Price and Manufacturing costs
Manufacturing overhead is a significant contributor to the overall product cost and
refers to all the production support costs incurred on or off the shop floor and generally
for the shared benefit of several products within a manufacturing facility. A traditional
approach to estimating such costs relies on summing up all the expenses, other than
those included in direct material and labour, and ascertaining a cost rate. The rate is then
used to estimate the overheads for a new product by multiplying it with the expected
manufacturing lead time (MLT) of the new product. A major disadvantage of this
methodology is a disproportionate allocation of the total manufacturing overheads to
different products with respect to the actual resources consumed by individual products.
Furthermore, since the overheads recovery is based only on MLT and ignores material
quantities, costs are underestimated for products with higher material requirements and
overestimated for those with lower material quantities.
89
Since, overhead is a major element of the overall product cost and estimating it
accurately has a direct bearing on the overall estimated cost value, the current chapter
will focus on developing a new overhead estimation methodology. The proposed
methodology will then form the basis for the development of a model for the estimation
of the overall product cost. The methodology developed in this chapter divides
overheads into time- and material-dependent cost elements and overcomes the above
mentioned limitations by providing a framework for apportioning the total
manufacturing overheads to individual products based on the product’s degree of
consumption of manufacturing activities and resources.
Product cost depends greatly on the geographical location of an enterprise owing to
factors (such as governmental policies, taxation, political situation, weather, labour cost,
and material availability) that could significantly alter the manufacturing cost of a
product on different locations around the globe. The contributions of individual sub-
elements (material costs, direct labour, and overheads) towards the overall
manufacturing cost also vary within this context. It is not surprising that many western
manufactures have moved their operations to Asian countries.
With this in mind, the present study analysed manufacturing cost and its various sub-
elements in an electrical engineering company in South Asia. It is also important to note
that any validation through industrial trials will not only make an attempt to endorse the
model but help to understand the implications of geographical locations. The developed
mathematical model for overhead estimation is implemented retrospectively in the
company for a comparative analysis covering a four-year period. The study sets
precedence for similar studies on other geographical locations with the aim of providing
90
a good foundation for comparative analysis in order to better understand manufacturing
costs and develop an effective CCS.
4.2 Cost estimation methodology at the selected company
A well-established ISO9001 certified electrical engineering company was selected for
the study as a representative case of the industrial sector in South Asia because it
follows the standardized industrial code of practises acknowledged internationally.
Because of the sensitivities around the confidentiality issues related to competitive cost
data provided by the company, the company’s name is not mentioned here. The
company designs, develops, manufactures, and markets a wide range of technologically
advanced electrical and electronic products and provides after sales services. The
transformers manufacturing unit of the Power Transmission & Distribution (PTD)
division of the company was chosen for the study, as it is the largest manufacturing unit
of the company based on a batch type manufacturing set up. The unit deals with the
manufacturing of distribution transformers (DT) and power transformers (PT). One of
the major reasons for selecting the company in general and the transformer unit in
particular is the unit’s existing method of cost estimation that estimates the entire
product’s cost based on a cost breakdown approach. Three major cost elements make-up
the total manufacturing cost of transformers at the company: material costs, direct
labour costs, and overheads. These are discussed in more detail below.
91
4.2.1 Material cost estimation
The material costs are estimated using the material quantities obtained directly from
BOM (see example in appendix A) which is developed based on the engineering design
details. Scrap margin is accounted for based on past trends. In most of the cases
standardized transformers are manufactured and thus engineering designs are already
established and so are BOMs. Even when a non-standard transformer is designed and
developed, the quantities of different materials required are similar to a closest standard
match even though the technical design of the product may be significantly different.
Thus, the match provides a good starting point for the non-standard transformer to
estimate material cost. As a result, material cost estimation is possible in the early stages
of the design and development stage. In such circumstances, material cost calculation,
based on past trends and certain commercial indicators, becomes standard. The
following commercial indicators were used based on past trends in order to calculate
material cost for the year 2003-2004:
• 15 percent increase in purchase price of locally sourced material;
• 10 percent increase in price of materials bought from foreign suppliers;
• Exchange rate = US $61.4 and 71 Euro (i.e. US $1.0 = Rupees (Rs) 61.4
and Euro 1.0 = Rs 71.0).
The cumulative material requirements (CMR) for transformer manufacturing for a given
period of time change in accordance with planned orders. Thus, the factory-wide
changes in material consumption over different time intervals can be used to reflect
changes in production levels.
92
4.2.2 Direct labour costs
Work centres in both DT and PT comprise operators and machines. Direct Labour here
accounts for the cost incurred on wages for total man-hours required for production. The
total manufacturing lead time (MLT) for a transformer is calculated using standard
routings and process times. These process times are established either by using standard
formulas or by deploying time and motion studies. The lead time for a non-standard
transformer is determined from the results for the closest standard match. In this way,
the total number of man-hours required for manufacturing a transformer can be
determined in the early stages of the design process and are subsequently multiplied by
the wage rate to give the labour cost incurred. Setting the wage rate for a particular
period of time is an organizational policy matter which is influenced by various factors
(such as minimum wage rate set by the government, geo-political conditions, economic
growth, etc.).
4.2.3 Overheads estimation
Overheads here refer to manufacturing overheads and account for costs other than direct
material and direct labour. Since such costs represent more than a third of total
manufacturing costs (see figure 4-1), estimating it accurately can significantly enhance
the accuracy of the overall PCE process. The new methodology proposed in the study is
implemented retrospectively for a four-year period starting from 1999-2000. The
methodology employed by the company is discussed below in order to facilitate
comparisons between the two methods and provide insight into the optimisation
achieved through the implementation of the proposed methodology.
93
The existing method is based on summing up all the expenses other than those incurred
on direct labour and material. For example, the costs incurred on handling,
transportation, inspection, storage, inventory control, and purchasing of different
materials are classified as overheads. Energy and other utility costs necessary to run
different workstations are also included in this category. Several small components such
as nuts, bolts, washers, etc. are classed as indirect materials and the costs incurred on
these items form part of the overheads as well. By summing all the above-mentioned
costs and dividing by the capacity hours, the overhead rate for the following year is
calculated. Capacity hours in this context are the hours available for production. The
rate is then used to estimate the overheads for a new transformer by multiplying by the
MLT of the transformer. The total transformer manufacturing cost is finally estimated
by adding the three cost elements: material cost, direct labour cost, and overheads.
The major drawback of this methodology is that the allocation of overheads to an
individual transformer in most cases is not in proportion to the material and energy
consumed for its production. For example, a transformer with a comparatively lower
lead time is allocated lower overheads even though the quantity of material required for
its manufacture is higher. Thus, the amount of material required for the manufacture of
a transformer is ignored with respect to overheads allocation. This results in
underestimation for transformers with higher material requirements and overestimation
for those with lower material requirements.
94
4.3 Proposed methodology for overheads estimation
A careful analysis of the overheads at the selected company revealed that they were
largely driven by either material quantities or lead times. In some cases, however,
material quantities and lead times may interchangeably influence the same overhead
cost element. For example, the cost incurred on tools and equipment is greatly affected
by lead times which in turn may be linked with the amount of material to be processed.
However, based on consistency, a more direct and stronger link can be established only
between tooling costs and lead times as in some cases lead times may be independent of
material quantities. The proposed methodology for overheads allocation and estimation
is, therefore, based on dividing the overheads into MRO and TRO. This allows more
realistic allocation of overheads to individual products by taking care of not only their
lead times but their respective material quantities. Mathematical models are developed
for the estimation of these overheads in the present study and are validated against
transformer data obtained from the selected company.
4.3.1 MRO estimation model
MRO include the costs incurred on material handling and transportation, material
inspection, material storage and inventory control, material purchasing etc. It is
important to note that activity times (such as time consumed for material handling,
inspection etc.) are not same as MLT that form the basis for TRO. These activity times
are effected by variations in material quantities and hence any costs associated with
them are considered MRO. The costs incurred on indirect materials are also included in
MRO. As the material consumption increases, MRO also go up. Total MRO consumed
95
during a certain period of time, therefore, can be expressed as a percentage of total
direct material cost. A fraction can be set to express the same at the end of the (n-1)th
year based on total MRO and the cumulative direct material cost obtainable at the end
of the (n-1)th year from the cost accountancy data. The resulting MRO fraction (which is
a reflection of the MRO as a percentage of the direct material cost) is then multiplied by
the total material cost of an individual product to calculate the MRO for a new product
(OM) in the nth year. Total material cost for the product is based on the estimated
material quantities for that product obtained from the BOM and their respective unit
costs, i.e.
(C imn−1 + C nft−1 + C Pn −1 + C in −1 + C sin −1 − S )

Om = (C d 1 m1 + C d 2 m2 + .... + C dn mn ) × n −1
(4-1)
C dmt
Where,
Cd1, Cd2, …., Cdn = Unit costs of direct material 1, direct material 2, …., direct material
n respectively used in the manufacture of the new product; m1, m2, …., mn = amounts of
material 1, material 2, …., material n consumed; Cimn−1 = overall indirect material costs in
the (n-1)th year, C nft−1 = overall freight & transportation costs in the (n-1)th year,
C Pn −1 = overall purchase department costs in the (n-1)th year, Cin−1 = overall inspection
costs in the (n-1)th year, Csin−1 = overall stores & inventory costs, S = Sale proceed of
n −1
scrap, Cdmt = overall direct material costs in the (n-1)th year.
96
It can be observed that the model can easily accommodate any other MRO to adapt to
the system where it is implemented. For example, if a system’s quality control costs are
material dependent, equation (4-1) can easily accommodate such costs to adapt to the
new system. The percentage fractions of MRO calculated based on cost data (PT & DT)
provided by the company for the year 2003-2004 are shown in Table 4-1. All the values
given in the table are in (000) Rs. and the scrap value is given in brackets to denote
subtraction in the model. The MRO for a new transformer can then be calculated using
the percentages obtained and the amount of material used for its manufacture.
Table 4-1: MRO for power and distribution transformers (2003-2004)
Material Related Overheads for Transformers (2003-2004)
Power Distribution Total

Transformers Transformers Transformers
(PT) (DT)
Indirect Materials 6,500 19,500 26,000
Sale Proceeds of Scrap (1,000) (3,000) (4,000)
Freight & Transportation 500 1,500 2,000
Purchase Department Costs 1,478 4,433 5,910
Incoming Inspection 790 2,370 3,160
Stores & Inventory Control 1,350 4,050 5,400
Total MRO 9,618 28,853 38,470
Total Material Consumption 203,000 445,700 648,700
MRO (2003-2004) 5% 6%
97
4.3.2 TRO estimation model
TRO refers to the overheads other than the MRO and are proportional to MLT. For
example, energy and other utility costs rise when the lead times increase. Different
workstations are identified and all the costs incurred on running these stations are
gathered together. For example, energy and other utility costs, repair and maintenance
costs, costs incurred on special tools and spares, financing expenses, etc. are grouped
under TRO. The total TRO can easily be obtained from cost accountancy data at the end
of the (n-1)th year and are divided by the total capacity hours to yield the budgeted TRO
rate for the nth year. The TRO for a new product (OT) can then be estimated by
multiplying the TRO rate with the MLT of the new product. i.e.
n −1
TROtotal
OT = MLT × n −1
(4-2)
C total
Where,
n −1 n −1
TROtotal = total TRO in the (n-1)th year; C total = total capacity in the (n-1)th year
TRO rates calculated based on cost data (PT & DT) provided by the company for the
year 2003-2004 are shown in Table 4-2. All the cost figures given in the table are in
(000) Rs. Since, the workstations running costs obtained from accounts include wages
they are excluded from the rates calculation. The TRO for a new transformer can be
calculated using the rates in Table 4-2 and the MLT for the transformer.
98
Table 4-2: Budgeted time related overhead rate calculation (2003-2004)
Power Distribution Total

Cost Centre No.
Transformers Transformers Transformers
8200 9,989 29,966 39,954
8210 31,110 13,333 44,443
8220 126,631 126,631
8230 2,032 6,097 8,129
8240 915 392 1,307
8250 2,248 963 3,211
8260 4,580 4,580
8290 3,126 3,126
Total 46,293 185,088 231,381
Less; Wages -4,400 -11,900 -16,300
41,893 173,188 215,081
Add; Financing Expenses 587 1,760 2,346
General Administration 1,665 4,994 6,658
Total (Rs. in 000) 44,145 179,941 224,086
Total Normal Capacity Hours 91,616 291,017 382,633
Budgeted TRO rate ≅ 482 619
The estimated MRO and TRO for the new product can be added together to give the
overall overheads for the new product. i.e.
O = OM + OT (4-3)
The total transformer cost is finally estimated by adding the three cost elements
discussed: i.e. material cost, direct labour and overheads (MRO and TRO).
99
Table 4-3: Summary of TRO rates, MRO percentage fractions and overhead rates for 4 years
Proposed Methodology Existing Methodology

Material
Capacity
Year consumed MRO
Hours TRO Total Overhead
(Rs.) TRO MRO Fraction
Rate Overheads Rate
(%)
99-00 206,200 511,591 104,400 506 13,114 3 117,514 570
00-01 206,200 440,000 103,000 500 30,000 7 133,000 645

DT
02-03 291,017 407,392 175,800 604 31,371 8 207,171 712
03-04 291,017 445,700 180,000 619 28,853 6 208,853 718
99-00 84,000 237,337 28,350 338 6,629 3 34,979 416
00-01 84,000 182,000 31,000 369 5,986 3 36,986 440

PT
02-03 91,616 112,608 41,300 451 7,359 7 48,659 531
03-04 91,616 203,000 44,145 482 9,618 5 53,763 587
In a similar way, TRO rates and MRO percentage fractions can be calculated for the
remaining three years of the validation period by using the data provided by the
company. Table 4-3 summarizes the TRO rates and the MRO percentage fractions for
the four-year period for both DT and PT obtained by the application of the proposed
methodology. The table also determines the overhead rates for the same duration using
the existing methodology for overheads estimation. The cost figures in the table are
given in (000) Rs. except the final rates. The values can be effectively used to calculate
the TRO, MRO and the overheads (company’s method) for a given product by using
MLT and material cost of the product as mentioned in Table 4-4 for 25kVA
transformer.
100
4.4 Model implementation and validation
Data were collected for a period of 4 years to test the proposed methodology for
overhead estimation. Results were systematically analysed using the cost elements
breakdown discussed earlier. Costs were estimated for a selected range of products
using both the existing and the proposed method for overheads estimation. The results
were compared with actual manufacturing costs obtained from the cost accountancy
data, and superiority of the proposed methodology was demonstrated.
As opposed to the two methods for cost estimation mentioned earlier, the actual
manufacturing costs are based on cost calculation after the actual production and at the
end of the financial year. Although, the precise methodology for the actual cost
calculation is unknown due to confidentiality issues involved, the cost results and a brief
guideline for the methodology involved was provided by the company. The actual costs
of the individual transformers are the results of the redistribution of the plant-wide total
costs incurred during the financial year in which they were manufactured. The actual
costs are based on the actual MLT and materials (including the actual scrap material)
consumed for the manufacture of a product. Total actual overheads are allocated to
individual products based on a combination of factors including order related costs,
tools and equipment utilization for specific products, traceable utility units to individual
products, exclusive packaging, etc. Measurable activities (such as testing, inspection,
quality control etc.) are accounted for by setting activity rates and determining activity
units. A single cost rate is determined for all the other costs based on total production
time. This cost is then allocated to individual products based on their MLT.
101
Table 4-4: Cost estimation for 25kVA transformer
% Variation
Actual manufacturing
Old method New Method from actual
Direct Labour Costs
Material Costs (Rs.)

cost
costs (Rs.)
Overheads (Rs.)
Total estimated
Total estimated
Year
(Rs.)
New method
Old method
MRO (Rs.)
TRO (Rs.)
costs (Rs.)
costs (Rs.)
99-00 250 19400 3962 23612 3520 582 23752 24290 2.79 2.22
00-01 250 21825 4484 26559 3472 1528 27075 28121 5.55 3.72
02-03 257 24662 4949 29868 4199 1973 31092 31682 5.74 1.88
03-04 257 27622 4989 32868 4300 1657 33836 34877 5.78 3.01
Table 4-4 presents cost estimates for the 25kVA transformer. The costs are based on a
MLT of 417 minutes and a wage rate of 36 Rs./hr for 1999-00 and 2000-01 and 37
Rs./hr for 2002-03 and 2003-04. A pictorial representation of the cost estimates
produced using the two methods against the actual costs is shown in Figure 4-2 (a and
b). It is apparent that the estimates obtained by applying the proposed methodology for
overhead estimation are closer to the actual manufacturing costs than those obtained by
the previously employed methodology. For example, the difference between the actual
and estimated costs based on the previously employed methodology is on average 5%
whereas the one resulting from the proposed methodology is on average 2.5%. This is
attributed to the fact that the previous method ignores material quantities in the
calculation of overheads and thus underestimates overheads and subsequently
manufacturing cost.
102
Similar results were obtained for 100kVA, 200kVA, 500kVA and 1000kVA
transformers as shown in Figure 4-2 (c to j).
25 kVA Transformers (a) 25 kVA Transformers (b)
36 7.00
Cost Value (Rs in 000)
% Variation from actual

34 5.55 5.74 5.78
6.00
32
5.00
30 3.72
4.00
28
cost
2.79 3.01
26 3.00 2.22
1.88
24 2.00
22 1.00
20
0.00
1999-00 2000-01 2002-03 2003-04
1999-00 2000-01 2002-03 2003-04
Year
estimated cost (old method) estimated cost (new method) Year
actual cost old method new method
100 kVA Transformers (c) 100 kVA Transformers (d)
55 6.00
5.60 5.48
50 5.00 4.44
45 4.00
2.70
cost
3.00 2.58
40
1.61 1.73
2.00
35 1.03
1.00
30
0.00
1999-00 2000-01 2002-03 2003-04
1999-00 2000-01 2002-03 2003-04
Year
200 kVA Transformers (e) 200 kVA Transformers (f)
60 7.00
5.96
55 6.00
5.10
5.00 4.52
50
4.00
cost
45 3.00 2.66
2.09 2.11 2.31
2.00 1.51
40
1.00
35
0.00
1999-00 2000-01 2002-03 2003-04
1999-00 2000-01 2002-03 2003-04
Year
Figure 4-2 Contd.
103
500 kVA Transformers (g) 500 kVA Transformers (h)
120 7.00

5.89
110 6.00 5.33
4.89
5.00
100
4.00
cost
3.04
90 3.00 2.40 2.54
1.82 2.04
80 2.00
1.00
70
0.00
1999-00 2000-01 2002-03 2003-04
1999-00 2000-01 2002-03 2003-04
Year
1000 kVA Transformers (i) 1000 kVA Transformers (j)
140 7.00

6.11
130 6.00 5.19 5.41
5.00
120
4.00
cost
3.34
110
3.00 2.62
2.26
100 2.01
2.00 1.43
90
1.00
80
0.00
1999-00 2000-01 2002-03 2003-04
1999-00 2000-01 2002-03 2003-04
Year
Year
estimated cost (old method) estimated cost (new method)
Figure 4-2: Cost estimation results for 25, 100, 200, 500 and 1000kVA transformers
These results clearly demonstrate the superiority of the proposed methodology in the
estimation of overheads and subsequently manufacturing costs. It can thus become an
important part of an effective CCS. Indeed, following the analysis presented in this
study, the company has adopted the proposed methodology since 2005-06.
The methodology can be further optimized and fine-tuned by considering the following
points. The gradual yearly increase in the overall estimated and actual costs shown in
Figure 4-2 is due to inflation. Since, the overheads for a particular year are used to
determine the rates for the following year; the rates are slightly underestimated due to
the effects of inflation which are not accounted for resulting in the underestimation of
104
the overall manufacturing costs. This explains why the estimated costs shown above are
lower than the actual costs. In addition, capacity hours are used to determine the
overhead rates instead of the actual manufacturing hours. This means, if the production
system remains under capacity for a particular year, the overhead rates are under
estimated for the following year and the same is true for the overall manufacturing
costs. Conversely, if the production works over capacity, overestimation may be
observed. Since the production for the observed period of four years remained under
capacity, underestimation is observed in all the cases presented in the study. To refine
the proposed method, the use of normal capacity hours during TRO rate estimation can
be replaced by the actual manufacturing hours followed by an adjustment to account for
inflation in the following year. If the actual material cost for a product is higher than
the estimated one, the actual MRO for the product would also be higher than the
estimated ones resulting again in an underestimation of the product cost. This is
especially true when the actual wastage on the shop floor exceeds the estimated one.
The case-based approach developed by Niazi et al. [110] could prove a useful tool for
more accurate material cost estimation.
The contribution of each of the three cost elements to the total manufacturing costs
company wide were also analysed in order to ascertain the manufacturing cost
breakdown statistics needed to better understand an effective CCS. Figure 4-3 shows the
breakdown of the overall manufacturing cost for DT and PT in a diagrammatic form and
illustrates the variation in the values of each cost element over the four-year period. The
peak values represent the total manufacturing costs for each year which can also reflect
the variation in production output over the four-year period examined. For example, the
lowest peak value observed in years 2003-04 for PT reflects the lowest level of
105
production during the four-year period. It can be seen clearly that direct labour cost
contributes the least to the total manufacturing cost (owing to the geographical location
of the company) whereas material cost is the most prominent cost element. The
overheads values shown in Figure 4-3 include both TRO and MRO.
Figure 4-3: Cost element values in DT and PT
The values can be expressed in percentages. For example, the material cost accounts for
about 65-85 percent of the total manufacturing cost whereas, labour cost is an
insignificant, almost negligible cost contributor (1-2 percent of the total cost). Indirect
106
labour and indirect material costs are included in overheads. Overheads that combined
both TRO and MRO varied between 13 and 33 percent.
The overheads breakdown showing MRO and TRO is plotted in Figure 4-4 for both DT
and PT over the four-year period investigated. The TRO turns out to be the dominant
component out of the two. One of the reasons is that the material handling costs that are
part of MRO are significantly lower than the workstation running costs that come under
TRO.
Figure 4-4: TRO and MRO values breakdown for DT and PT
Since the labour cost is insignificant, a comparison is only made between material costs
and overheads in the form of cost trends over the four-year period. Figure 4-5 (a)
illustrates the cost trends for DT. It can be seen that the total overheads kept on
increasing despite a decline in material consumption until 2002-2003. However, this
increase reflects an increase in TRO rather than MRO. The decline in material
107
consumption indicates the decline in productivity. The lower productivity, thus, resulted
in increased overhead rates due to the allocation of company wide overheads
expenditure towards the under capacity manufacture of transformers. Another factor
that resulted in increased overheads was an inefficient cost control during the low
productivity period.
It can be seen that MRO also increased but their effect on the total overheads was
minimal. This was attributed to the fact that energy and other utility costs which are
elements of TRO jumped up during that period. Figure 4-5 (b) shows a similar trend for
PT. However, the TRO does not exhibit a steep increase as in DT despite a much
sharper decline in material consumption compared to DT. This is because the lead times
for the PT transformers are much higher than those for DT and so are the TRO. The
variation in production levels, therefore, has more profound effect on TRO in PT than in
DT. As a result, the decline in production in PT led to low figures for TRO.
108
Cost trend over time (DT) (a)
600
Cost Figure (Rupees in million)

500
400
300
200
100
0
1999-00 2000-01 2002-03 2003-04
Year
time related overheads material related overheads

Total Overheads Parts and material cost
Cost trend over time (PT) (b)
300
Cost Figures (Rupees in million)
200
100
0
1999-00 2000-01 2002-03 2003-04
Year
time related overheads material related overheads

Total overheads Parts and material cost
Figure 4-5: Cost trends in (a) DT and (b) PT
109
Figure 4-6 compares the cost trends for both DT and PT by considering material costs
against overheads. Because of the reasons discussed earlier, more irregularities are
found in DT than in PT.
Material Cost against Overheads
600
Cost Value (Rupees in million)
500
400
300
200
100
0
1999-00 2000-01 2002-03 2003-04
Year
Material Cost in DT Overheads in DT

Material Cost in PT Overheads in PT
Figure 4-6: Cost trends comparison in DT and PT
The cost element breakdown analysis and the discussion presented above highlight
some key points that are useful for establishing an effective CCS. This is particularly
relevant where manufacturing cost is calculated using material cost, direct labour costs
and overheads. These points are listed below:
• Estimating direct material cost of a product accurately is vital as this is the
most dominant cost element with a significant impact on the overall
manufacturing cost
110
• Material cost of a product can be estimated precisely using an effective
estimation methodology such as the case-based approach [110]
• Material cost is directly linked with the product design; however,
minimising wastage on the shop floor is an important aspect of an effective
CCS because it not only reduces direct material costs but also minimizes
overheads
• Direct labour cost is affected by factors like MLT, wage rate, geographical
locations, etc.
• Unlike material costs, overheads are not an integral part of product design
and hence play an important role in controlling the overall product cost
• Controlling overheads during low productivity periods is especially crucial
to minimizing product costs
• Identifying overheads correctly and adopting an effective estimation
methodology can significantly alter product cost estimates
• Determining overhead rates based on material and time consumed to
manufacture a product and using the rates to estimate overheads for a new
product is an easy, effective and accurate method to produce product cost
estimates and is applicable in the early stages of the design and
development process.
111
4.5 Conclusions
This chapter laid foundation for developing a PCE modelling methodology by
establishing an overhead estimation method. In order to predict the entire product cost
instead of only part or a component cost, breakdown approach was considered. The
chapter considered the existing breakdown of a product cost from a published model in
order to portray the potential areas of importance. This led to the identification of
overheads as an area worthy of exploration. Problems with the existing methods of
overhead estimation were then identified. Some of the existing methods that partially
resolved the problems were also discussed. However, this led to the identification of
possible solutions for estimating overheads more accurately. It was decided to validate
the developed model for overhead estimation through industrial trial. A batch
manufacturing set up with the estimation methodology based on a breakdown approach
was chosen for the implementation and validation purposes.
The manufacturing costs in an electrical engineering company in South Asia were
systematically analysed with emphasis on the estimation of overheads which contribute
about a third of total manufacturing costs. The analysis identified problems with the
previous practise of estimating overheads and a new methodology was proposed. The
new methodology was retrospectively implemented for a four-year period in the
company and a cost analysis was carried out. The estimates obtained with the previously
employed and the proposed methodologies were compared against the actual costs. It
was found that the proposed methodology leads to more accurate manufacturing cost
estimation results.
112
The study identified irregularities in the cost trends of the selected products over the
period of observations, especially during low productivity periods. Recommendations
were made to improve the proposed methodology further such as the use of actual
manufacturing hours instead of normal capacity hours during TRO rate estimation
followed by an adjustment for inflation in the following year. Job shops and batch
manufacturing set up based on flexible manufacturing and MRP based systems are some
of the examples of potential candidates for the proposed methodology.
The study also revealed that material cost accounts for 75 percent, labour costs for only
2 percent and overheads for almost 23 percent of the overall manufacturing costs at the
company. This cost breakdown, however, is representative of that region characterized
by lower wage rates. Thus, extensive testing of the proposed methodology in other
geographical locations is ongoing in order to further fine-tune it and make it an integral
part of an efficient CCS.
113
Chapter 5 PCE Hybrid Model
The aim of the Chapter 5 is to develop mathematical models for accurate product cost
estimation with a focus on batch manufacturing systems. The modelling methodology
presented is a hybrid approach consisting of a breakdown technique and an activity-based
costing system and focuses on an effective utilization of cost data obtained at the end of a
given period to predict a product’s cost for the following phase. The modelling framework
proposed is based on time- and material-dependent cost elements.
5.1 Introduction
Chapter 5 is aimed at modelling the overall product cost from the foundation already
provided in the previous chapter. It was established that the development of a hybrid
system for PCE in a batch type manufacturing environment is a viable research option in
the area. Breakdown approach was then selected as one of the hybrid element. The other
elements from the classification system for the hybrid approach will now be identified in
this chapter. However, since, the breakdown approach was selected and a complete
overhead analysis was already carried out in the previous chapter, the focus is first given
here to establish the breakdown elements for the proposed model. This would require an in-
depth analysis of what constitutes a product’s cost.
114
Chapter 5: PCE Hybrid Model
Manufacturing is a transformation process that consumes material and resources. The
degree of consumption of these resources is reflected in a product’s cost. The origin of the
consumed resources and those of the factors influencing a product’s cost may not only be
traced to the shop floor but outside it. The emerging picture of a product’s cost is, therefore,
a complex blend of distinct and indistinct contributions made to the product. The
contributions or the cost elements are often easier to be allocated to individual products
once they are produced. However, predicting these cost contributions even beforehand
requires efficient and accurate predictive tools.
The ever-increasing pressure on estimators to predict the cost early (sometimes even before
the conceptual phase) and accurately is also accentuating the need to develop innovative
techniques combining knowledge, experience, resources and historical data. Often the
techniques are customized to suit the needs of a system and/or a product. Whether, the
methodology is customized or generic, the final product cost determines selling price. An
effective cost estimation methodology, therefore, bears a potential of setting competitive
price without compromising profits yet securing commercial advantage to the enterprise.
In a batch manufacturing environment, where the available facilities are effectively utilized
to carry out a wide range of operations, estimating manufacturing costs accurately requires
an in-depth analysis of the shop floor resources and the way they are consumed. For this
reason, determining manufacturing cost in addition to the overall product cost does not
often receive enough attention. The estimation techniques available, although predict the
overall product cost, fail to yield accurate estimates of the manufacturing cost component.
115
An opportunity to make any optimization into the same or other components may, thus, be
lost. The chapter presents a comprehensive model for the estimation of manufacturing cost
and its sub-elements in addition to the overall product cost. However, before developing the
mathematical models for PCE, an analysis of the product cost and modelling methodology
is drawn.
5.2 Product cost and modelling approach
The previous chapter considered manufacturing cost as a combination of direct material,
direct labour and overheads (also termed as indirect manufacturing costs). Manufacturing
cost, on the other hand, is a sub-element of the overall product cost which is a combination
of engineering costs (costs incurred during design and development phases starting with
customer requirements, design of the part, its process planning and prototyping),
manufacturing costs and production overheads. Manufacturing cost is the largest element of
the overall product cost and is largely determined during the product design phase. This
element accounts for all the costs associated with the resources that can be traced to the
manufacturing floor where the actual production takes place.
Overall product cost can also be divided into direct and indirect costs incurred on a product
during its transformation from raw material to finished entity. Direct costs refer to the cost
elements that can be directly traced to a product and are generally incurred for the benefit
of the individual product. For example, direct material costs incurred on a product refer to
116
the costs incurred on material that is integral to the product design. Similarly, direct labour
costs refer to the costs incurred on labour spent purely on manufacturing the product and is
easily determined using lead times obtainable from process planning details. Direct material
and labour are the elements of manufacturing cost. Indirect costs refer to all the production
support costs incurred on or off the shop floor and generally for the shared benefit of
several products within a manufacturing facility. These refer to all the costs other than
those included in direct costs. These include manufacturing overheads, production
overheads and engineering costs.
From the analysis presented in the previous chapter, it was found that manufacturing
overheads (part of indirect costs) can be divided into time– and material–dependent cost
elements. However, after the implementation and validation analysis, areas for further
optimization were discovered. It was found out that a further two classes of manufacturing
overheads can be identified. Some of the indirect costs depend on both processing time and
material quantities to be produced. Cost spent on tooling, for example, depends on both
processing time and the amount of material to be processed and can not be grouped with
either time– or material–dependent cost groups. Some of the other indirect costs, on the
other hand, are spent on keeping the manufacturing shop floor in a running condition. A
realistic approach to allocate such costs on different products is required that is based on
the product’s level of consumption of resources from the manufacturing shop floor. If a
product stays for a longer period of time and occupies a larger space on the manufacturing
floor, it is highly likely to consume more resources than a product that stays for a shorter
duration and occupies little space. Such indirect costs can be termed as building space cost
117
and are, thus, based on lead times and occupied spaces for individual products. Yet another
set of indirect costs are those that can not be directly traced to a manufacturing shop floor
or only partially. For example, costs incurred on building security, general administration
costs and financing expenses etc. can be grouped together. These costs need to be allocated
to individual products using a pragmatic approach. It is highly likely that a product with
higher manufacturing costs should incur these costs in greater proportion than those with
lower manufacturing costs. It is because a product with a higher manufacturing cost
consumes more resources on the manufacturing shop floor and is thus likely to incur higher
proportions of other costs also. Such indirect costs can, therefore, be allocated to individual
products in proportion to their manufacturing costs. These indirect costs refer to the only
set of overheads not directly traceable to the manufacturing shop floor and can be termed as
production overheads. The other four indirect costs can be traced to the manufacturing shop
floor and can be termed indirect manufacturing costs. These include: time–dependent cost
(also referred to as processing cost), material–dependent cost, tooling cost and building
space cost. Engineering cost is also a form of indirect cost and refers to the costs incurred
during design and development phases starting with customer requirements, designing of
the part, its process planning and prototyping.
Based on the breakdown approach, product cost can now be expressed as a summation of
the individual cost elements (manufacturing cost, engineering cost and production
overheads). A pictorial representation of the detailed product cost breakdown followed in
the present study is shown in Figure 5-1 together with the cost distribution for a typical
product [109]. The estimated cost for a new product, C pn+1 , can thus be expressed as the
118
n +1
sum of the three cost elements: estimated manufacturing cost , CGp , estimated engineering
n +1
cost, C Ep , and estimated overheads, O pn +1 :
C pn +1 = C Gp
n +1 n +1
+ C Ep + O pn+1 (5-1)
Selling Price
Selling Price
20%
25%
Product Cost Profit
15%
40%
Manufacturing Production
Engineering
Cost Overheads
Cost Manufacturing Cost
38%
Direct Indirect 50%
Material Direct
Labour Manufacturing
Cost Cost
12%
Processing Material – Dependant Tooling Building

Cost Cost Cost Cost
Figure 5-1: Pictorial representation of the mathematical model for product cost estimation
Manufacturing costs refer to the costs associated with the activities taking place on the
manufacturing shop floor and the cost elements that can be traced to it. The manufacturing
119
n +1
cost, CGp , for an individual product, p, can be estimated by summing the direct cost
n +1 n +1
elements (material cost, Cmp , and labour cost, C Lp ) and the indirect manufacturing cost
n +1 n +1
elements (processing cost, Ctdp , material-dependent cost, Cmdp , tooling cost, CTpn+1 , and
n +1
building cost, C Bp ) for the product, i.e.:
n +1 n +1 n +1 n +1 n +1
C Gp = C mp + C Lp + C tdp + C mdp + CTpn +1 + C Bp
n +1
(5-2)
Determining indirect costs accurately and early for an individual product in a
manufacturing environment where resources are shared between several different products
is challenging and requires sound engineering knowledge of product design, planning,
production and costing. A traditional approach to estimating such costs relies on summing
up all the expenses other than those included in the direct costs and ascertaining a cost rate.
The rate is then used to estimate the indirect costs for a new product by multiplying it with
the expected manufacturing lead time (MLT) of the new product. However, a major
drawback of this methodology is a disproportionate allocation of indirect costs to different
products. Furthermore, since the indirect costs recovery considers only MLT, the amount of
material required for the manufacture of a product is ignored. This often results in cost
underestimation for products with higher material requirements and overestimation for
those with lower material quantities.
The ABC system was presented [93] as a useful means to distribute the indirect costs in
proportion to the activities performed on a product during manufacturing. The system
120
proved a good alternative to the traditional estimation techniques since it provided more
accurate manufacturing cost estimates [95]. However, the effectiveness of the methodology
depends on the way the activity rates are determined and the accuracy of the estimated
activity units. Moreover, time-based activity units are generally employed ignoring material
related costs during PCE using the ABC system. For example, the ABC model developed
by Aderoba [48] only considered activity times ignoring material quantities for the cost
estimation of a new product. Material dependent costs such as transportation, packaging,
storage and inventory, etc. may, therefore, not be accurately estimated for the new product.
The model presented in the current chapter overcomes the limitations of the existing
methodologies, as mentioned above, by providing a framework for PCE based on time– and
material–dependent cost elements. The modelling methodology is a hybrid approach
combining a breakdown model (including manufacturing costs, engineering costs and
production overheads) with a modified ABC system and skilfully exploits the already
validated models in Chapter 4. The above mentioned limitations associated with
determining activity rates and activity units using the conventional ABC system are
overcome based on an effective utilization of the cost data obtained at the end of a given
year ‘n’ to predict a product’s cost in the beginning of the following year ‘(n+1)’ by
determining the rates for cost elements. It means that the developed method incorporates
the attributes of a case-based system into the ABC system by making use of past data. The
modified ABC system, therefore, forms second element of the Hybrid Model. Figure 5-2 is
a pictorial representation of the overall modelling framework within the context of the
developed technique classification system. It is evident that the developed model is also
121
hybrid at the top level of the classification system by making use of the attributes of the
qualitative and the quantitative techniques. The breakdown approach helps to estimate the
overall product cost. The ABC takes care of the accuracy of the results and the use of the
case-based approach helps to produce early results. Although, at the lower levels of the
classification system, three elements appear to contribute towards the proposed model, the
modified ABC system and the breakdown elements are the two elements of the Hybrid
Model. The modified ABC component is itself hybrid in nature. The developed model is a
comprehensive and an integrated costing tool that not only estimates the overall product
cost but predicts the essential individual cost elements. A low to medium volume batch
type production environment where various indirect costs are involved is a highly suitable
environment for the application of the proposed methodology.
5.3 Direct cost elements
Direct costs refer to the costs associated with the elements directly identifiable to and
linked with a product and are generally incurred for its exclusive benefit. Costs associated
with material and labour are largely considered direct manufacturing costs. These are
examined in more detail below.
122
5.3.1 Direct material costs
The material costs can be estimated using the material quantities obtained directly from the
BOM developed on the basis of the product structure and engineering design details. A
scrap margin is usually accounted for in BOMs based on past trends. In a manufacturing
environment where products are standardized, their engineering designs and BOMs are also
established. Even when a non-standard product is designed and developed, the required
quantities of different raw materials are similar to those in a closest standard match even
though the technical designs of the two may be different. Thus, the closest match provides a
good starting point to estimate material cost for the non-standard product. As a result,
material cost estimation is possible in the early stages of the design and development stage.
n
Under such circumstances, the material cost calculation, based on past product cost, Cmp ,
(already stored in a database) and certain commercial indicators termed as material cost
deviation index (MCDI), φ n +1 , (such as inflation rate, foreign exchange rate, price increase
etc.) for the following year, becomes a standard practice. i.e.,
n +1
C mp = (1 + φ n +1 ) × C mp
n
(5-3)
Past product costs can be established using the unit cost, Ckn , of the kth material in the
product and its corresponding quantity, mk , i.e.
123
n
C mp = ∑ (C kn )(mk ) (5-4)
k
n
The total cost based on the cumulative material requirement, Cmt , for manufacturing, p,
products for a given period of time changes in accordance with the number of units
produced, N pn , for a given product. Thus, the factory-wide changes in material
consumption over different time intervals can be used to reflect changes in production
levels and can be obtained by the following equation.
C mtn = ∑ (C mp
n
)( N pn ) (5-5)
p
The total planned cost, Cmtn+1 , based on cumulative material consumption can be obtained
from the planned orders, N pn+1 .i.e.
C mtn+1 = ∑ (C mp
n +1
)( N pn +1 ) (5-6)
p
124
Product Cost Estimation
Techniques
Intuitive Techniques Analogical Techniques Parametric Techniques Analytical Techniques
Regression Analysis Back-Propagation Neural

Model Network Model
Decision Support
Techniques Operation-Based
Breakdown
Approach
Approach
Feature-Based Cost
Case-Based Estimation
Technique
Expert System
Rule-Based System Tolerance- Activity-Based
Based Cost Cost Estimation
Fuzzy Logic System Models
Hybrid Model for PCE
Figure 5-2: Development of the Hybrid Model within the framework of the technique classification system
125
5.3.2 Direct labour
The manufacturing shop floor layout consists of work centres which can be either labour-
intensive or machine-intensive centres or both. Fabrication, manual assembly and painting
are examples of labour-intensive centres sometimes referred to as labour centres also. CNC
machines, automatic assembly lines, industrial robots are examples of machine-intensive
centres. A combination of the two is also common and can be termed as hybrid centres;
examples include lathes, mills, semiautomatic assembly lines, etc. More commonly the
term machine centre is also used to refer to either machine-intensive or hybrid centres i.e.
any work centre that is not a labour centre can also be termed a machine centre. Figure 5-3
illustrates the three different work centres.
Work Centre
Labour Centre/ Machine Centre

Labour-intensive
centre
Fabrication,
manual assembly,
painting etc. Machine-intensive Hybrid centre
centre Lathes, mills,
CNC machines, automatic semiautomatic
assembly lines, industrial assembly lines
robots etc. etc.
Figure 5-3: Types of work centres
126
The direct labour cost incurred on a product accounts for the fraction of the wages of the
workers in the work stations through which the product is routed. The time spent by the
workers to manufacture individual products and their respective wage rates (generally set
on the basis of their skill levels) can be used to determine this cost. However, in a batch
manufacturing environment where multi-skilled workers are placed on work centres
intermittently and handle multiple tasks, a more pragmatic approach is required to ascertain
these costs.
The direct labour in the proposed methodology is based on the shop floor wide aggregate
labour rate and the labour units spent to manufacture a product. Thus, the direct labour cost
n
already incurred on an individual product, C Lp , can be obtained as a product of the actual
n n
labour rate, RLA , and the labour units consumed, U Lp , to manufacture it, i.e.
n
C Lp = ( RLA
n
) × (U Lp
n
) (5-7)
n +1
The estimated labour cost, C Lp , can be given as a product of the estimated labour rate,
n +1 n +1
RLE for a given period and the estimated labour units, U Lp , for a given product.
n +1 n +1 n +1
CLp = ( RLE ) × (U Lp ) (5-8)
127
Total planned labour units, U Ltn+1 , can be used to estimate the total labour cost, C Ltn+1 , for
planned orders, i.e.
C Ltn +1 = ( R LE
n +1
) × (U Ltn +1 ) (5-9)
Labour units
The MLT of a product is the summation of the individual lead times in the relevant work
centres. The lead time for a job on an individual work centre (sometimes referred to as
work centre cycle time WCCT) is the time spent by the job on the work centre before
moving on to the next work centre. The lead time is different from the man-units consumed
which are referred to as the total time spent by all the workers on the job in that work
centre. The man-units on an individual work centre can be converted into labour units by
considering the output levels (skilled, semi-skilled and non-skilled levels) of the individual
workers. The total time spent by skilled labour, tx, semi-skilled labour, ty and non-skilled
labour, tz, can be expressed in a unified form of labour units, Lnjp , by applying skill indices,
α, for semi-skilled labour with ranging from 0.4 to 0.8 and, β, for non-skilled labour with
ranging from 0.25 to 0.4.
Lnjp = ∑ t x + α ∑ t y + β ∑ t z (5-10)
x y z
128
The values for skill indices can be set based on feedback form shop floor supervisors. For
example, the value of α = 0.6 can be set for a semi-skilled labour whose skill or output
level is almost 60 percent of a standard skilled-labour. Wages of a worker on the shop floor
may also be a criterion to set the indices as variations in wages of the workers can be in
accordance with their skill or output levels.
This approach is particularly helpful in a computer integrated manufacturing (CIM)
environment where time cards are filled at individual work centres with job codes,
employee numbers, etc. Based on the relative skill levels of semi- and non-skilled labour
compared to those of the skilled ones, line supervisors are generally in a good position to
ascertain skill indices for individual workers. If an individual lead time, t njp , is taken as the
time spent by all the workers working on a job in a work station, then the number of
skilled, x, semi-skilled, y, and non-skilled, z, labour can be used to obtain labour units on
the work centre, using the formula:
Lnjp = t njp ( x + αy + βz ) (5-11)
The total labour units consumed for the product can then be given as:
n
U Lp = ∑ Lnjp (5-12)
j
129
This in turn allows the estimation of the shop floor wide labour units consumed, U Ltn , for a
given period of time based on the number of different products produced during that period:
U Ltn = ∑ (U Lp
n
)( N pn ) (5-13)
p
The determination of labour units for a product, therefore, requires lead times at individual
work centres and number of different workers with their skill levels working on the
product. The lead times can be obtained from standard routings and are established either
by using standard formulas or by deploying time and motion studies. For example, a range
of time estimation models and time standards can be found in [111]. In a CIM environment,
a feedback mechanism ensures that the actual lead times are standardised for future use.
The lead time for a non-standard or a new product can be established from the results for
the closest standard match and incorporating the changes in consultation with the shop floor
supervisors and planning engineers. The labour units required to manufacture the product
can then be estimated based on the number of operators required on individual work
centres.
Labour rate
The shop floor-wide direct labour cost incurred during a given period, C Ltn , reflects the
wages paid to the shop floor workers during that period and is the summation of all the total
wages in the individual work centres, Gtjn , i.e.:
130
C Ltn = ∑ Gtjn (5-14)

j
The total wages and the total labour units consumed in a given period can be combined to
calculate the actual labour rate, i.e.:
∑G n
tj
n
RLA = j
(5-15)
U Ltn
The labour rate calculated in this way differs from the conventional wage rate for a given
worker. The wage rate for a particular period of time is a matter of organizational policy
and is influenced by various factors (such as minimum wage rate set by the government,
geo-political conditions, economic growth, and an individual worker’s skill level etc.). The
actual labour rate during a period n can be used to ascertain an estimated labour rate for
period n+1 based on the expected variance called labour cost deviation index (LCDI), ε n +1 ,
(such as the effects of inflation, forecasted wage differential, etc.):
n +1
RLE = (1 + ε n +1 ) × RLA
n
(5-16)
131
5.4 Indirect cost elements
These are costs associated with contributions that can be traced to the manufacturing shop
floor but not to a specific product or only partially. The costs incurred on indirect material
(material shared for the benefit of many different products), utilities, machine repair and
maintenance, quality control, tooling and equipment, building space, etc. comprise indirect
costs. The indirect cost elements for an individual product can be determined by adopting a
suitable methodology for apportioning the total indirect costs based on the product’s degree
of consumption of manufacturing activities and resources. The study identifies four kinds
of indirect costs that can be linked with manufacturing shop floor and presents
comprehensive models to estimate them for a specific product.
5.4.1 Processing cost
The processing costs incurred during a given period on the manufacturing shop floor, Ctdn ,
refer to all the processing-time-dependent costs (excluding labour costs) that contribute to
the smooth running of the machine centres on the shop floor. Thus, this cost can be
expressed as a summation of all the individual time-dependent cost elements, Cdn , (such as
utility cost, maintenance cost, repair cost, machine depreciation, machine insurance etc.) for
a given period:
Ctdn = ∑ Cdn (5-17)

d
132
n +1
The processing cost for an individual product, Ctdp , can be determined from the estimated
n +1 n +1
processing rate, RME , and processing units for the product, U Mp :
n +1 n +1 n +1
C tdp = ( RME ) × (U Mp ) (5-18)
Processing units
The job processing time, tipn , on the, ith, machine centre (time spent by a job between its
arrival and departure from a machine centre while the machine is running) can be used to
determine processing units consumed by the job, M ipn , on the centre by applying a machine
index, ηi .
M ipn = η i × t ipn (5-19)
The machine index accounts for repair, maintenance, utilities etc. and takes values ranging
from 1.25 to 2.0 [9]. Older machines requiring more maintenance, repair, etc. will have a
higher index value. The job processing time is generally the same as lead time and can be
obtained directly from time cards in a CIM environment. By summing up all the processing
n
units at individual machine centres, the total processing units for the product, U Mp , can be
obtained:
133
n
U Mp = ∑ M ipn (5-20)
i
n
The shop floor-wide processing units consumed, U Mt , for a given period of time can be
calculated based on the number of different products produced during that period:
n
U Mt = ∑ (U Mp
n
)( N pn ) (5-21)
p
The job processing times for a new product can be estimated using closest standard matches
allowing the estimation of the processing units for the new product.
Processing rate
n
A shop floor-wide aggregate processing rate, RMA , is proposed from the processing cost
data collected during a given period and the corresponding processing units consumed:
C tdn
R n
MA = n (5-22)
U Mt
However, the same methodology can be used to establish separate rates for individual
machine centres if their corresponding processing costs can be traced and cumulative
processing units obtained easily. The estimated rate for the following period can be set
134
based on a processing cost deviation index (PCDI), µ n +1 , that reflects the effects of
inflation, utility costs variations, maintenance and insurance forecasts, etc.:
n +1
RME = (1 + µ n+1 ) × RMA
n
(5-23)
5.4.2 Material dependent cost
Several indirect costs incurred in a manufacturing enterprise can be attributed to material
quantities used to manufacture different products, called material-dependent costs (MDC).
Examples of such costs include: indirect material, purchasing, stores & inventory, freight &
transportation, material inspection, packaging, and quality costs, etc. These costs do not
depend on the lead times or processing times but on the quantity of materials consumed.
Thus when the direct material quantities go up, MDC also go up. As a result, MDC is also a
significant contributor to the manufacturing costs. The total material-dependent costs for a
n
given period, Cmd , can be expressed as:
 
n
C md =  ∑ C wn  − S n (5-24)
 w 
Where, C wn refers to the total cost incurred on an individual MDC element (such as
packaging) and S n is the salvage value for scrap material. Since the MDC increase with the
rise in direct material quantities, the fraction based on the total MDC and the cumulative
135
direct material cost for a given period can be used to estimate the MDC for a specific
n +1 n +1
product, Cmdp , using its estimated direct material cost, Cmp and MDC deviation index
(MDCDI), ρ n+1 (the effect of inflation, variations in material, freight and transportation
prices etc.):
 Cn 
n +1
C mdp =  mdn
 × (1 + ρ n+1 ) × C mp
n +1
(5-25)
 C mt 
5.4.3 Tooling cost
Tools range from manual equipments to jigs, fixtures, moulds, dies etc. The total tooling
cost incurred on the manufacturing floor for a given period, CTtn , comprises the cost
n
incurred by tools utilization and replenishment in machine centres, CMT , and labour
n
centres, CLT .
CTtn = CMT
n
+ CLT
n
(5-26)
Tooling cost depends not only on the amount of material processed on the manufacturing
floor but also on the job processing times on machine and/or labour centres. The proposed
methodology, therefore, allocates the total tooling cost incurred on the shop floor to
individual products on the basis of their material costs and corresponding processing and/or
labour units. Tooling cost rates in labour and machine centres for a given period n are
136
determined for this purpose. Estimated tooling cost rates for the following period n+1
n +1 n +1
based on machine centres, RMT , and labour centres, RLT , can then be effectively utilized to
predict the tooling cost for an individual product, CTpn+1 , i.e.:
(
CTpn +1 = C mp ) [(
n +1 n +1
× RMT )(
n +1
U Mp ) (
n +1
+ RLT )( n +1
U Lp )] (5-27)
Determining the tooling cost for an individual product based on an aggregate shop floor-
wide tooling cost rate often results in disproportionate allocation. This is due to the cost
variances between the tools at machine centres and labour centres and the different time
spent by the product on the respective centres. The proposed model overcomes this problem
by providing separate rates for tools used at machine and labour centres respectively.
Machine tool rate
n
The machine tool rate, RMT , is associated with the total tooling cost incurred in machine
n
centres, C MT , during a given period. It is given by the formula:
n
(CMT )
n
RMT = (5-28)
(U Mt ) × (Cmtn )
n
The total machine tool cost for a given period can be expressed as the sum of total
depreciation, DMn , of M’ number of machine tools during that period. The costs associated
137
with number of b machine tools that are broken down can be incorporated by considering
their initial purchase prices, Pb , and respective depreciated values, Db .
   
n
CMT =  ∑ DMn  +  ∑ ( Pb − Db )  (5-29)
M   b 
Standard techniques can be applied to ascertain the depreciation values for individual tools.
The estimated machine tool rate for the following period can then be set based on a
machine tool cost deviation index (MTCDI), ψ n +1 , to account for the effects of inflation,
expected variation in tool utilization, life expectancy factor of the current tools, etc.:
n +1
RMT = (1 + ψ n +1 ) × RMT
n
(5-30)
Labour tool rate
n n
The labour tool rate, RLT , considers the total tooling costs incurred in labour centres, C LT ,
for a given period. It is determined using the formula:
n
(C LT )
n
R LT = (5-31)
(U Lt ) × (C mtn )
n
The total labour tool cost for a given period can be expressed as the sum of total
depreciation, D Ln , of L number of labour tools during that period. Similarly to the machine
138
tools, the costs linked with number of a broken labour tools with their initial purchase
prices, Pa , and respective depreciated values, Da , can be incorporated as follows:
   
n
CLT =  ∑ DLn  +  ∑ ( Pa − Da )  (5-32)
 L   a 
A labour tool cost deviation index (LTCDI), σ n +1 , is then employed to account for
inflation, expected variation in tool utilization, life expectancy factor of the current tools,
etc. and thus the labour tool rate for the following period will be:
n +1
RLT = (1 + σ n +1 ) × RLT
n
(5-33)
5.4.4 Building space cost
The building space cost includes all the essential costs incurred to keep the manufacturing
shop floor and the overall plant in a usable condition. Plant depreciation, building
insurance, maintenance, repair, and utilities (excluding those supplied to the shop floor) are
some of the examples of building space cost elements. The number of h building space cost
elements, Chn for a given period can be added together to determine the total building space
cost, C Btn , for that period.
C Btn = ∑ C hn (5-34)
h
139
The proposed methodology allocates the total building space cost to individual products
based on the areas they occupy on the manufacturing shop floor and their lead times.
Products requiring bigger spaces and higher lead times incur larger proportion of building
space costs. Building space rate for a given period is determined for this purpose. An
estimated building space rate, RBn +1 , can then be set for the following period. This allows
the estimation of the building space cost for an individual product by multiplying the
n +1
estimated rate with the manufacturing space, SGp , and units (processing and labour) for the
product, i.e.:
n +1
C Bp ( ) (
= RBn +1 × S Gp
n +1
) (
n +1
× U Mp n +1
+ U Lp ) (5-35)
Building space rate
Building space rate for a given period, RBn , can be set by dividing the total building space
n
cost incurred during that period from the total manufacturing units (processing, U Mt , and
labour, U Ltn ) and the total area of the manufacturing shop floor, S Gtn .
C Btn
R Bn = (5-36)
n
(U Mt + U Ltn ) × S Gtn
140
Again a building cost deviation index (BCDI), δ n +1 , can be defined to account for any
variations due to inflation, forecast variations for building maintenance, repair, utilities etc.
and the final adjusted rate will be:
RBn +1 = (1 + δ n +1 ) × RBn (5-37)
Manufacturing space
Manufacturing space here refers to the overall area of the manufacturing shop floor and is a
combination of the space occupied by all work centres, S otn , and the total unoccupied space
on the floor, Sutn .
SGtn = S otn + Sutn (5-38)
Shop floor cell layout records can be used to find out occupied spaces by individual work
centres, O nj , that can be summed together to give the total occupied space.
S otn = ∑ O nj (5-39)
j
The ratio of the total manufacturing space to the total space occupied can be effectively
n +1
used to allocate the manufacturing space for an individual product, SGp .
141
n +1  Sn 
S Gp = S opn +1  Gtn  (5-40)
 S ot 
Where, S opn +1 , refers to the total occupied space for the product and can easily be obtained
by summing up the number of r different work centre spaces through which the product is
routed, Orpn +1 .
Sopn +1 = ∑ Orpn +1 (5-41)

r
5.5 Production overheads
Overheads here refer to the production overheads (PO) and are different from indirect
manufacturing costs. These include costs incurred on elements like security services,
computer software, general administration, financing, sales and marketing, etc. They are
not normally traced to manufacturing shop floor. The total PO, Otn , can be found by adding
the number of q different PO cost elements, H qn , i.e.:
Otn = ∑ H qn (5-42)
q
142
Since, the POs are incurred to maintain manufacturing activities; they can be effectively
allocated to individual products as fractions of their manufacturing costs. The fraction
calculated based on the total PO and the total manufacturing cost, CGtn , incurred during a
given period can, therefore, be used to estimate the POs for an individual product for the
following period, O pn +1 , using its estimated manufacturing cost, C Gp

n +1
, and PO deviation
index (PODI), τ n+1 (the effect of inflation, variations in selling expenses, general and
administration costs etc.):
 On 
O pn +1 =  nt  × (1 + τ n +1 ) × C Gp
n +1
(5-43)
 C Gt 
The total manufacturing cost incurred for a given period is a summation of all the
individual total manufacturing cost elements, i.e. material, labour, time-dependent,
material-dependent, and building space costs.
CGtn = Cmt
n
+ CLtn + Ctdn + Cmd
n
+ CTtn + C Btn (5-44)
5.6 Conclusions
This chapter presented a comprehensive modelling methodology for PCE in a batch type
manufacturing environment. The overhead estimation method developed in the previous
chapter formed the basis for extending the modelling framework. The limitations associated
143
with the existing methods of overheads estimation resulted in defining sets of new indirect
costs which were modelled later on. The only element that could not be modelled was
engineering cost due to the insufficient experience, data and knowledge obtained from the
industrial domain. However, the nature and the extent of the theories developed for the
other elements and the clarity of the developed models contain the intrinsic characteristics
that can serve as guidelines to model the engineering cost.
The adopted modelling methodology is a hybrid approach combining the attributes of a
breakdown technique with the ABC system in order to estimate the total product cost. In
this respect, the model was presented as a summation of individual cost elements
(manufacturing cost, engineering cost and production overheads) which were broken down
further to their sub-elements. Activity rates and units were defined and modelled for
individual elements based on an effective utilization of historical cost data in order to
predict product costs early and accurately. Thus comprehensive models for estimating
manufacturing cost and indirect cost elements for individual products were given.
The proposed modelling framework is based on time- and material-dependent cost elements
and overcomes the current limitations in cost estimation. The model has been tested and
validated using data from a crane assembling unit and the results are presented and
discussed in the next chapter.
144
Chapter 6 Industrial
Implementation and Analysis of the PCE
Hybrid Model
The aim of Chapter 6 is to implement the developed PCE Hybrid Model in a batch
manufacturing environment in the UK. The process will focus on developing an
implementation algorithm, modelling cost deviation indices and obtaining the estimated
cost values from the model’s implementation for a given product range. The estimated
results will also be obtained from the company’s own method.
6.1 Introduction
Chapter 6 establishes basis for a thorough validation analysis for the developed PCE
Hybrid Model. This is achieved by the industrial implementation and application of the
Hybrid Model developed and presented in the previous chapter. When considering the
validation process, it is important to determine the criteria for validation and that if the
adopted criteria are widely acceptable and recognizable standards within the framework of
a specialism. Although, mathematical models can be simulated to validate the results, those
designed for industrial applications need to be backed up by industrial trials to best serve
145
Chapter 6: Industrial Implementation and Analysis of the PCE Hybrid Model
the purpose. Industrial trials can be time consuming requiring attention to details, the
results based on such attempts can only strengthen the confidence in the entire validation
process itself.
The developed model as part of the current research study was not only conceptualized with
the aim of theoretical development in the field but to serve the wider industrial needs also.
The criterion for its validation, therefore, is fittingly set to proceed with the industrial
implementation and application. The process involves implementing the model and
validating it by comparing the results with those generated by any existing system in place.
Setting the objectives for the validation process and defining how they will be achieved are
fundamental to the success of the entire process. Before any of the objectives can be set, the
active conditions have to be considered. Within the context, getting the right balance
between the duration of the trials and the consistency of the results generated during that
time is important. The trial period, therefore, is set for three years (2003 – 2005)
retrospective analysis requiring field data from 2001 to 2005 based on the requirement of
the developed model. The validation period of three years is a reasonable time to allow the
process to generate reliable and consistent results. Since the model is developed to mainly
fulfil a batch manufacturing system’s requirements, the selection of such an industrial
environment for the implementation and validation analysis is yet another active condition.
A careful consideration was given to the conditions to initiate the validation process.
146
The main objective of a PCE process is to predict the likely product cost early and
accurately. Accurate estimation refers to minimizing the difference between the actual costs
and the predicted values. The developed model should aim to predict product costs more
accurately than any of the existing systems in use at the selected company. Although cost
estimation has no direct influence on cost control, the estimated results should help the
decision-making process to devise cost control strategies. Cost estimation methods that can
predict not just the entire product cost but the elemental values can highlight the potential
areas for better cost control. The developed model should, therefore aim to present better
structured and more elaborate elemental results than those generated by the existing system.
Section 6.2 details the procedure and methodology for the industrial implementation of the
PCE Hybrid Model. A comprehensive algorithm is developed to facilitate the
implementation process. Section 6.3 details the PCE process at the company. The necessary
business information is presented followed by the cost estimation for the given product
range using the company’s own method of cost calculation. Section 6.4 deals with the
implementation of the PCE Hybrid Model in the selected company with the help of the
modelled indices and based on the developed algorithm in section 6.2 in order to ascertain
the estimated costs for the given product range for the duration of the trial period. Section
6.5 concludes the chapter.
147
6.2 HMI algorithm and the implementation methodology
The Hybrid Model Implementation (HMI) algorithm is developed and presented in the
section to facilitate the industrial implementation of the model. The overall aim of the
implementation of the Hybrid Model in return is to facilitate the entire validation process.
With this in mind, it is essential to compare the results generated by the implementation of
the Hybrid Model against those generated by the company’s own method. The company’s
method of cost estimation with the relevant results would, therefore be discussed to support
the implementation of the Hybrid Model. But first the necessary steps and the overall
procedure of the Hybrid Model’s implementation will be discussed.
The Hybrid Model is implemented in the selected manufacturing company for a
retrospective analysis of three years (2003 – 2005). The developed model makes use of the
available data at the end of a year (n) and the preceding year (n-1) to predict the costs for
the following year (n+1). Data obtained from 2001 to 2004 could therefore be used to
predict costs from 2003 to 2005. Since, the comparison of the estimated costs predicted at
the beginning of a year is made with the cost data obtained at the end of that year; cost data
for 2005 were also required in order to make comparison with the estimated costs for that
year. Therefore, the data obtained from 2001 to 2005 were used for the implementation and
validation analysis for (2003 – 2005). The estimated results will then be compared against
the actual product costs in the next chapter as part of the validation analysis. The current
chapter details the entire implementation process. Figure 6-1 to Figure 6-7 represents the
Hybrid Model Implementation (HMI) Algorithm.
148
The stepwise preliminary implementation of the Hybrid Model is presented in Figure 6-1. It
is clear that the entire PCE process for a product is outlined in 13 different steps starting
from material cost estimation to the estimation of a product’s overall cost. The figure cross
references relevant sections, figures and equations. The logical sequence of necessary steps
is shown with arrows. The dotted arrows follow sub-systems detailed in separate cross
referenced figures with step numbers.
149
Start
Step 1
Step 2 (refer to Figure 6-2)
Material Cost Estimation (Section 5.3.1) 1
Step 3 1A

Labour Cost Estimation (Section 5.3.2) 2
Step 5 2A

Processing Cost Estimation (Section 5.4.1) 3
Step 7 3A

MDC Estimation (Section 5.4.2) 4
4A
Step 9
Manufacturing Cost Estimation Step 10 (refer to Figure 6-6)

5
Equations (5-2), (5-44)
Step 11 5A

Production Overhead Estimation (Section 5.5) 6
6A
Step 13
Product Cost Estimation (Section 5.2)

Figure 5-1, Equation 5-1
Stop
Figure 6-1: PCE Hybrid Model implementation phase
150
Detailed implementation steps are facilitated with necessary links in the preliminary
implementation phase. For example material cost estimation process is presented with links
1 and 1A and is elaborated in Figure 6-2. The logical steps presented in rectangular boxes
are connected with arrows. Inputs to the system are presented in parallelograms together
with necessary references to the equations developed as part of the Hybrid Model. For
example, equation (5-4) can be used to calculate material cost for the ‘pth’ product in the
‘nth’ and the (n-1)th years using the input provided in the same parallelogram. The input in
this case refers to material cost data provided by the company from its financial accounts.
1
Material cost data for the nth
and (n-1)th years
Material cost calculation for pth
Eq (5-4)
product in the nth and (n-1)th years
Number of units produced in the nth

Cumulative material costs in the nth and (n-1)th years
and (n-1)th years
Eq (5-5)
MCDI calculation
Eq (B-5)
for the (n+1)th year
Estimated material cost for the

Eq (5-3)
pth product in the (n+1)th year
1A Refer to Figure 6-1 (step 3)
Figure 6-2: Material cost estimation implementation process
151
The inputs to the system are either primary or secondary. Primary inputs are the details
provided by the company. Secondary inputs to the system are the results obtained during
the implementation phase and are fed to the system to generate further results. For
example, during the calculation for PCDI mentioned in Figure 6-4, the inputs to equation
(B-9) are aggregate processing rates (secondary input) and inflation rates (primary input).
Aggregate processing rates are termed secondary inputs because they were calculated in the
preceding step using equation (5-22). It is, therefore, fitting to sequence the execution steps
in the HMI algorithm to maintain the logical flow and to effectively utilize the secondary
data. The algorithm is designed to facilitate the Hybrid Model implementation at the
selected company and as such adapts to the system. However, minor changes can make the
algorithm more generic. Some of the customizations are considered in light of the
requirements mentioned in section 6.4.
The HMI algorithm facilitates the generation of the results by effectively utilizing the
primary and secondary inputs and the PCE Hybrid Model. The results include not only the
estimated product and manufacturing costs but the elemental values including the estimated
material, labour, processing and material-dependent costs. In addition, the cumulative costs
and the deviation indices are effectively calculated. All the results generated through the
implementation of the HMI algorithm are discussed in detail in section 6.4.
152
Labour units (LU) calculation

for pth product in the nth year
Lead times at jth WC, number of
skilled, semi-skilled and non-skilled
LU calculation at jth work labour, labour indices
centre (WC)
Eq (5-10) or (5-11)
No
LU calculated
at all ‘j’ WC ?
Yes
Total LU calculation for pth product Eq (5-12)
No Total LU for all ‘p’ Number of units produced in

products calculated? the nth year for all ‘p’ products
Eq (5-13)
Yes
Shop floor-wide LU
calculation in the nth year
Total wages paid in the
nth year to all workers
Direct labour cost calculation in the nth year Eq (5-14)
Actual labour rate (LR) calculation for the nth year Eq (5-15)
LCDI calculation for skilled, semi-skilled Average per month wages of

skilled, semi-skilled and non-skilled
and non- skilled labour and then average
labour in the nth and (n-1)th years,
LCDI for the (n+1)th year inflation in the nth and (n+1)th years
Eqs (B-6), (B-7) and (B-8)
Estimated LR calculation for the (n+1)th year
Eq (5-16)
Estimated labour cost for pth product in the
(n+1)th year
Eq (5-8)
Figure 6-3: Labour cost estimation implementation process
153
Processing units (PU) calculation for

pth product in the nth year
PU calculation at ith machine Processing time at the ith

centre (MC) MC, machine index
Eq (5-19)
No PU calculated
at all ‘i’ MC?
Eq (5-20)
Yes
Total PU calculation for pth product

Number of units produced in the
nth year for all ‘p’ products
No Total PU for all ‘p’ Eq (5-21)
products calculated?
Yes
Financing expenses and insurance,
Shop floor-wide PU maintenance, repair and energy
calculation in the nth year costs incurred in the nth year
Eq (5-17)
th
Processing cost calculation in the n year
Eq (5-22)
Aggregate processing rate (PR) calculation
for the nth year
Aggregate processing rates for the

nth and (n-1)th years, inflation in
PCDI calculation for the (n+1)th year the nth and (n+1)th years
Eq (B-9)
Estimated PR calculation for the (n+1)th year

Eq (5-23)
Estimated processing cost for pth product in the

Eq (5-18)
(n+1)th year
Figure 6-4: Processing cost estimation implementation process
154
4 Indirect material, stores & inventory,

freight & transportation, material
inspection, purchasing, packaging
Total MDC calculation for the nth and material handling costs and
and the (n-1)th years scrap for the nth and (n-1)th years
Eq (5-24)
MDCDI calculation for (n+1)th year

Total MDC and cumulative material
costs for the nth and (n-1)th years,
inflation in the nth and (n+1)th years
Eq (B-10)
Estimated MDC for the pth
product in the (n+1)th year
Eq (5-25)
Figure 6-5: MDC estimation implementation process
5
Estimated material, labour,
processing and material-
dependent costs for the pth
Manufacturing cost estimation (MCE) product in the (n+1)th year
for pth product in the (n+1)th year
Eq (5-2)
Cumulative material, labour,

Cumulative manufacturing costs in processing and material dependent
the nth and the (n-1)th years costs for the nth and (n-1)th years
Eq (5-44)
Figure 6-6: Manufacturing cost estimation implementation process
155
6
Computer related and general
admin costs and selling expenses
in the nth and the (n-1)th years
Total PO in the nth and the (n-1)th
years Eq (5-42)
PODI calculation for the (n+1)th Total PO and cumulative

year manufacturing costs in the nth
and (n-1)th years, inflation in
the nth and (n+1)th years
Eq (B-14)
Estimated PO for the pth product in
the (n+1)th year
Eq (5-43)
Figure 6-7: PO estimation implementation process
In order to fully appreciate and understand the PCE Hybrid Model, the HMI algorithm and
the generated results, it is essential to understand the company’s method of cost estimation
and the estimated results. The next section, therefore, details the company’s estimation
method and the generated results before elaborating the Hybrid Model’s results.
156
6.3 PCE at the company
A well established crane and ship engineering company based in the UK (identity is
shielded due to confidentiality issues) was selected for the study as a representative case in
the country. The company designs, develops, manufactures, and markets a wide range of
cranes and parts for cranes and ships. It also provides after sales, installation and
commissioning services. The organizational structure of the company is divided into crane
manufacturing division (CMD) and ship engineering division (SED). The structural
verticalization within the organization meant that the divisions and sub-divisions carry out
their businesses independently. The total industrial output in the year 2005 was reported in
excess of £100m with £25m and £75m divided between the CMD and the SED
respectively. The CMD is divided into three sub-divisions namely spares & parts (SPSD),
assembly & services (ASSD) and installation & commissioning (ICSD) with the respective
industrial output in excess of £3m, £12m and £10m in the year 2005.
6.3.1 Information and details
The SPSD of the CMD was selected for the study. The main reason for the selection of this
sub-division in particular and the company in general was that the developed model’s scope
in terms of its parameters was likely to be covered there. The company was keen on
initiating the validation process from a relatively smaller unit. Table 6-1 highlights the
industrial output values in (£s) from 2002-2005 for the CMD.
157
Design of parts and products at the company takes place at the divisional level. New
product designs are normally generated for the ASSD where crane assemblies and sub-
assemblies are manufactured from the outsourced parts and/or procured from the SPSD.
The designs for products manufactured at the SPSD are already made, product range being
the standard. However, any minor changes are customized to the actual product designs.
Table 6-2 outlines the product range for the SPSD for the duration of the study.
The SPSD comprise fabrication and assembly units. There are a total of 20 different
machines in the sub-division. During its manufacture a product is first routed on the
machines within the fabrication unit before being routed through the assembly unit
machines until its final completion. The products route through a set of predetermined
machines within each unit based on its processing requirements. On its completion, a
product is handed over to the CMD sales from where it is either shipped to the customers or
sent to the ASSD.
Table 6-1: Industrial output values for the CMD (2002 – 2005)
Industrial Output Values 2002 2003 2004 2005
SPSD 2,192,384 2,534,222 2,895,843 3,200,353
ASSD 10,983,782 11,625,460 12,098,570 12,595,000
ICSD 9,901,843 9,955,987 9,980,350 10,048,380
Total (CMD) 23,078,009 24,115,669 24,974,763 25,843,733
158
Table 6-2: Product range at the SPSD
S. No. Product Number Product Description
1. MQ4033 Tripod crane
2. MQ4030 Tripod crane
3. MQ4024 Overhead mounted bridge crane gripping manipulator
4. MQ3522 Overhead mounted gripping manipulator
5. MQ2538 Overhead trolley mounted gripping manipulator
6. MQ1033 Overhead mounted linear track gripping manipulator
7. GQ4026 Column mounted double wire gripping manipulator
8. GQ4024 Column mounted gripping manipulator
Products within a unit are manufactured either in small batches or independently. Product
quantities are planned based on quarterly sales forecast released a quarter in advance of the
actual production. For example, the quantities for production in the third quarter are
released at the end of the first or the beginning of the second quarter in order to make
necessary arrangements for material procurement and production planning and control.
Product cost evaluation and review is carried out yearly. Table 6-3 gives out the details of
the product quantities produced in the SPSD from 2001 to 2005.
Product cost at the company is made up of material cost, labour cost and overheads. Actual
unit product costs are obtained at the end of a financial year from the cost accountancy data
159
based on the cumulative production cost of a given product and its respective number of
units produced in a given year. The values are given in the Table 6-3 and were provided by
the company.
Table 6-3: Yearly production quantities for the product range in the SPSD (2001 – 2005) and Actual
unit product cost for the given product range (2002 – 2005)
Product No of units produced Actual Product Cost per unit (£)
2001 2002 2003 2004 2005 2002 2003 2004 2005
MQ4033 430 450 400 485 324 1068 1197 1344 1520
MQ4030 265 250 230 200 265 955 1082 1226 1428
MQ4024 145 150 165 190 175 1948 2155 2370 2665
MQ3522 175 150 135 165 185 1267 1388 1551 1740
MQ2538 280 325 360 345 378 1492 1659 1851 2100
MQ1033 432 375 450 319 316 487 540 628 729
GQ4026 265 245 298 335 285 807 911 1005 1174
GQ4024 318 325 350 245 329 386 435 477 555
Material cost here refers to the direct material costs. Material quantities are ordered based
on the planned orders. The actual material quantities take into account the actual scrap and
160
wastage. The actual material costs are, therefore, only available after the production and
can be obtained from the financial data at the end of the production year. Table 6-4 presents
the actual material costs in (£s) for the product range during the period of study. The
cumulative material costs were obtained at the end of a given financial year and used to
furnish the unit material costs for the respective products for that year.
Table 6-4: Actual material cost (Cumulative and per unit costs)
Product Total material cost (Actual) (£) Unit material cost (Actual) (£)
2001 2002 2003 2004 2005 2001 2002 2003 2004 2005
MQ4033 214140 228100 211200 265780 181440 498 507 528 548 560
MQ4030 110240 105120 101430 92000 129850 416 420 441 460 490
MQ4024 176610 185210 215325 259540 252350 1218 1235 1305 1366 1442
MQ3522 123550 112125 104625 135300 158175 706 748 775 820 855
MQ2538 225400 280520 327600 331200 383670 805 863 910 960 1015
MQ1033 101520 95240 117900 92510 98592 235 254 262 290 312
GQ4026 84535 87620 110856 123950 112575 319 358 372 370 395
GQ4024 52470 54200 59500 40425 56588 165 167 170 165 172
Total 1,088,465 1,148,135 1,248,436 1,340,705 1,373,240
The material cost calculation for a specific product in a given year is based on the amounts
of different materials in that product and their respective unit costs for that year. Material
quantities are obtainable from BOMs (see example in appendix A). Since the products are
161
standardized, BOMs are established. Actual material costs incurred for a product can vary
based on a number of factors and are available after its production and generally at the end
of a financial year.
Table 6-5: Estimated unit material costs for the given product range (2003 – 2005)
Product Estimated unit material cost (£)
2003 2004 2005
MQ4033 530 554 574
MQ4030 439 463 482
MQ4024 1290 1370 1430
MQ3522 781 814 859
MQ2538 902 956 1005
MQ1033 265 275 304
GQ4026 374 391 387
GQ4024 174 179 173
Material cost estimation at the SPSD is based on considering the actual product costs at the
end of a given year to predict for the following year. Average inflation values are used to
estimate the material costs for the following year. A cost uncertainty index (CUI) value is
also used on top of the inflation to account for any uncertainties. The average inflation
values of 2.4, 2.5, 3.0, 3.5 and 3.2 percent were used respectively for 2001 – 2005 along
162
with a fixed CUI value of 1.5 percent. For example, the actual material cost value of £507
for MQ4033 obtained at the end of year 2002 was used to estimate for the year 2003 to be
£530. Table 6-5 gives out the estimated material cost values for the given product range
predicted by the company for 2003 – 2005.
6.3.3 Labour cost estimation
The total labour cost in a given year for the SPSD is the summation of the wages paid to all
the workers in both the fabrication and the assembly units for that year. The yearly capacity
is given in man-hours and is calculated by multiplying the number of working days in a
year, the average number of shifts per day, the number of hours per shift and the average
number of workers working per shift. The total labour cost obtained at the end of a given
year is divided by the capacity to provide an aggregate shop floor-wide labour rate. Table
6-6 provides the aggregate labour rate values for the respective years.
163
Table 6-6: Aggregate labour rate calculation
2001 2002 2003 2004 2005
Shifts per day 1 1 1 1 1
No. of working days per year 250 250 250 250 250
No. of hours per shift 8 8 8 8 8
No. of workers per shift 21 26 26 26 26
Capacity (man hours) 42,000 52,000 52,000 52,000 52,000
Total Wages (£) 240,120 336,600 387,000 460,080 507,600
Aggregate Labour Rate (£/hr) 5.72 6.47 7.44 8.85 9.76
Based on the lead times and number of workers working on different products in each of
the shop floor units, the total man-hours for individual products can be calculated as shown
in Table 6-7. The labour rate available at the end of a given year is used for the following
year to estimate the labour costs for individual products by multiplying the rate with the
products’ corresponding man-hour values. For example, the labour rate of 6.47 £/hr
obtained at the end of 2002 was set as an estimated rate for the following year giving an
estimated labour cost value of £337 in the year 2003 for MQ4033. Table 6-7 tabulates the
estimated labour costs also.
164
Table 6-7: Total man-hours and estimated labour costs (2003 – 2005) for the given product range
Estimated Labour
Product Fabrication Unit Assembly Unit Total
Cost (£)
Lead Lead Lead Lead

No. of man- No. of man- man-
time time time time 2003 2004 2005
workers hours workers hours hours
(min) (hr) (min) (hr)
MQ4033 240 4.0 5 20.0 240 4.0 8 32.0 52.0 337 387 460
MQ4030 330 5.5 5 27.5 200 3.3 5 16.7 44.2 286 329 391
MQ4024 180 3.0 6 18.0 220 3.7 4 14.7 32.7 211 243 289
MQ3522 150 2.5 6 15.0 180 3.0 6 18.0 33.0 214 246 292
MQ2538 220 3.7 7 25.7 180 3.0 6 18.0 43.7 283 325 386
MQ1033 120 2.0 6 12.0 100 1.7 4 6.7 18.7 121 139 165
GQ4026 220 3.7 6 22.0 230 3.8 5 19.2 41.2 266 306 364
GQ4024 120 2.0 3 6.0 180 3.0 3 9.0 15.0 97 112 133
6.3.4 Overhead estimation
Overhead costs at the company refer to all the costs other than material and labour costs.
Overheads incurred in a given year are known at the end of a financial year. The overheads
incurred in that year are summed up and divided by the man-hour capacity to give the
overhead rate. The rate is used for the following year to estimate the overheads likely to be
incurred on individual products by multiplying it with the corresponding man-hour values
for the products. Table 6-8 details the yearly overheads in £s obtained at the end of each
165
financial year. Scrap values are presented in negative to show the salvage value. The
individual overhead cost elements represent categories with like costs. Indirect material
here not only represents the cost on material used for shared benefits of products but the
tools utilized. Since the tooling is not heavy with comparatively lower costs, the company
has not assigned a separate cost pool for the tooling cost. Insurance costs here refer to the
costs of insuring machinery, equipment, building and all other insurable items. However,
the costs of insuring the equipment and machinery are higher than any other insurance
costs. Any costs not incurred for the direct benefit of the SPSD but for the shared benefit of
the company are charged proportionately.
Total actual production cost for a year can be calculated by summing the total material,
wages and overheads in that year and are also given in Table 6-8. The estimated values for
product costs using the company’s method can now be given by summing the three
estimated elements: material, labour and overheads as mentioned in Table 6-9. The table
also gives the cumulative values for the estimated production costs (2003 – 2005).
166
Table 6-8: Overhead costs for individual elements (2001 – 2005)
2001 2002 2003 2004 2005
Indirect material 55050 56500 64200 75200 75800
Stores & Inventory control 45600 48900 54200 56240 62450
Freight & Transportation 35200 39900 58200 62400 85240
Material Inspection 24500 25600 26500 27000 27560
Purchasing 17800 18970 19250 21130 25360
Packaging 26050 27700 32540 45230 51230
Material handling 16890 17290 19260 19265 21450
Insurance cost 67300 69500 72000 88000 98400
Financing Expenses 24500 28000 36000 42000 45000
Maintenance Cost 20850 25424 38586 51384 62598
Repair Cost 18853 21220 38106 25814 41852
Energy cost 57200 65000 98000 130000 160000
Computer related cost 44280 45550 64870 85925 105425
General admin cost 139660 145650 182560 210220 238954
Selling Expenses 95255 109245 132364 194500 256450
(Scrap) -34350 -36800 -37850 -39250 -38256
Total Overheads (actual) 654,638 707,649 898,786 1,095,058 1,319,513
Overhead rate (£/hr) 15.59 13.61 17.28 21.06 25.38
Total production cost (actual) 1,983,223 2,192,384 2,534,222 2,895,843 3,200,353
167
Table 6-9: Estimated product cost values (2003 – 2005)
Product Estimated product cost per unit Estimated production cost (Cumulative)
2003 2004 2005 2003 2004 2005
MQ4033 1574 1840 2129 629579 892490 689762
MQ4030 1326 1555 1802 305059 311029 477660
MQ4024 1946 2178 2407 321139 413818 421251
MQ3522 1444 1630 1845 194918 268905 341409
MQ2538 1779 2035 2311 640397 702155 873572
MQ1033 640 737 862 288117 234996 272356
GQ4026 1200 1409 1619 357726 471852 461284
GQ4024 475 549 621 166425 134603 204425
Total 2,903,361 3,429,848 3,741,718
6.4 Implementation of the PCE Hybrid Model
The developed model presents the estimated product cost as a summation of manufacturing
cost, engineering cost and production overheads. Since the engineering designs for the
SPSD are already established, the engineering costs in terms of design and development
expenditures are negligible. Any minor costs within the context are incurred at the ASSD
and subsequently charged to the SPSD within general administration overheads. These
costs were not separately provided by the company. Manufacturing cost and production
overheads would, therefore, add up to the estimated product cost for an individual product.
168
Manufacturing cost is presented as material, labour and indirect manufacturing costs.
Indirect manufacturing costs comprise processing, material-dependent, tooling and building
space costs. Any cost incurred on tooling at the company is not separately recorded but
combined with the indirect material cost and would form part of material-dependent cost
(MDC) estimation. Plant-wide building maintenance, repair and energy costs are allocated
in proportion to the industrial output and charged to the division and subsequently
combined with equipment & tool maintenance, repair and energy costs respectively. These
costs would, therefore, be dealt with part of processing cost estimation. Manufacturing cost
estimation in this case would be the sum of material, labour, processing and MDC.
The application of the HMI algorithm presented in section 6.2 can be followed from the
following sub-sections in order to facilitate the implementation of the Hybrid Model and
generate estimated cost values.
Material cost estimation using the presented Hybrid Model required calculation of material
cost deviation index, φ n +1 . Material cost deviation trends from the past years could be used
to predict the index values for the following years based on the model derived (see
appendix B) as follows:
 C mp
n

φ pn +1 = I n +1 + (1 + I n +1 ) ×  −1− I n  (6-1)
C n −1 
 mp 
169
Where, φ pn +1 is the deviation index for the pth product estimated for the following year. I n
and I n+1 are the interest rates in the current and the following year respectively. C mp
n
and
n −1
C mp are the material costs for the pth product in the current and the preceding year
respectively.
Table 6-10: Material cost deviation index and estimated unit material cost values (2003 – 2005)
Estimated unit material cost (£) Material cost deviation index (%)
Product
(Hybrid Model) φ pn +1
2003 2004 2005 2003 2004 2005
MQ4033 518 553 567 2.26 4.71 3.50
MQ4030 427 465 479 1.53 5.45 4.03
MQ4024 1257 1387 1426 1.84 6.29 4.41
MQ3522 796 808 866 6.48 4.20 5.58
MQ2538 931 965 1010 7.86 6.01 5.26
MQ1033 276 272 321 8.74 3.67 10.62
GQ4026 404 389 366 12.90 4.55 -0.97
GQ4024 169 174 159 1.53 2.40 -3.45
Average index ( φ
n +1
) 5.39 4.66 3.62
By using all the known values, the model can predict the indices for the product range for
the following year. The values can then be used to predict the estimated material cost
170
values. An average index value can also be used for the purpose but it may compromise the
accuracy of the individual results. However, the purpose of the average value is best served
for a new product with no known past data. The material cost values available at a given
time and the average index value can then be used to predict the product cost for the
following year. Table 6-10 gives out the indices and the estimated material cost values.
6.4.2 Labour cost estimation
Labour cost estimation using the hybrid method requires the estimation of shop floor-wide
aggregate labour rate and the labour units consumed for each product. The lead times for
the individual products in the two work centres and the corresponding data obtained from
the company for the number of skilled (x), semi-skilled (y) and non-skilled (z) workers
required to manufacture the respective products can be used to determine labour units for
each product. Skill indices of 0.75 and 0.35 were set for semi-skilled and non-skilled
workers respectively based on an average skill level of around 75 and 35 percent with
respect to their skilled counterparts. Table 6-11 tabulates the labour units in hours required
for the manufacture of the individual products. The table also calculates the total number of
labour units consumed in each year using the data for the number of units produced in
corresponding years.
The total labour units and the total wages paid in a year can be used to determine the labour
rate for that year. The rate can then be used to determine the estimated labour rate for the
following year by considering the LCDI, ε n +1 for the following year. The LCDI values can
171
be found by using the equations (B-6), (B-7) and (B-8). Average monthly wages of the
skilled, semi-skilled and non-skilled workers obtained from the company were used for that
purpose. The average index values as obtained can be used to determine the estimated
labour rates. The estimated rates can be used to generate estimated labour costs for
individual products based on their respective labour units as shown in Table 6-12.
Table 6-11: Labour units calculation
Product Fabrication Assembly Labour Total labour units consumed

Units
x y z x y z (hr) 2001 2002 2003 2004 2005
MQ4033 1 2 2 2 3 3 34.00 14620 15300 13600 16490 11016
MQ4030 2 2 1 2 1 2 32.68 8659 8169 7515 6535 8659
MQ4024 1 2 3 2 1 1 22.02 3192 3303 3633 4183 3853
MQ3522 2 1 3 2 2 2 22.10 3868 3315 2984 3647 4089
MQ2538 2 3 2 1 2 3 28.80 8064 9360 10368 9936 10886
MQ1033 1 2 3 1 1 2 11.18 4831 4194 5033 3567 3534
GQ4026 2 1 3 2 1 2 27.16 7197 6654 8093 9098 7740
GQ4024 1 1 1 1 1 1 10.50 3339 3413 3675 2573 3455
Total 53770 53706 54900 56029 53231
α 0.75 0.75
β 0.35 0.35
172
Table 6-12: Labour cost deviation index and estimated labour costs
Average wages per month (single Labour cost deviation Estimated unit
Product
employee) (£) index (%) labour cost (£)
2001 2002 2003 2004 2005 2003 2004 2005 2003 2004 2005
Skilled 980 1100 1290 1580 1700 13.04 18.27 22.79 MQ4033 248 273 326
Semi 810 950 1080 1280 1400 18.23 14.56 18.70 MQ4030 248 272 326
Non-
650 800 910 1050 1200 24.19 14.63 15.46 MQ4024 165 181 217
skilled
Average Index ( ε
n +1 18.49 15.82 18.98 MQ3522 167 183 219
)
Labour Rate (Hybrid Model) (£/hr) MQ2538 210 231 277
Actual 4.47 6.27 7.05 8.21 9.54 MQ1033 82 90 107
Estimated 7.43 8.16 9.77 GQ4026 208 228 273
GQ4024 79 86 103
6.4.3 Processing cost estimation
The total processing costs spent on the shop floor refer to the necessary costs to keep the
floor facilities in the running condition. Financing expenses, insurance, maintenance, repair
and energy costs obtained at the end of the financial year add up to the total processing
costs for the year. Processing costs for a product can be estimated by multiplying the
estimated rate with the processing units of the product. Processing units refer to the total
time the product spends on the work centres while being processed. The combined lead
time, therefore, for a given product spent on the fabrication and the assembly units would
173
refer to the processing units for the product. The total processing costs for a given year and
the total processing units consumed in that year can be effectively utilized to produce the
processing rate for the year.
Table 6-13: Estimation of processing units, deviation indices, rates and costs
Estimated processing
Processing units (Total)
Processing cost per unit (£)
Product
Units (hr)
2001 2002 2003 2004 2005 2003 2004 2005
MQ4033 8.00 3440 3600 3200 3880 2592 128 198 216
MQ4030 8.83 2341 2208 2032 1767 2341 142 219 238
MQ4024 6.67 967 1000 1100 1267 1167 107 165 180
MQ3522 5.50 963 825 743 908 1018 88 136 148
MQ2538 6.67 1867 2167 2400 2300 2520 107 165 180
MQ1033 3.67 1584 1375 1650 1170 1159 59 91 99
GQ4026 7.50 1988 1838 2235 2513 2138 120 186 202
GQ4024 5.00 1590 1625 1750 1225 1645 80 124 135
Total 14738 14638 15109 15028 14578
Processing costs (Total) 188703 209144 282692 337198 407850
Processing rate (actual) 12.80 14.29 18.71 22.44 27.98
Processing cost deviation index ( µ

n +1
) (%) 12.37 32.42 20.15
Processing rate (estimated) (£/hr) 16.06 24.78 26.96
174
However, predicting the rate for the following year would require considering any
processing costs deviations. The method for calculating such deviations mentioned earlier
can be used to predict the PCDI values, µ n +1 , for the following year by using the equation
(B-9). The PCDI values can then be used to generate the estimated processing rates.
Estimated processing costs can then be determined for the range of products. Table 6-13
covers the estimation of processing units per unit and the total processing units required in
a given year. It also calculates the rates and the deviation index values and then gives out
the estimated processing cost per unit for the given products.
6.4.4 MDC estimation
The total MDC for the SPSD for a given year is obtained by summing up indirect material,
stores and inventory, freight and transportation, material inspection, purchasing, packaging
and material handling costs and subtracting the scrap value for that year. MDC per unit
material consumed can then be used to determine the MDC deviation index ( ρ n+1 ) values
using the method described earlier and given by equation (B-10). This allows the estimation
of the MDC per unit material cost for the following year resulting in the estimation of the
MDC per unit of the corresponding products for that year. Table 6-14 outlines all these
values.
175
Table 6-14: MDC deviation index and estimated MDC per unit values
2001 2002 2003 2004 2005
n
Total MDC ( C md ) (£) 186,740 198,060 236,300 267,215 310,834
n
Cumulative material cost ( C mt ) (£) 1,088,465 1,148,135 1,248,436 1,340,705 1,373,240
n n
MDC fraction ( C md / C mt ) 0.172 0.173 0.189 0.199 0.226
MDC deviation index ( ρ

n +1
) (%) 0.99 10.46 5.06
Estimated MDC fraction

0.174 0.209 0.209
( C md / C mt )x( 1 +
n n
ρ n+1 )
Product Estimated MDC per unit (£)
MQ4033 90 116 119
MQ4030 74 97 100
MQ4024 219 290 299
MQ3522 139 169 181
MQ2538 162 202 212
MQ1033 48 57 67
GQ4026 70 81 77
GQ4024 29 36 33
176
6.4.5 Production overheads estimation
The total PO for the SPSD in a given year can be obtained by adding selling expenses,
computer related costs and general administration costs. The total actual manufacturing
costs for the SPSD in a year refers to the summation of the cumulative material costs, total
wages, total MDC and the total processing costs incurred in that year.
Table 6-15 PO fractions (actual & estimated) and deviation indices
2001 2002 2003 2004 2005
Total manufacturing cost (actual)

n
( CGt ) (£) 1,704,028 1,891,939 2,154,428 2,405,198 2,599,524
n
Production overheads ( Ot ) (£) 279,195 300,445 379,794 490,645 600,829
n n
PO fraction ( Ot / CGt ) 0.164 0.159 0.176 0.204 0.231
PO deviation index ( τ
n +1
) (%) -2.74 11.79 15.81
Estimated PO fraction ( Ot / CGt )x( 1 + τ

n n n +1
) 0.154 0.197 0.236
The total PO and the total manufacturing costs can be used to establish the PO fractions and
subsequently the estimated values of PO fractions. The estimated values of the fractions
require the calculation of the PO deviation indices, τ n+1 using the method mentioned earlier
for the calculation of the indices and given by equation (B-14). Table 6-15 explains the
calculation of the fractions and the indices. By adding the already estimated manufacturing
cost elements, the estimated values for manufacturing costs for individual products in a
177
given year can be obtained. These values and the corresponding estimated PO fractions can
be used to estimate the production overheads for the individual products. Estimated product
cost can then be determined by combining the manufacturing and the PO cost elements
estimated for the product. Table 6-16 tabulates the estimated manufacturing, PO and
product costs per unit.
Table 6-16: Estimated per unit values for manufacturing, PO and product costs
Estimated
Estimated production Estimated product cost per
product manufacturing cost per
overheads per unit (£) unit (£)
unit (£)
2003 2004 2005 2003 2004 2005 2003 2004 2005
MQ4033 990 1144 1234 153 225 291 1142 1370 1525
MQ4030 886 1048 1136 137 207 268 1023 1254 1405
MQ4024 1747 2022 2120 270 398 501 2017 2420 2621
MQ3522 1187 1293 1411 183 255 333 1370 1548 1745
MQ2538 1414 1567 1683 218 309 398 1633 1875 2081
MQ1033 466 511 596 72 101 141 538 611 737
GQ4026 796 878 911 123 173 215 919 1051 1126
GQ4024 357 420 430 55 83 102 412 503 532
178
6.5 Conclusions
This chapter discussed the implementation of the Hybrid Model in a batch manufacturing
environment in the UK as part of the validation process for the model. The implementation
phase involved the retrospective application of the PCE Hybrid Model along with the
company’s method of cost estimation. The implementation of the Hybrid Model was
facilitated by the HMI algorithm. Cost data at the end of a given year were effectively used
with the help of the proposed deviation indices, resulting in furnishing the estimated costs
for the given product range well before the actual manufacturing started.
It was noted that the Hybrid Model with the help of the HMI algorithm generated detailed
cost breakdown values for every product. The company’s method was not able to predict
such breakdown values. The provision of breakdown values is helpful in identifying high
cost elements and subsequently helps in making cost control decisions. The cumulative cost
values were also obtained that can be helpful in making statistical and financial analysis of
the cost data at the organizational level. Such analyses support key operational and strategic
decisions for middle and senior management.
The Hybrid Model attempted to overcome the limitations associated with TRO and MRO
models. For example, inflation and other cost deviations were incorporated through the
introduction of deviation indices that also optimized the use of past data. The use of normal
capacity hours during labour cost and overheads (indirect costs) calculation was replaced
by a new method based on calculating labour and processing units etc.
179
Deviation indices are crucial to the overall accuracy of the estimated results and as such
demand further investigations. For example, there is a need to investigate the effects of
more than two years past cost trends. The HMI algorithm can also be made more generic by
adding components for tooling and building space costs calculation in line with the Hybrid
Model. Care should be taken while categorizing the indirect elements to different sub-
categories. For example, material handling costs are categorized under material-dependent
costs based on its relation with the increasing material quantity. Increased material
quantities result in increased handling and thus in increased costs. However, an increased
handling time should not be confused with processing time and the resulting costs should
not be categorized under processing costs. Because processing costs are based on job
processing times at work or machine centres. On the other hand, an increased material
handling time is accounted for in increased material quantities. Similar care should be taken
for other overheads elements.
In short, previous limitations found in TRO and MRO estimation methods were removed.
The obtained results are compared in the next chapter for the validation analysis.
180
Chapter 7 Comparisons and
Validation Analysis
The aim of Chapter 7 is to validate the developed PCE Hybrid Model by comparing the
estimated costs obtained in the previous chapter and in the current chapter against the
actual costs for a range of products. The process will involve analysing the costs at
different levels including the entire product costs, cumulative production costs, overheads
for an individual product and overheads at cumulative level. The chapter will also present
a breakdown analysis.
7.1 Introduction
The route to successful validation is inscribed in achieving the objectives. This in turn
requires outlining the validation method. In Chapter 4, the MRO and TRO estimation
methods were presented, implemented and validated by comparing the results obtained
from the application of the model with those generated by the company’s method. The
model presented in Chapter 5 is more comprehensive and requires more than just
scrutinizing it against an in-house methodology. A decision, therefore, was made to
implement a published model and obtain the estimated cost results from its application
alongside the already obtained results in Chapter 6 from the application of the PCE Hybrid
181
Chapter 7: Comparisons and Validation Analysis
Model and the company’s method. The three results will then be compared against the
actual product costs. Comparing the estimated product costs from the three methods against
the actual costs can only strengthen the argument to justify the superiority of the developed
methodology. The newly developed Hybrid Model was therefore, implemented alongside a
published model in the selected manufacturing company for a retrospective analysis of
three years (2003 – 2005).
Chapter 7 presents the comparisons and validation analysis. This is based on comparing the
estimated cost values (2003 – 2005) obtained using the three methods (the company’s
method, the selected published model and the Hybrid Model) against the actual cost values
provided by the company and given in Table 6-3, Table 6-4, Table 6-6 and Table 6-8. An
analysis of the cost variations from the actual costs is presented by not just considering the
cumulative production costs but the individual product costs and the cost elements, where
possible. Almost every level of the analysis demonstrates the superiority of the developed
model. An analysis for cost elements breakdown based on the actual production costs is
also drawn. The breakdown results obtained highlight the typical tendencies of the region
through the analysis of a representative case. The analysis when compared with the results
obtained for the South Asian region presented in Chapter 4 provides the basis for
understanding and developing an effective CCS.
The actual product cost values were provided by the company and calculated based on the
actual cost figures obtained from the cost accountancy data at the end of each year. These
considered the actual material and labour costs. Material costs considered the actual
182
material quantities and wastage. Labour costs were based on the actual time spent on a
product or a batch of products taken from the time cards on the shop floor. Any other costs
were allocated to the products based on either activity rates and units or their traceability to
the products. For example, material handling, stores & inventory, inspection and quality
control costs were allocated based on activity rates and units. Freight & transportation,
packaging, financing expenses etc. were in most cases traceable to the products. However,
any exact method for their calculation was not given by the company.
Before the validation analysis, the results from a published model would be obtained in
order to compare the three estimated results against the actual costs. Section 7.2 details the
cost estimation results for the given product range using the selected published model as a
representative case from the domain. The section, therefore serves the preparation before
comparison analysis. Section 7.3 presents the product cost analysis based on comparing the
estimated product costs from the three methods against the actual costs. Section 7.4
compares the estimated and the actual cumulative costs. Section 7.5 analyses estimated and
actual overheads at product and production level. Section 7.6 presents a statistical analysis
of cost breakdown results. Section 7.7 concludes the chapter.
7.2 Preparations for comparisons
The purpose of the implementation of a published model is to validate the developed model
by not just comparing it against the industrially established tool but to present its
183
superiority against an already established and recognized representative case from the
published research domain. Therefore, the selected model should already have been tested.
In order to draw comparison between the developed and the published models, it is
essential that the published model should have the necessary parameters to be applicable
within the framework of the manufacturing enterprise. The available field data should form
the basis for the selection of the model and to further evaluate costs.
The model presented by Jung [84] was selected for the implementation and validation
purposes as it not only complied with the above criteria but is established in the field. Jung
described major cost elements as material, labour, engineering and burden cost. He did not
model engineering costs but manufacturing costs were presented as follows:
Mfg cost = (Ro + Rm)[( Tsu/Q)Tot + Tno] + material cost + factory expenses (7-1)
where, Ro = operator’s rate, Rm = machine rate, Tsu = set-up time, Q = batch size, Tot =
operation time, Tno = non- operation time.
Three different times are considered but refer to the manufacturing lead time (MLT) time
for a product. Labour and machine related costs were described using the MLT. Burden
cost or factory expenses refer to the costs other than material, labour and machine related
costs and can be considered as the overheads spent for the combined benefit of all the
products. In an industrial environment where engineering costs are relatively low,
negligible or accumulated with the overheads, Jung’s model refers to the overall product
184
cost. In the given conditions where engineering costs, if any, are part of the general
administration costs, the model can predict product costs. The major cost elements,
therefore, comprise labour and machine costs, material cost and factory expenses.
7.2.1 Labour and machine cost estimation
Labour and machine cost estimation using Jung’s model requires the operator’s and
machine rates calculation. The operator’s rate (Ro) is based on the shop floor’s direct labour
cost. The total wages can, therefore, be divided by the shop floor-wide cumulative MLT to
obtain the rate. The floor-wide MLT in a given year can be obtained by lead times of the
individual products and the respective number of units produced in that year. Similarly,
machine running costs (maintenance, repair and energy costs) can be divided by the total
MLT to determine the machine rates (Rm).
The obtained rates at the end of a given year can be used for the following year’s labour
and machine cost estimation process by considering the effect of inflation in that year.
Since, Jung’s model does not describe the ways to consider any cost variances, only the
known variation (i.e. average inflation) is considered for the estimation purposes. The
summation of the estimated rates can be multiplied by the individual lead times to estimate
the labour and machine costs for individual products. Table 7-1 tabulates the lead times
(individual and cumulative), machine running costs (the summation of maintenance, repair
and energy costs), the rates (operator and machine) and the estimated labour and machine
costs per unit.
185
Table 7-1: Lead times, machine running costs, rates (operator and machine) and labour and machine
costs
Manufacturing lead time (MLT) (total) Labour and machine cost

Product Lead time
(hr) per unit (£)
per unit
(hr)
2001 2002 2003 2004 2005 2003 2004 2005
MQ4033 8.00 3440 3600 3200 3880 2592 252 308 367
MQ4030 8.83 2341 2208 2032 1767 2341 279 340 405
MQ4024 6.67 967 1000 1100 1267 1167 210 257 305
MQ3522 5.50 963 825 743 908 1018 173 212 252
MQ2538 6.67 1867 2167 2400 2300 2520 210 257 305
MQ1033 3.67 1584 1375 1650 1170 1159 116 141 168
GQ4026 7.50 1988 1838 2235 2513 2138 237 289 344
GQ4024 5.00 1590 1625 1750 1225 1645 158 192 229
Total (hr) 14738 14638 15109 15028 14578
Operator's rate (Ro )

(£/hr) 16.29 23.00 25.61 30.61 34.82
Machine running cost

(£) 96,903 111,644 174,692 207,198 264,450
Machine rate (Rm)

(£/hr) 6.57 7.63 11.56 13.79 18.14
186
7.2.2 Material cost and factory expenses
The amounts of all the material used for the manufacture of a product and their unit prices
can be used to ascertain the total material cost of the product. The only known cost
deviation (i.e. the average inflation) can then be accounted for in order to predict the total
material cost for the product for the following year. For example, the material cost of £522
can be set for MQ4033 for 2003 based on the known value of £507 from 2002 and the
inflation rate of 3 percent for 2003.
Ascertaining the factory expenses for an individual product requires the calculation of an
expenses rate. The rate can be calculated by dividing the total factory expenses from the
cumulative MLT for a given year. The rates after taking into account any deviations effect
(inflation in this case) for the following year can be multiplied by the total lead times of the
products to furnish their corresponding estimated factory expenses values in that year.
Whereas, the total factory expenses for a given year can be obtained by summing all the
costs other than material and labour costs. Since, the maintenance, repair and energy costs
were considered in the calculation for machine related costs; they would also be excluded
from the factory expenses. The estimated product costs can then be given by adding the
corresponding factory expenses, material, labour and machine cost components. Table 7-2
calculates the factory expenses rate, the estimated factory expenses per unit and the
estimated product costs per unit.
187
Table 7-2: The total factory expenses, the expenses rate, the estimated per unit values (factory expenses,
product cost) (2003-2005)
2001 2002 2003 2004 2005
Factory expenses
557,735 596,005 724,094 887,860 1,055,063
(Total) (£)
Estimated product cost per

Factory expenses
37.84 40.72 47.92 59.08 72.37 unit (£)
rate (£/hr)
Product
Estimated factory expenses
per unit (£)
2003 2004 2005
MQ4033 336 397 488 1110 1251 1420
MQ4030 370 438 539 1082 1234 1418
MQ4024 280 331 406 1762 1938 2122
MQ3522 231 273 335 1174 1287 1434
MQ2538 280 331 406 1379 1529 1703
MQ1033 154 182 224 531 594 691
GQ4026 315 372 457 919 1046 1183
GQ4024 210 248 305 539 616 704
188
7.3 Comparison analysis for product cost
The estimated product costs using the Hybrid Model (new method), the company’s method
and the Jung’s model mentioned in Table 6-16, Table 6-9 and Table 7-2 respectively can be
plotted against the actual product costs given in Table 6-3.
Figure 7-1 gives the cost comparison for the entire product range for the three year period
(2003 – 2005). The green vertical columns represent the actual costs for respective
products. The other three dotted lines represent the estimated product costs. The analysis
clearly shows the superiority of the Hybrid Model as the dotted line representing the results
from the Hybrid Model remains closer to the green vertical bars (the actual costs). For
example, the estimated costs of £1110, £1574, £1142 were given by the Jung’s model, the
company’s method and the Hybrid Model respectively for the year 2003 against the actual
product cost of £1197 for the same year.
189
Figure 7-1: The comparison of the actual costs against the estimated costs
190
Any deviation of the estimated cost values from the actual cost represent the estimation
error linked with a given estimation method. In addition to the cost analysis presented
earlier, there is a need to analyse the estimation errors using the three methods. The
estimation error trends using the three methods for the investigation period are given in
Figure 7-2. Error estimations are based on deviations from the actual costs and presented in
percentages. The actual cost is represented by the x-axis with zero error. It is clear that on
almost every occasion the estimated values predicted by the Hybrid Model are closer to the
actual values than those predicted by the other two methods. Although on an isolated case
(MQ4030), the Jung’s model predicted slightly better values, the Hybrid Model’s
prediction was not only within the acceptable limit for that case but the overall accuracy is
far more consistent. It can be seen that a cautious approach by the company’s adopted
method resulted in not just overestimation on almost every occasion but resulted in as much
as 40 percent error. This was mainly due to an effort to avoid under estimating and setting a
price more than the actual cost value. However, the loss of goodwill and sales as a result of
overestimation could far outweigh any apparent profits as a result of overestimation
resulting in net loss of potential profits. A yearly cost rise in the actual values and the
estimated values is also visible due to an increase in almost all the cost element values.
Jung’s model underestimated on many occasions partly because of no provisions in the
model for considering cost variances that could be useful in keeping up with the cost trends.
Only the known variances were accounted for. Both under- and overestimations were
observed within acceptable limits using the Hybrid Model. It is also clear that the use of
deviation indices has eliminated the problem of underestimation associated with the TRO
and MRO methods presented in chapter 4.
191
Figure 7-2: Estimation error trends for the three methods (2003 – 2005)
192
It is not uncommon in an industrial environment to focus on net error estimation analysis.
While some of the products during a given period may be highly overestimated, the others
remain underestimated. The net estimation error may be within acceptable limits. However,
the approach can be dangerous as some of the products may be priced very low while the
others could be highly priced. This could lead to losses on individual products due to
underestimation. On the other hand, those overestimated and thus over-priced products
could also result in loss of consumer confidence and loss of market. Therefore, instead of
relying on net error estimation, there is a need to estimate individual product’s cost
accurately.
An analysis based on a yearly consolidated representation for the investigated period
considering the entire product range is given in Figure 7-3. Error estimation values are also
given. Big error distributions across the product range are seen in a given year using both
the company’s and the Jung’s method. The company’s method overestimated for almost the
entire product range throughout the investigation period. Jung’s method mostly under
estimated for the given product range throughout the investigation period. The values
estimated by the Hybrid Model were either underestimated or overestimated within very
close range of the actual costs. In any given year, the values were evenly spread across the
entire product range.
193
Figure 7-3: Percentage cost estimation variations from actual product costs.
194
In order to visualise better the error distribution trends across the product range for a given
year using the three methods, results are presented in Figure 7-4. The error trends for a
given year demonstrate the effectiveness of the Hybrid Model when compared with the
other two methods. Huge fluctuations are seen for the estimation error values using either
the company’s method or the Jung’s model across the entire product range for the three
year period. The Hybrid Model’s estimation trends, on the other hand, are clearly even and
stay close to the actual values. In terms of not just the accuracy, but the consistency of the
results, the Hybrid Model can be seen as a superior methodology. Estimation errors
displayed on various occasions are almost negligible unlike those displayed by the other
two methods. The effective use of the past data along with accurate prediction of the
various indices is behind the accurate prediction of the overall product costs among other
factors. Such precision is mainly down to eliminating the already observed shortcomings in
the methodology presented in Chapter 4.
195
Figure 7-4: Estimation error trend across the product range
196
Due to huge fluctuations of estimation error across the entire product range using the two
methods, some of the products may be highly over estimated while the others highly
underestimated resulting in unrealistic price settings. Such a situation could have serious
implications for the business objectives of an enterprise. Many companies tend to even out
the errors by linearizing the error distribution. Even any error linearization could do little to
help the situation. For example, the linearization of the errors for the company’s method
could help to even out the error distribution across the product range, but the overall error
still stays unreasonably high (around 20 percent) as shown in Figure 7-5. Most importantly,
such a linearization has little practical implications as such a comparison requires the actual
costs which are not discovered until usually the products are produced with price quotes
already furnished. However, for statistical analysis and accuracy optimization (based on the
rectification of the consistently inaccurate trends) purposes, such results can be helpful.
Past error distribution trends, however, may have to be used and relied upon at increased
risk of compromising the accuracy for future results.
Similarly, the linearization carried out for both the Jung’s model and the Hybrid Model are
shown in Figure 7-6 and Figure 7-7 respectively. The linearization of the errors for the
Jung’s model finds the underestimation of up to around 20 percent and an overestimation of
around 10 percent. The values for the Hybrid Model, on the other hand, do not exceed 5
percent for both under and overestimation.
197
Figure 7-5: Error linearization for the results given by the company’s method
198
Figure 7-6: Error linearization for the results given by the Jung’s method
199
Figure 7-7: Error linearization for the results given by the Hybrid Model
200
Another point noticeable from Figure 7-7 is a changing response for error estimation trend
on yearly basis as opposed to the ones presented in Figure 7-5 and Figure 7-6. This is due
to the fact that the Hybrid Model takes into account the changing factors to estimate costs.
Factors like, inflation rate, expected cost deviations etc. are accounted for very well and by
adjusting the response, the estimated values are always kept close to the actual costs. The
incorporation of the case-based technique into the modelling framework is also responsible
for sensitizing the model by keeping it aligned with the changing trends. While the model is
sensitized to the trends, its robustness remains unquestionable. The overall validation
analysis spread over a three-year period covering a range of products under varying
conditions and huge fluctuations of cost deviations still produced consistent results within
acceptable limits of error estimations. The estimated values are, therefore, reliable and can
serve the operational purposes (such as price setting, cost control, lot sizing, etc.) well.
More strategic decisions can also be relied upon (such as make or buy decisions, business
resizing, etc.). The results produced by the other two methods are hugely desensitized to the
changing cost trends and consistently produced results with huge estimation errors and
unchanged trends. Although, the statistical data analysis can help to exploit the error
consistencies to manually adjust the estimated values, the process would not be risk free
and the essence of using an estimation method will be lost.
201
7.4 Comparison analysis for cumulative costs
In addition to the analysis for individual products, it is essential to carry out the validation
for the cumulative cost values. This is also necessary to quell any notion of a possible self-
adjustment of individual under and overestimation on the cumulative level. This means an
attempt will be made to determine if any under or overestimation at product level even out
each other at cumulative level. In other words, the current analysis is aimed at finding out if
a methodology less accurate at estimating individual product costs can be more accurate at
cumulative level. This would be done by comparing the total actual costs incurred in the
SPSD during a given period against the cumulative values for the estimated results obtained
by using the three methods.
Figure 7-8 (a) compares the three estimated cumulative costs with the actual costs at the
cumulative level. The values are given in million £. The black line representing the actual
costs stays closer to the green vertical bars representing the estimated values given by the
Hybrid Model (new method). Figure 7-8 (b) quantifies the estimation error. The estimation
error values represent the percentage deviations from the actual costs. The estimated values
predicted by the Hybrid Model at the cumulative level fluctuated between -3 and 2 percent
for the overall period of the investigation. The company’s method overestimated on the
cumulative level by as much as around 19 percent despite fluctuating under and
overestimation values at individual levels. The Jung’s model, on the other hand,
underestimated costs by around 9 percent at the cumulative level despite fluctuating values
at the individual products. This demonstrate that even after taking into account any
202
adjustments of fluctuating values for individual products at the cumulative level, high
degree of under and overestimations were observed using the two methods.
Actual Cost Vs Estimated Cost (a)
4.0
Millions
3.5
Cost Values (£s)
3.0
2.5
2.0
2003 2004 2005
Production Year
Total Estimated Production Cost (Jung's) Total Estimated Production Cost (Comapny)
Total Estimated Production Cost (New ) Total Production Cost (Actual)
Estimation errors from cumulative actual costs (%) (b)
20
14.57 16.92
18.44
Estimation variation (%)
10
1.79
0
-3.01 -1.17
2003 2004 2005
-10 -7.66 -8.40 -9.29
Production Year
-20
Estimation error using Jung's model (%) Estimation error using company's method (%)
Estimation error using hybrid model (%)
Figure 7-8: (a) Cumulative actual costs against total estimated values, (b) estimation errors for
cumulative costs
203
The Hybrid Model remained close to the actual costs at the cumulative level of prediction.
Even if the cumulative results for the two methods were closer to the actual costs after
considering any adjustments, the impact of huge under and overestimation on individual
products could not be ignored. However, more accurate results for the Hybrid Model even
at the cumulative level demonstrate the ultimate superiority of the model over the two
methods considered. In addition, the highly accurate values predicted by the Hybrid Model
both at the individual and the cumulative levels with consistency would be very difficult for
any other method (not considered in the current validation analysis) to rival.
The Hybrid Model gives out optimized values at the cumulative level (i.e. more accurate
than those predicted by the other two methods). The quantification of the optimized values
against the two methods is shown in Figure 7-9. The results are based on considering the
estimation error values for the cumulative costs. Only the error values in the respective
years are considered ignoring the symbolic negative signs, if any (meant for showing
underestimation), thus allowing the quantification of the deviations from the actual costs.
The differences between the estimation error values for a given year using the Hybrid
Model and the company’s method resulting in the optimized values are mentioned in the
first part of the figure. Those based on the Jung’s model are presented in the second part. A
maximum of over 16 and 8 percent optimizations were obtained against the company’s and
the Jung’s methods respectively. Linearized trends are also included in both cases showing
similar results even after adjusting the errors over the duration of the investigation. The
results show significant improvements achieved by the Hybrid Model against the two
methods.
204
Optimization achieved 2003-2005 (%) (a)
18.00
16.65
15.75
15.00
y = 2.09x + 10.47
12.00
Optimization (%)
11.56
9.00
6.00
3.00
2003 2004 2005
0.00
Production Year
Optimization from company's method (%)
Linear optimization trend from company's method (%)
Optimization achieved 2003-2005 (%) (b)
10.00
8.12
8.00
6.61
Optimization (%)
6.00 y = 1.73x + 2.99
4.65
4.00
2.00
2003 2004 2005
Production Year
0.00
Optimization from Jung's model (%)
Linear optimization trend from Jung's model (%)
Figure 7-9: Optimization for the estimation accuracy achieved by the Hybrid Model against (a) the
company’s method (b) the Jung’s model
205
7.5 Comparison analysis for product and production
overheads
An important aspect of the validation analysis is to investigate the effects of the three
methods on the accuracy of the breakdown elements. The accurate estimation of the
breakdown elements would eventually lead to the accurate prediction of the overall product
costs. Identifying the least accurate elements could help to better understand the problem
areas in the existing methods and reason why the developed model is more efficient.
Direct and indirect elements are estimated in different ways using the three methods. Direct
elements are easier to analyse because they are mainly estimated using the standard
procedures. Since, the indirect costs were estimated with different approaches and comprise
a larger portion of the overall product cost, it warrants investigation. The company’s
method termed indirect costs as overheads. Factory expenses were considered using the
Jung’s model and are part of the indirect costs. However, the machining costs have to be
considered also for the total indirect costs. Processing cost, material – dependent cost and
production overheads comprise the total indirect cost using the Hybrid Model. Since, the
three methods divide the indirect costs into different elements; the best way to make an
analysis is to consider the overall indirect cost or total overheads instead of considering any
sub-levels.
In order to compare the estimated overheads using the three methods, the actual values of
overheads incurred by different products are required. The actual values were obtained
206
from the accountancy data at the end of a financial year and are allocated to individual
products based on the ABC method. Activity rates are set and the activity times are based
on the recorded values during a manufacturing phase. Although, the exact method was not
disclosed by the company, the allocated overhead values for individual products were
supplied.
The comparison of the estimated values for overheads using the three methods against the
actual values is presented in Figure 7-10. The comparison is based on analysing the
estimation error trends for the given product range. The error trends represent the
percentage deviations from the actual costs. The solid lines represent the estimation error
trends for the Hybrid Model. On almost every occasion, the estimation error produced by
the Hybrid Model is not only lower than those presented by the other two methods but
remains in acceptable limits. The company’s method largely overestimated overheads for
individual products based on a cautious approach, whereas, huge fluctuations can be
observed in the estimation errors for the product range using the Jung’s model. It is also
noticeable that the estimation errors for overheads are greater than those for the overall
product costs using all the methods. This refers to the complexities in predicting overheads
accurately. The other elements must also be more accurately predicted such that the overall
results stay more accurate than those for just overheads. Despite that the overheads
predicted by the Hybrid Model stay within reasonable limits unlike those predicted by the
other two methods.
207
Figure 7-10: Estimation error trends for overheads using the three methods (2003 – 2005)
208
Due to the inconsistent results shown by the two methods for the individual products, the
cumulative overhead values are also analysed in Figure 7-11. The results are based on
comparing the estimated and actual overhead values at the cumulative level using the three
methods. The effects of adjustment result in the overall underestimation of overheads
predicted by the Jung’s model by around 15 percent and those for the company’s method
result in an overestimation of up to 32 percent. The values for the Hybrid Model varied
between -12 and 4 percent. The results clearly demonstrate the superiority of the Hybrid
Model over the other two methods for overheads estimation at both individual and
cumulative levels of estimation.
Estimation errors for cumulative overhead (%)
40
30
Estimation variation (%)
20
10
2003 2004 2005

-10
-20
Production Year
Estimation error using company's method (%) Estimation error using hybrid model (%)
Estimation error using Jung's model (%)
Figure 7-11: Overheads estimation analysis for the cumulative values
209
The above analysis leads to the quantification of the optimization achieved in the
estimation of overheads as a result of using the Hybrid Model in comparison to the other
two methods. The results are presented in Figure 7-12 (a) & (b) along with linearization.
The optimization is based on considering the deviation values from the actual costs without
considering the negative signs (meant for representing underestimation). The differences in
the values of error estimation of the two methods against those shown by the Hybrid Model
presented in the respective years result in the optimization values. For example, the
overhead estimation values predicted by the Hybrid Model in 2004 were almost 29 percent
more accurate as compared to those estimated by the company’s method. The optimization
of around 14 percent was also achieved against the Jung’s model estimated values for
overheads. For the purpose of statistical analysis, the linearized results are also incorporated
and demonstrate linear trends of the optimization for the investigation period.
Every aspect of the validation analysis demonstrated the superiority of the developed
Hybrid Model for PCE. Individual product’s estimated costs and the cumulative costs were
analysed and checked for not just accuracy but consistency of the results. The analysis for
overhead estimation at product’s level and cumulative level also demonstrated the
developed method’s superiority.
210
Overhead estim ation optim ization based on Com apny's m ethod (a)
30 28.79
24 23.97
% Optimization
y = 5.58x + 10.70
18
12 12.82
0
2003 2004 2005
Production Year
Estimation Optimization (%) Linear estimation optimization trend
Overhead estim ation optim ization based on Jung's m odel (b)
15
14.10
11.41
% Optimization
10
y = 4.91x + 0.11
5
4.28
0
2003 2004 2005
Production Year
Estimation Optimization (%) Linear estimation optimization trend
Figure 7-12: Optimization achieved for overhead estimation based on (a) the Company’s method (b)
the Jung’s model
211
7.6 Cost breakdown analysis
One of the advantages of incorporating the breakdown approach in the Hybrid Model is its
ability to present a cost breakdown analysis that could lead to the identification of the cost
elements requiring better cost control. A cost breakdown analysis is also presented in this
study based on the actual costs incurred. All the three methods discussed here consider
different breakdown structures. However, since the Hybrid Model is already validated for
its accuracy and consistency and found to be superior to the other two models, it will be
applied to the actual costs to get the breakdown statistics and for further analysis.
Figure 7-13 shows the production cost (actual) breakdown analysis from 2002 to 2005.
Overheads refer to the total indirect costs and are made up of processing costs, material-
dependent costs and production overheads. The values are based on the cumulative
expenditure in respective years for a given element and are represented in percentages. The
breakdown can be considered a typical representation of the UK manufacturing sector
based on the analysis of a representative case. Material cost is found to be the highest share
(47 percent on the average) of the overall product cost followed by overheads and then
labour costs with 37 and 16 percent average values respectively. It can be noted that labour
costs in UK are significantly higher than those obtained from the South Asian region.
Overhead is a big share of the overall product cost and its accurate estimation is important
to the overall accuracy of an entire product’s cost.
212
Figure 7-13: Production cost (actual) break down analysis (2002 – 2005) presented in values and
percentage
213
An analysis of the three elements based on yearly trends is presented in Figure 7-14. The
values represent the cumulative costs incurred on material, labour, and overheads. A
gradual yearly rise for the direct elements (material and labour) is not surprising and is due
to inflation and other variants. A steeper rise in overheads needs further investigation.
However, one explanation could be a less-effective control on overheads resulting in a
steeper rise. This is due to the company’s adopted cautious approach of overheads
estimation resulting in overestimation and thus setting cost control strategies accordingly.
Production Cost Element Trends
1600
Thousands
1400
1200
1000
Values (£s)
800
600
400
200
0
2002 2003 2004 2005
Production Year
Total Overheads (actual) Total material cost (Actual)
Actual Labour Cost (Total)
Figure 7-14: Production cost elements trends
214
Before analysing in detail why overheads rose more steeply, it is necessary to see their
effect on the overall production and manufacturing cost. Figure 7-15 (a) gives out the
yearly production and manufacturing costs in million £. Part (b) considers the effect of
elemental costs on these values. It is clear that production and manufacturing costs kept
rising relative to a steeper rise in overheads. Although material and labour costs are
proportionately higher in the overall product cost, their effect on the overall manufacturing
and production costs remained unaffected throughout the duration of the analysis.
However, a direct effect of a steep rise in overhead values on production and manufacturing
cost necessitates its further analysis.
Material cost is directly associated with design attributes and although cost control aspects
consider it, the discussion is beyond the scope of the thesis. Similarly, labour cost control
require effective process planning and operations management strategies, the scope of the
study (cost estimation) does not cover such discussions. However, an analysis of the
overhead values could lead to the identification of the elements with higher cost values in
order to keep in place an effective CCS. Figure 7-16 is a breakdown of the actual overheads
based on the Hybrid Model application. Production overheads make the largest portion with
an average value of 44 percent from 2002 to 2005. Processing cost and MDC make up 31
and 25 percent respectively. The figure also describes the values for the four-year period.
215
Production and Manufacturing Costs (2002 - 2005) (a)
Millions
3.20
2.90
3
2.53 2.60
2.41
Values (£s)
2.19
2.15
2 1.89
1
2002 2003 2004 2005
Production Year
0
Total Manufacturing Cost (Actual) Total production cost (actual)
Cost elements effect on production & manufacturing costs (b)
3500
Thousands
3000
2500
Values (£s)
2000
1500
1000
500
0
2002 2003 2004 2005
Production Year
Total Overheads (actual) Total material cost (Actual)
Actual Labour Cost (Total) Total Manufacturing Cost (Actual)
Total production cost (actual)
Figure 7-15: (a) Production and manufacturing costs (b) elemental costs effect on production and
manufacturing costs
216
Figure 7-16: Overheads break down analysis (2002 – 2005) presented in values and percentage
217
The analysis of the three sub-elements presented in Figure 7-17 shows a steeper rise in the
values for production overheads after 2003. While the processing costs displayed a steady
yearly rise for the four – year period, MDC were more controlled from 2003 – 2005.
Overhead Elements Trends
700
Thousands
600
500
Values (£s)
400
300
200
100
0
2002 2003 2004 2005
Production Year
Material Depandent Cost Processing Cost Production Overheads
Figure 7-17 Overhead elements trends (2002 – 2005)
A further breakdown of the largest overhead element (production overhead), into its
constituent elements is presented in Figure 7-18. General administration costs ranged from
40 to 49 percent of the overall production overheads and mainly remained the largest
constituent element. Selling expenses with 35 to 42 percent and computer related costs with
15 to 18 percent of the production overheads were the other two constituents.
218
Figure 7-18: Production overheads breakdown (2002 – 2005)
An analysis of the cost trends for the constituents presented in Figure 7-19 clearly shows
that selling expenses increased rapidly and became the largest constituent in 2005 leaving
behind general administration costs. General administration costs remained the largest
value constituent until 2004 but its rise remained gradual throughout four-year period.
Similar results were noted for computer-related costs. Although, the analysis revealed the
219
selling expenses as a possible area for cost control, a steeper rise in its values in the later
half of the trend was partly a result of the business policy decisions to boost sales.
Production Overhead Elements Trends
300
Thousands
250
200
Values (£s)
150
100
50
0
2002 2003 2004 2005
Production Year
Computer related cost General administration cost Selling Expenses
Figure 7-19: Production overhead elements trends analysis
A similar analysis can be carried out for the constituent elements for processing costs and
MDC. For example, the one carried out for processing costs in Figure 7-20 reveals a sharp
rise in energy costs. Repair costs were kept under control and went down between 2003 and
2004. Maintenance costs showed a steadier trend. Financing expenses can also be seen to
be kept under control. It is clear that energy costs were a major contributor to the
processing cots.
220
180
Thousands
Processing Cost Elements Trends
160
140
120
Values (£s)
100
80
60
40
20
0
2002 2003 2004 2005
Production Year
Insurance cost Financing Expenses Maintenance Cost
Repair Cost Energy cost
Figure 7-20: Processing cost elements trends
An analysis for the MDC constituent elements presented in Figure 7-21 shows that freight
and transportation costs could have been better controlled especially between 2002 and
2003 and then 2004 and 2005. Packaging costs between 2003 and 2004 could also be
controlled more efficiently. Costs associated with material handling, inspection and scrap
were largely kept under control.
221
MDC Elements Trends

90
Thousands 80
70
60
Values (£s)
50
40
30
20
10
0
2002 2003 2004 2005
Production Year
Indirect material Stores & Inventory control Freight & Tranportation
Material Inspection Purchasing Packaging
Material handling Scrap
Figure 7-21: MDC elements trends
The cost element analysis presented is a result of the Hybrid Model’s application and can
be effectively used to identify high cost elements providing opportunities for better cost
control. In that sense, the developed model can not only furnish early cost estimates for an
entire product with accuracy and consistency but provide cost control opportunities by
identifying high cost elements.
222
7.7 Conclusions
This chapter presented the validation analysis that comprised the comparisons of the actual
costs of the given product range against the estimated costs for the same range given by the
developed Hybrid Model, the company’s method and a published model. The Jung’s model
was selected as a representative case from the published domain being a proven case that
could be implemented within the framework of the selected manufacturing set up. Every
level of the analysis performed demonstrated the superiority of the developed model.
The cost estimates for individual products given by the Hybrid Model were found to be
more accurate and more consistent. The company’s method adopted a cautious approach
and as a result overestimated on almost every occasion with as much as 40 percent
estimation error on isolated cases. Although the Jung’s model predicted slightly better
values for an isolated case, the Hybrid Model’s prediction was not only within the
acceptable limit for that case but the overall accuracy was found to be more consistent. The
incorporation of the case-based technique into the modelling framework sensitized the
model without compromising its robustness. The overall validation analysis spread over a
three-year period covering a range of products under varying conditions and huge
fluctuations of cost deviations still produced consistent results within acceptable limits of
error estimations.
Cumulative cost analysis showed similar results. The company’s method overestimated by
as much as around 19 percent and the Jung’s model underestimated by around 9 percent at
223
the cumulative level. The Hybrid Model remained close to the actual costs even at the
cumulative level of prediction. The quantification of the optimization against the two
methods showed that a maximum of over 16 and 8 percent optimizations were obtained
against the company’s and the Jung’s methods respectively.
Another aspect of the validation analysis considered analysing the overhead estimation
accuracy at product and cumulative levels. On almost every occasion, the estimation error
(overheads for individual product) produced by the Hybrid Model was found to be not only
lower than those presented by the other two methods but remained in acceptable limits. The
cumulative overhead values were also analysed and resulted in the overall underestimation
of overheads predicted by the Jung’s model by around 15 percent and those for the
company’s method in an overestimation of up to 32 percent. The values for the Hybrid
Model varied between -12 and 4 percent.
A cost breakdown analysis was also presented for a four-year period (2002 – 2005) based
on the actual costs incurred. Material cost was found to be 47 percent; labour costs 37
percent and overheads 16 percent of the overall production cost. A further analysis of the
overheads breakdown values was carried out. The cost element analysis presented was a
result of the Hybrid Model’s application and helps to identify high cost elements providing
opportunities for better cost control.
224
Chapter 8 Conclusions
Chapter 8 concludes the thesis by summarizing the overall work and contributions. The
methodology and approaches developed in this thesis are evaluated by comparing the aims
and objectives of the work set at the beginning of the study with the outcome of the
research. The chapter describes current and future trends in the field of cost estimation and
describes how the developed model conforms to those trends. The chapter also identifies
future research avenues within the specific context of the established research.
8.1 Summary
The main aim of the study was to develop a comprehensive methodology for product cost
estimation in a batch type manufacturing environment. In order to achieve that, a genuine
theoretical development in the area was maintained followed by industrial trials and
validation. The overall process of the theoretical development was based on identifying
problems that the cost estimators face with the existing methods of PCE. This led to an
extensive review of the existing methods for PCE and identification of possible solutions.
The objectives were set in light of the possible solutions and were followed up with a
careful plan throughout the study. New theories, methods, techniques and mathematical
models were developed in line with the objectives. The developed models were validated
225
Chapter 8: Conclusions
through industrial trials due to the practical nature of the overall research area. The overall
research work was carefully structured in the thesis.
Chapter 1 described the scope of the research study. To do that, an overview of the research
area was provided. It was highlighted that a good pricing system requires the application of
cost engineering. The significance of cost estimation on cost control was discussed. The
role of cost estimation in the set up of a business enterprise with reference to maximizing
its profits was then noted. The chapter outlined and briefly discussed the problems in the
area of PCE. The aims and objectives of the research study were set and briefly discussed.
Finally, the chapter described the overall structure of the thesis.
Chapter 2 established background information in the area of cost estimation. The concepts
of cost and costing were introduced and the implications of cost control on manufacturing
systems were discussed. Three kinds of manufacturing systems were discussed: mass, batch
and job shop production. The rationale for developing a cost estimation methodology for a
batch type manufacturing environment was then presented. The chapter also presented a
comprehensive literature review with a focus on the techniques for different applications of
cost estimation. One of the application area noted was for generic systems. It was found out
that developing a methodology for generic systems could widen the scope of the developed
methodology. It was, therefore, decided to develop the methodology for generic systems
applications. Since, a batch type manufacturing environment is also a form of generic
system, and was already considered a viable option to develop a methodology for; a
226
decision was made to develop a methodology to suit the needs of a batch type environment
to satisfy the needs of a generic system also.
Chapter 3 extensively reviewed the literature on manufacturing and product cost estimation
and provided a detailed survey along with a critical evaluation of some of the techniques
developed in the area. Based on the similarities and the differences of the existing methods,
a comprehensive hierarchical classification system was suggested. The developed system
along with the key conditions for the application of a methodology from a given category
helped in developing a decision support model. The support model was developed with the
aim of assisting designers and estimators in selecting a methodology. During the course of
the classification system development, a framework for case-based methodology was
proposed in order to make an effective use of past product details. The proposed
classification system was broadly categorized in qualitative and quantitative techniques.
Qualitative techniques were found to be useful in providing early estimates and quantitative
techniques were noted for their accuracy. The chapter concluded with the suggestion of
developing a methodology for cost estimation based on combining concepts from both
qualitative and quantitative techniques for early and accurate estimation of a product’s cost.
Chapter 4 formed the basis for developing a PCE modelling methodology by establishing
an overhead estimation method. In order to fulfil the objectives of developing a method for
predicting an entire product’s cost, a cost breakdown approach was considered. An analysis
of the existing breakdown techniques identified overhead as a key element for the overall
accuracy of a product cost. Problem identification with the existing method of overhead
227
estimation led to the development of a new overhead estimation methodology based on
material– and time–dependent costs. The proposed methodology was validated in a batch
type manufacturing environment in South Asia. The validation analysis was based on
comparing the estimated product costs obtained with the application of the proposed
methodology and the company’s method against the actual product costs. The proposed
model for overhead estimation was found to be superior to the company’s method of
overhead estimation as the former allowed more accurate estimation of product costs. The
chapter also revealed manufacturing cost breakdown statistics at the company. The cost
breakdown can be considered a representative case for that region characterized by lower
wage rates. The chapter finally discussed room for further optimization in the proposed
method.
Chapter 5 provided a complete framework for estimating a product cost early and
accurately in a batch type manufacturing environment. The overhead method developed in
Chapter 4 formed the basis for extending the work. The limitations identified in the already
developed method were overcome. In order to make use of the attributes of the qualitative
and the quantitative techniques, a hybrid modelling approach was considered. Case-based
approach from the qualitative techniques was selected to facilitate an effective use of past
product details to allow an early estimation of new products costs. This was achieved by
making use of the past product costs obtained at the end of a given year to predict the costs
in the following year. Two methods from the quantitative techniques were selected. The use
of breakdown approach allowed the estimation of an entire product’s cost instead of only
part or component costs. The activity-based costing method was incorporated to get
228
accurate results. Although, the activity rates and units were modelled, the modelling
framework was established on time– and material–dependent cost elements giving the
Hybrid Model even better accuracy than the conventional ABC systems.
Chapter 6 was based on the industrial implementation and application of the Hybrid Model
in a batch manufacturing environment in the UK selected as a representative case of the
manufacturing sector in the country. The implementation presented the procedure of using
the Hybrid Model. This was achieved by developing a comprehensive Hybrid Model
Implementation (HMI) algorithm. The algorithm provided all necessary steps in logical
sequence to implement the model. The provisions of primary and secondary data input to
the system were accommodated in the algorithm along with inclusions of the Hybrid
Model’s comprehensive set of equations. Equations for cost deviation indices were
introduced in order to effectively utilize past data. The algorithm successfully implemented
the model and the cost estimate results were obtained. Key benefits of using the algorithm
and the Hybrid Model were demonstrated. The cost estimates from the company’s method
were also obtained. The implementation phase facilitated the generation of cost estimates
and is part of the entire validation process. The estimated results were obtained
retrospectively from 2003 to 2005 by using the field data from 2001 to 2004.
Chapter 7 was based on a follow up of the implementation phase and compared the results
obtained in the preceding chapter in order to validate the Hybrid Model. The Jung’s model
was also selected as a representative case from the published domain based on its already
established validity within its domain and its compatibility with the field data and the
229
selected manufacturing set up. Estimated cost values using the Jung’s model were obtained.
All the estimated cost values for the given product range were compared against the actual
costs for the same range. The estimated costs for that purpose were obtained by the
application of the Hybrid Model, the company’s own method and the selected published
model. Every level of the validation analysis demonstrated the superiority of the developed
model. For example, the costs for individual products were more accurately and more
consistently estimated using the Hybrid Model. Cumulative production costs were also
estimated more accurately and more consistently. The optimization of over 16 and 8
percent were obtained against the company’s and the Jung’s methods respectively by using
the Hybrid Model. The developed model was also checked for its overhead estimation
accuracy at product and cumulative production levels. The optimization achieved for
overhead estimation was found to be up to 29 percent against the company’s method and 14
percent against the Jung’s model.
A cost breakdown analysis was also presented for a four-year period (2002 – 2005) based
on the actual costs incurred. The breakdown values revealed can be considered a typical
representation of the UK manufacturing. A further analysis of the overhead breakdown
values was also carried out based on the Hybrid Model’s application. The analysis proved
helpful in identifying the high cost elements thus providing opportunities for better cost
control. The developed model, therefore, forms part of a cost control system development
also.
230
8.2 Contributions
In order to evaluate the outcomes from this thesis, the contributions from the overall
research work are outlined below.
8.2.1 Development of a technique classification system
Following a comprehensive literature review in the area of PCE, methodologies were
categorised into qualitative and quantitative. An extensive analysis of the categorized
techniques led to further sub-division of up to four levels and thus resulting in a
comprehensive hierarchical classification system. It was found that qualitative techniques
mainly focussed on early estimates and quantitative ones delivered accuracy by making use
of product design and process planning details. The developed system was found helpful in
identifying that an accurate overhead estimation methodology is essential for accurate
estimation of a product’s cost. The proposed system is also helpful in visualizing the nature
of an estimation technique and comparing it with the other from the same category or a
different one.
8.2.2 Development of a decision support model (DSM)
Following a comprehensive literature review and the developed classification system, key
conditions for the implementation of a technique from a given category were identified.
This led to the development of a decision support model. The aim of the system is to assist
231
designers and estimators in selecting an estimation methodology to suit the needs of a
system and given conditions. This is particularly necessary when several cost estimation
techniques may be applied in a given condition. However, due to the availability of a large
number of estimation techniques, estimators either find the selection process time-
consuming or erroneous. The developed model overcomes the limitations by providing
consistency in the selection process.
8.2.3 Development of time- and material-based overhead estimation
methodology
In order to develop a methodology that could predict an entire product’s cost, breakdown
approach was considered. Following the review and the proposed classification system,
overhead was identified as an important breakdown element. The significance of accurate
estimation of this element led to the identification of any existing methods for its
prediction. The ABC system was found to overcome some of the problems but still left few
areas for improvement. In light of the identified problems with the existing method of
overhead estimation, an improved overhead estimation methodology was proposed based
on time– and material–dependant cost elements. The industrial validation based on a four-
year retrospective analysis in a batch manufacturing environment in South Asia revealed
the superiority of the developed methodology against a representative method from the
manufacturing domain.
232
8.2.4 Development of a PCE methodology for batch production
A main objective of the study was to develop a methodology that could predict an entire
product’s cost early and accurately. This was achieved by developing the Hybrid Model for
PCE in a batch type environment. The model was based on combining attributes from
qualitative and quantitative cost estimation techniques from the proposed classification
system. Case-based approach from qualitative techniques provided means of utilizing past
cost details in order to furnish estimates for similar new products in future. It allowed the
early estimation by eliminating the need for product design details necessary to furnish
estimates from scratch. The already developed methods for overhead estimation (TRO and
MRO) provided basis for further improvement by incorporating the concepts of improved
ABC system features. Overheads were sub-divided and calculated based on improved
concepts of activity rates and activity units. In this way, the accuracy attributed to
quantitative techniques was achieved by making use of the improved and modified
concepts of the ABC system. The use of breakdown approach not only allowed the
estimation of the entire product cost, but the representation of the cost into elements and
sub-elements.
The developed model can also predict manufacturing cost in addition to the overall product
cost. The representation of a product cost into its elements and sub-elements allows the
identification of any high cost areas. The identification in return facilitates cost control
opportunities. The developed model in that sense goes beyond just furnishing early and
233
accurate cost estimates for a product. It constitutes a key component of a cost control
system.
8.2.5 Development of cost deviation indices
The introduction of the case-based framework in the Hybrid Model was facilitated by cost
deviation indices. The indices were based on effectively utilizing past two years data to
predict future costs. Inflation and other cost deviations were also incorporated in the indices
for optimized use of past data and accurate prediction. This also resulted in overcoming the
limitations of TRO and MRO methods. Deviation indices are crucial to the overall accuracy
of the estimated results from the developed Hybrid Model.
8.2.6 Development of HMI algorithm and industrial implementation
Development of the PCE Hybrid Model could serve no real purpose if it could not be
successfully implemented in industries. One of the objectives was to implement the
developed model in an industrial environment. This thesis devised a comprehensive
procedure for the industrial implementation of the developed PCE Hybrid Model. This
resulted in the development of HMI algorithm to facilitate the implementation procedure.
The Hybrid Model was subsequently implemented in the selected UK batch manufacturing
industry. The Hybrid Model with the help of the HMI algorithm and deviation indices
generated the estimation results for the given product range. The HMI algorithm after
minor changes can be made more generic for a wider scope of industrial implementation.
234
8.2.7 Comparison and validation
The validity of the developed model could only be established after the comparisons of its
generated results against the ones from the already recognized and validated methods.
Therefore, the model was validated after a three-year retrospective industrial
implementation trial. The cost estimation results that were obtained after a successful
industrial implementation were not only compared against an industrially recognized and
representative method but against an already established published model. The results from
the Hybrid Model were found to be more accurate as the furnished estimated costs were
found to be closer to the actual product costs than those predicted by the other two
methods. However, the validation analysis went even further ahead by comparing not just
the product costs but the cumulative costs and the elemental costs. Every aspect of the
comparison analysis validated the Hybrid Model as the costs estimated were not only more
accurate but more consistent than those predicted by the other two methods.
8.2.8 By – products
The overall study resulted in not just achieving the set objectives but in delivering valuable
by-products.
A comprehensive literature review carried out in the field not only presented a critical
analysis of the developed techniques with the key advantages and limitations but provided
useful work with reference to the applications in the area. Similarly, the developed case-
235
based model serves as a valuable component for not just the Hybrid Model but provides a
framework for developing a data retrieval system for other applications of production
planning and control system. The developed model makes use of past product details in
order to furnish ones for similar new designs. Only the changes in the new designs can be
incorporated thereby eliminating the need to furnish details from scratch. Another valuable
outcome of the study was the revelation of statistical cost breakdown figures on two
different geographical locations that were typical of the regions and reflected contemporary
trends. The figures reflect the geographical implications on cost occurrences and are helpful
in creating a better understanding for developing an effective CCS.
The overall contributions resulted in theoretical development to the field of PCE. This
development was fully backed up by industrial trials and validation analyses. All in all the
work presented in the thesis resulted in some pioneering contributions to the field of cost
engineering in general and PCE in particular.
8.3 Current Trends and Future Work
Before exploring the possibilities for further research avenues in the already established
research domain, it is befitting to analyse how it conforms to the current and future trends
in the area.
236
With the advent of computers and the advancement in technology, non-conventional
approaches such as knowledge-based techniques and neural network models have been
applied effectively utilizing past knowledge to predict the future costs in the early design
phases. Current trends in cost estimation exploit the feature technology and a simpler trend
is based on estimating the costs by calculating the amount of activities performed to
manufacture a product. However, more recent works in the field focus on getting quicker
and more accurate results by developing integrated systems combining two or more
approaches. For example, a mix of neural network approach and feature technology is an
emerging trend. Applying neural networks in CBR [112], cost-tolerance analysis using
neural networks [113] and fuzzy activity-based costing [114] are also among some of the
new concepts. Yet another area of ongoing research activities combines rule-based, fuzzy
logic based, and feature-based methodologies together. An approach that blends some of
these techniques could provide more promising results. For example, there is a need to
combine the feature technology with the ABC method to study the effects in detail. An
approach dealing with the ABC systems and neural networks at the same time may yet be
another research area.
In line with the recent and ongoing research trends in the area of PCE, the developed
methodology exploits the benefits of the three individual techniques from two branches
(qualitative and quantitative) to develop a Hybrid Model that conforms to the current and
future trends. The established research, therefore, opens new avenues for exploring further
in the field. Following are some of the suggested projects.
237
Validation and optimization analysis of the Hybrid Model for PCE through
simulation
The aim is to analyze and investigate thoroughly the behaviour of different parameters of
the models under varying conditions and compare the results. This should lead to further
optimization of the Hybrid Model.
Development of a data retrieval system based on the case-based methodology
The aim of the proposed project is to develop a matching algorithm to facilitate the
adaptation of a past product design with closest similarities to a new design. The developed
algorithm will form part of the overall data retrieval system by interfacing it into the
already proposed case-based system.
The development of a comprehensive prototype computer-based cost estimation
system based on the developed Hybrid Model
The already established research when combined with the elements from the above
mentioned proposed projects could lead to the development of a comprehensive prototype
computer-based cost estimation system. The optimization analysis through simulation and
the development of a data retrieval system based on the proposed case-based approach
would form integral elements for the proposed computer-based system. The project should
aim at interfacing the system with product design data-base. The aim of such an integrated
238
system would be to generate product cost estimates for new product designs at design
conceptualization stages. Such a system would facilitate designers, estimators, planners and
managers to make necessary decisions from design changes and price quotes to devising
production planning and control strategies. More strategic decisions like make or buy and
business resizing etc. would also be possible.
8.4 Concluding remarks
The validation of the Hybrid Model demonstrated through various aspects of the analyses
can be attributed to a number of factors. A comprehensive study of the available techniques
through a well-structured literature review helped to identify the problem areas. This led to
the identification of ways to develop solutions based on an effective utilization of strengths
of some of the best known methods and eliminating their shortcomings. The initial methods
proposed as part of this thesis that were already found to be more accurate than some of the
existing techniques, were further analysed. As a result, a framework for a more
comprehensive methodology based on further optimizations was presented. The self-
assessment procedure fine-tuned the developed methodology.
The work presented in the thesis resulted in contributions to the field of PCE and forms part
of the pioneering work at King’s College London. The encouraging results from the
industrial validation analysis open avenues for further exploration in the field. The
established work is also an attempt to provide a platform for researchers and practitioners
239
alike to explore further in the field. For example, Professor Frank J Fabozzi of Yale School
of Management (USA) and editor of the Journal of Portfolio Management not only cited
the published work as excellent but incorporated the article as a basic chapter of his
recently published book [115]. Finally, the overall work, therefore, provides contributions
to the field of applied engineering and opens up new channels for further explorations.
240
Publications Arising from the PhD Study
The research carried out as part of the current PhD study and covered in the thesis has
resulted in the following leading journal publications:
• Niazi A, Dai JS, Balabani S, and Seneviratne L. Product cost estimation:
Technique classification and methodology review. Transactions of the ASME.
Journal of Manufacturing Science and Engineering, May-2006, 128(2), 563-
75, Publisher: ASME, USA, 2006.
• Niazi, A., Dai, J. S., Balabani, S., and Seneviratne, L. D. A new overhead
estimation methodology: A case-study in an electrical engineering company.
Proceedings of the Institution of Mechanical Engineers, Part B: Journal of
Engineering Manufacture, Jan-2007, 221: 699-710, 2007.
• Niazi, A., Dai, J. S., and Balabani, S. PCE Hybrid Model for cost estimation in
a batch type manufacturing environment. Transactions of the ASME. Journal
of Manufacturing Science and Engineering, 2007, (submitted).
• Niazi, A., and Dai, J. S., HMI Algorithm and industrial implementation
framework for PCE Hybrid Model. Transactions of the ASME. Journal of
Manufacturing Science and Engineering, 2007, (submitted).
241
Publications Arising from the PhD Study
• Niazi, A., and Dai, J. S., Methodology comparisons and validation analysis for
PCE Hybrid Model. Transactions of the ASME. Journal of Manufacturing
Science and Engineering, 2007, (submitted).
242
References
[1]. Clark FD and Lorenzoni AB. Applied Cost Engineering, New York, M. Dekker; 1985.
[2]. Cavalieri S, Maccarrone P and Pinto R. Parametric vs. neural network models for the
estimation of production costs: A case study in the automotive industry. International
Journal of Production Economics, 91(2): 165-177, 2004.
[3]. Rehman S and Guenov MD. A methodology for modelling manufacturing costs at
conceptual design. Computers & Industrial Engineering, 35(3-4): 623-626, 1998.
[4]. Gayretli A and Abdalla HS. A feature-based prototype system for the evaluation and
optimisation of manufacturing processes. Computers & Industrial Engineering, 37(1-
2): 481-484, 1999.
[5]. Pahl G and Beitz W. Engineering Design: A Systematic Approach, (2nd edition),
Springer; 1999.
[6]. OuYang C and Lin TS. Developing an integrated framework for feature-based early
manufacturing cost estimation. International Journal of Advanced Manufacturing
Technology, 13(9): 618-629, 1997.
243
References
[7]. Sheldon DF, Perks R, Jackson M, Miles BL, and Holland J. Designing for whole life
costs at the concept stage. Journal of Engineering Design, 1(2): 131-45, 1990.
[8]. Hicks BJ, Culley SJ, and Mullineux G. Cost estimation for standard components and
systems in the early phases of the design process. Journal of Engineering Design,
13(4): 271-292, 2002.
[9]. Sheldon D, Huang G, and Perks R. Specification and development of cost-estimating
database for engineering design. Design for Manufacturability, 52: 91-96, 1993.
[10]. Bode J and Fung RYK. Cost engineering with quality function deployment.
Computers & Industrial Engineering, 35(3-4): 587-90, 1998.
[11]. BenArieh D and Li Q. Activity-based cost management for design and development
stage. International Journal of Production Economics, 83(2): 169-183, 2003.
[12]. Feng C, Kusiak A, and Huang C. Cost evaluation in design with form features.
Computer-Aided Design, 28(11): 879-885, 1996.
[13]. Gupta S, Nau DS, Regli WC, and Zhang G. A methodology for systematic generation
and evaluation of alternative operation plans. In: Shah JJ, Mantyla M, and Nau DS,
editors. Advances in Feature Based Manufacturing, Amsterdam, Elsevier Publishing
Company; 161-184, 1994.
244
References
[14]. Kiritsis D, Neuendorf KP, and Xirouchakis P. Petri net techniques for process
planning cost estimation. Advances in Engineering Software, 30(6): 375-387, 1999.
[15]. Aldrich RL. Costing a bill of material/route of thermoformed part using integrated
MRPII software. 1995.
[16]. Roztocki N and Needy KL. Integrating activity-based costing and economic value
added in manufacturing. Engineering Management Journal, 11(2): 17-22, 1999.
[17]. Park CS and Son YK. Computer-assisted estimating of non-conventional
manufacturing costs. Computers in Mechanical Engineering, 6(1): 16-25, 1987.
[18]. Diplaris SC and Sfantsikopoulos MM. Cost–tolerance function: A new approach for
cost optimum machining accuracy. International Journal of Advanced Manufacturing
Technology, 16(1): 32-38, 2000.
[19]. Weustink IF, Ten Brinke E, Streppel AH, and Kals HJJ. A generic framework for cost
estimation and cost control in product design. Journal of Materials Processing
Technology, 103(1): 141-148, 2000.
[20]. Koonce D, Judd R, Sormaz D, and Masel DT. A hierarchical cost estimation tool.
Computers in Industry, 50(3): 293-302, 2003/4.
245
References
[21]. Boothroyd G, Dewhurst P, and Knight W. Product Design for Manufacture and
Assembly, New York, Marcel Dekker, 1994.
[22]. Daabub AM and Abdalla HS. A computer-based intelligent system for design for
assembly. Computers & Industrial Engineering, 37(1-2): 111-115, 1999.
[23]. Dewhurst P and Blum C. Supporting analyses for the economic assessment of die-
casting in product design. CIRP Annals, 38(1): 161-164, 1989.
[24]. Luong, L. H. S. and Spedding TA. An integrated system for process planning and cost
estimation in hole making. International Journal of Advanced Manufacturing
Technology, 10(6): 411-415, 1995.
[25]. Ramirez JC and Touran A. An integrated computer system for estimating welding
cost. Cost Engineering, 33(8): 7-14, 1991.
[26]. Taiber JG and Ishii K, editors. Development of an optimization method for
determination of process sequences considering prismatic work pieces. Proceedings of
International Computers in Engineering Conference (1994) USA. ASME, Vol. 1, pp
(271-280), 1994.
246
References
[27]. Schreve K, Schuster HR, and Basson AH. Manufacturing cost estimation during
design of fabricated parts. Proceedings of the Institution of Mechanical Engineers,
Part B: Journal of Engineering Manufacture, 213(7): 731-735, 1999.
[28]. Boothroyd G and Radovanovic P. Estimating the cost of machined components during
the conceptual design of a product. CIRP Annals, 38(1): 157-160, 1989.
[29]. French MJ. Function costing: A potential aid to designers. Journal of Engineering
Design, 1(1): 47-53, 1990.
[30]. Ulrich KT and Eppinger SD. Product Design and Development, (3rd edition),
McGraw Hill, 2003.
[31]. Kendall K, Mangin C, and Ortiz E. Discrete event simulation and cost analysis for
manufacturing optimisation of an automotive LCM component. Composites Part A:
Applied Science and Manufacturing, 29(7): 711-720, 1998.
[32]. Gutowski TG, Henderson R, and Shipp C. Manufacturing costs for advanced
composite aerospace parts. SAMPE Journal, 27(3): 37-43, 1991.
[33].Mileham AR, Currie GC, Miles AW, and Bradford DT. A parametric approach to cost
estimating at the conceptual stage of design. Journal of Engineering Design, 4(2):
117-125, 1993.
247
References
[34]. Shing ON. Design for manufacture of a cost-based system for moulded parts.
Advances in Polymer Technology, 18(1): 33-42, 1999.
[35].Chen Y and Liu J. Cost-effective design for injection moulding. Robotics and
Computer-Integrated Manufacturing, 15(1): 1-21, 1999.
[36]. Lee P-and Sullivan WG. Integrating concurrent engineering and activity-based costing
in a knowledge-based support system. International Journal of Flexible Automation
and Integrated Manufacturing, 6(1-2): 1-28, 1998.
[37]. Ong NS. Manufacturing cost estimation for PCB assembly: An activity-based
approach. International Journal of Production Economics, 38(2-3): 159-172, 1995.
[38]. Ong NS and Lim LEN. Activity-based cost-modelling procedures for PCB assembly.
International Journal of Advanced Manufacturing Technology, 8(6): 396-406, 1993.
[39]. Arenz DR. Using AI to estimate developmental equipment costs. AACE Transactions,
F.2.1-F.2.5, 1991.
[40]. Zhang YF, Fuh JYH, and Chan WT. Feature-based cost estimation for packaging
products using neural networks. Computers in Industry, 32(1): 95-113, 1996.
248
References
[41]. Zhang YF and Fuh JYH. A neural network approach for early cost estimation of
packaging products. Computers & Industrial Engineering, 34(2): 433-450, 1998.
[42]. Poli C, Escudero J, and Fermandez R. How part design affects injection-moulding
tool costs. Machine Design, (101-104), November, 24, 1988.
[43]. Bernet N, Wakeman MD, Bourban PE, and Manson, J. A. E. An integrated cost and
consolidation model for commingled yarn based composites. Composites Part A:
Applied Science and Manufacturing, 33(4): 495-506, 2002.
[44]. Olofsson K and Edlund A. Manufacturing parameter influences on production cost.
International Conference on Advance Composites. SICOMP Technical Report TR 98-
009, 1998.
[45]. Mayer C, Hartmann A and Neitzel M. Cost-conscious manufacturing of tailored
thermoplastic composite intermediates using a double belt press. 18th International
SAMPE Europe Conference, Proceedings SAMPE Europe, (339-351), 1997.
[46]. Foley M and Bernardon E. Thermoplastic composite manufacturing cost analysis for
the design of cost effective automated system. SAMPE J, 6(4): 67-74, 1990.
249
References
[47]. Walls KO and Crawford RJ. The design for manufacture of continuous fibre-
reinforced thermoplastic products in primary aircraft structures. Composites
Manufacturing, 6(3-4): 245-254, 1995.
[48]. Aderoba A. A generalised cost-estimation model for job shops. International Journal
of Production Economics, 53(3): 257-263, 1997.
[49]. Lizhang Z and Burns G. Activity-based costing in non-standard route manufacturing.
International Journal of Operations & Production Management, 12(3): 38-60, 1992.
[50]. Ozbayrak M, Akgun M, and Turker AK. Activity-based cost estimation in a push/pull
advanced manufacturing system. International Journal of Production Economics,
87(1): 49-65, 2004.
[51]. Takakuwa S. The use of simulation in activity-based costing for flexible
manufacturing systems. Proceedings of the 1997 Winter Simulation Conference
(1997) USA. (793-800), 1997.
[52]. LaScola Needy K, Billo RE, and Warner RC. A cost model for the evaluation of
alternative cellular manufacturing configurations. Computers & Industrial
Engineering, 34(1): 119-134, 1998.
250
References
[53]. Wen Hsien T. Activity-based costing model for joint products. Computers &
Industrial Engineering, 31(3-4): 725-729, 1996.
[54]. Haedicke J and Feil D. In a DOD environment: Hughes aircraft sets the standard for
ABC. Management accounting, 72(8): 29-33, 1991.
[55]. Merz CM and Hardy A. ABC puts accountants on design team at HP. Management
accounting, 75(3): 22-27, 1993.
[56]. Miller SH. The view from inside: GM’s general auditor looks back. Journal of
Accountancy, 177(3): 44-46, 1994.
[57]. Soloway L, J. Government contracting: Using activity-based cost management
systems in aerospace and defence companies. Journal of Cost Management, 6(4): 56-
66, 1993.
[58]. Hobdy T, Thomson J, and Sharman P. Activity-based management at AT&T.
Management Accounting, 75(10): 35-39, 1994.
[59]. Porter T, J. and Kehoe J, G. Using activity-based costing and value analysis to take
the pain out of downsizing at a naval shipyard. National Productivity Review, 13(1):
115-125, 1994.
251
References
[60]. Ostwald PF. Engineering Cost Estimating (3rd Edition), Prentice Hall, 1991.
[61]. Clark F and Lorenzoni AB. Applied Cost Engineering, (3rd edition), CRC, 1996.
[62]. Brimson JA and National Association of Accountants. Activity Accounting: An
Activity-Based Costing Approach, New York, John Wiley and Sons, 1991.
[63]. Shehab EM and Abdalla HS. Manufacturing cost modelling for concurrent product
development. Robotics and Computer-Integrated Manufacturing, 17(4): 341-353,
2001.
[64]. Shehab EM and Abdalla HS. A design to cost system for innovative product
development. Proceedings of the Institution of Mechanical Engineers, Part B (Journal
of Engineering Manufacture), 216(B7): 999-1019, 2002.
[65]. Lihua X and Yunfeng W. Research on rapid cost evaluation based on case-based
reasoning. Computer Integrated Manufacturing Systems, 10(12): 1605-1609, 2004.
[66]. Ficko M, Drstvensek I, Brezocnik M, Balic J, and Vaupotic B. Prediction of total
manufacturing costs for stamping tool on the basis of CAD-model of finished product.
Journal of Materials Processing Technology, 164-165: (1327-1335), 2005.
252
References
[67]. Balarman V and Vattam SS; Sasikumar M, Rao DD, et al. (editors). Finding common
ground in case-based systems. Knowledge Based Computer Systems. Proceedings of
the International Conference: KBCS'98, (1998) India, (25-37), 1998.
[68]. Kingsman BG and De Souza AA. A knowledge-based decision support system for
cost estimation and pricing decisions in versatile manufacturing companies.
International Journal of Production Economics, 53(2): 119-139, 1997.
[69]. Shehab E and Abdalla H. An intelligent knowledge-based system for product cost
modelling. International Journal of Advanced Manufacturing Technology, 19(1): 49-
65, 2002.
[70]. Gayretli A and Abdalla HS. An object-oriented constraints-based system for
concurrent product development. Robotics and Computer-Integrated Manufacturing,
15(2): 133-144, 1999.
[71]. Ostwald PF. AM Cost Estimator: Cost Estimating Database, (4th edition), Ohio,
Penton Publishing, 1998.
[72]. Musgrove JG. Expert conceptual estimator. A multi-media expert system cost
estimating tool. New Generation Knowledge Engineering. IAKE '92 Proceedings of
Third Annual Symposium of the International Association of Knowledge Engineers
(1992) USA, 480-489, 1992.
253
References
[73]. Venkatachalam AR, Mellichamp CM, and Miller DM. A knowledge-based approach
to design for manufacturability. Journal of Intelligent Manufacturing, 4(5): 355-366,
1993.
[74]. Waring CW. Product costing automation: The impact of the learning curve.
Computers & Industrial Engineering, 21: 313-317, 1991.
[75]. Hundal MS. Design to cost. In: Parsaei HR, Sullivan WG, editors. Concurrent
Engineering: Contemporary Issues and Modern Design Tools, (1st edition), pp (330-
351), Springer, 1993.
[76]. Lewis J. Metrics mapping cuts estimating time. Design News, 55(18): 107-110, 2000.
[77]. McKim RA. Neural network applications to cost engineering. Cost Engineering,
35(7): 31-35, 1993.
[78]. Shtub A and Zimerman Y. A neural-network-based approach for estimating the cost
of assembly systems. International Journal of Production Economics, 32(2): 189-207,
1993.
[79]. ManYi C and DingFang C. Early cost estimation of strip-steel coiler using BP neural
network. Proceedings of 2002 International Conference on Machine Learning and
Cybernetics (2002) USA, 1326-1331, IEEE, 2002.
254
References
[80]. Hajare AD. Parametric costing—steel wire mill. Proceedings of the Annual
Convention of the Wire Association International, (172-178), 1998.
[81]. Roberts CA and Hermosillo EP. An automated machining cost estimator. Journal of
Engineering Valuation and Cost Analysis, 3(1): 27-42, 2000.
[82]. Boothroyd G and Reynolds C. Approximate cost estimates for typical turned parts.
Journal of Manufacturing Systems, 8(3): 185-93, 1989.
[83]. Dewhurst P and Boothroyd G. Early cost estimating in product design. Journal of
Manufacturing Systems, 7(3): 183-91, 1988.
[84]. Jong Yun J. Manufacturing cost estimation for machined parts based on
manufacturing features. Journal of Intelligent Manufacturing, 13(4): 227-238, 2002.
[85]. Wei Y and Egbelu PJ. A framework for estimating manufacturing cost from
geometric design data. International Journal of Computer Integrated Manufacturing,
13(1): 50-63, 2000.
[86]. Son YK. A cost estimation model for advanced manufacturing systems. International
Journal of Production Research, 29(3): 441-452, 1991.
255
References
[87]. Singh N. Integrated product and process design: A multi-objective modelling
framework. Robotics and Computer-Integrated Manufacturing, 18(2): 157-168, 2002.
[88]. Yeo SH, Ngoi, BKA, Poh LS, and and Hang C. Cost-tolerance relationships for non-
traditional machining processes. The International Journal of Advanced
Manufacturing Technology, 13(1): 35-41, 1997.
[89]. Sfantsikopoulos MM, Diplaris SC, and Papazoglou PN. Concurrent dimensioning for
accuracy and cost. International Journal of Advanced Manufacturing Technology,
10(4): 263-268, 1995.
[90]. Leibl P, Hundal M, and Hoehne G. Cost calculation with a feature-based CAD system
using modules for calculation, comparison and forecast. Journal of Engineering
Design, 10(1): 93-102, 1999.
[91]. Brimson JA. Feature costing: Beyond ABC. Journal of Cost Management,
January/February 1998: (6-12), 1998.
[92]. Wierda LS. Linking design, process planning and cost information by feature-based
modelling. Journal of Engineering Design, 2(1): 3-19, 1991.
[93]. Cooper Rand Kaplan RS. How cost accounting distorts product costs. Management
Accounting, 69(10): 20-27, 1988.
256
References
[94]. Hundal. Product costing: A comparison of conventional and activity-based costing
methods. Journal of Engineering Design, 8(1): 91-103, 1997.
[95]. Andrade MC, Filho RCP, Espozel AM, Maia LOA, and Qassim RY. Activity-based
costing for production learning. International Journal of Production Economics,
62(3): 175-180, 1999.
[96]. Beaujon GJ and Singhal VR. Understanding the activity costs in an activity-based cost
system. Journal of Cost Management, 4(1): 51-72, 1990.
[97]. Noreen E. Conditions under which activity-based cost systems provide relevant costs.
Journal of Management Accounting Research, 3: 159-168, 1991.
[98]. Borjesson S. What kind of activity-based information does your purpose require?:
Two case studies. International Journal of Operations & Production Management,
14(12): 79-99, 1994.
[99]. Malik SA and Sullivan WG. Impact of ABC information on product mix and costing
decisions. IEEE Transactions on Engineering Management, 42(2): 171-176, 1995.
[100]. Kee R and Schmidt C. A comparative analysis of utilizing activity-based costing
and the theory of constraints for making product-mix decisions. International
Journal of Production Economics, 63(1): 1-17, 2000.
257
References
[101]. Spedding TA and Sun GQ. Application of discrete event simulation to the activity-
based costing of manufacturing systems. International Journal of Production
Economics, 58(3): 289-301, 1999.
[102]. Kaplan RS. In defence of activity-based cost management. Management
Accounting, 74(5): 58-63, 1992.
[103]. Tornberg K, Jamsen M, and Paranko J. Activity-based costing and process
modelling for cost-conscious product design: A case study in a manufacturing
company. International Journal of Production Economics, 79(1): 75-82, 2002.
[104]. Tseng YJ and Jiang BC. Evaluating multiple feature-based machining methods
using an activity-based cost analysis model. International Journal of Advanced
Manufacturing Technology, 16(9): 617-623, 2000.
[105]. Yang YN, Parsaei HR, Hamid R, Leep HR, and Wong JP. A manufacturing cost
estimating system using activity-based costing. International Journal of Flexible
Automation and Integrated Manufacturing, 6(3-4): 223-243, 1998.
[106]. Ong NS. Activity-based cost tables to support wire harness design. International
Journal of Production Economics, 29(3), 271-289, 1993.
258
References
[107]. Karbhari VM and Jones SK. Activity-based costing and management in the
composites product realization process. International Journal of Materials and
Product Technology, 7(3): 232-244, 1992.
[108]. Koltai T, Lozano S, Guerrero F, and Onieva L. A flexible costing system for
flexible manufacturing systems using activity-based costing. International Journal
of Production Research, 38(7): 1615-1630, 2000.
[109]. DeGarmo EP, Black JT, and Kohser RA. Materials and Processes in
Manufacturing, (8th edition), New York, John Wiley & Sons, 1999.
[110]. Niazi A, Dai JS, Balabani S, and Seneviratne L. Product cost estimation: Technique
classification and methodology review. Transactions of the ASME, Journal of
Manufacturing Science and Engineering, 128(2), 563-75, 2006.
[111]. Boothroyd G, Dewhurst P, and Knight W. Product Design for Manufacture and
Assembly, (2nd edition), New York, Marcel Dekker, 2002.
[112]. Lotfy EA and Mohamed AS. Applying neural networks in case-based reasoning
adaptation for cost assessment of steel buildings. International Journal of
Computers & Applications, 24(1): 28-38, 2002.
259
References
[113]. ZoneChing L and DarYuan C. Cost-tolerance analysis model based on a neural
networks method. International Journal of Production Research, 40(6): 1429-1452,
2002.
[114]. Nachtmann Hand Lascola Needy K. Fuzzy activity-based costing: A methodology
for handling uncertainty in activity-based costing systems. Engineering Economist,
46(4): 245-273, 2001.
[115]. Frank J. Fabozzi, Pamela Peterson Drake and Ralph S. Polimeni, The Complete
CFO Handbook: From Accounting to Accountability, New Jersey, John Willey &
Sons, 2007.
260
Appendix A Bill of Material (BOM)
The purpose of appendix A is to elaborate BOM. This is carried out by taking the example
from a hammer drill. The drill is gradually dismantled and assemblies, sub-assemblies and
parts & components are identified. Material and material quantities for different parts are
identified. The information along the process is recorded and helps to establish BOM for
the drill.
A.1 Introduction
Bill of material (BOM) is a representation of different material with their respective
quantities used in a product. Such details are extracted from product design documents.
Details contained in design documents help in establishing a product structure based on
creating sub-assemblies, assemblies and eventually product levels.
In order to better understand BOM and product structure a hammer drill presented in Figure
A-1 is selected and dismantled gradually. The process of dismantling the product identifies
assemblies and sub-assemblies and hence results in the overall product structure as
presented in Figure A-2. Figure A-3 shows the dismantled drill with the assemblies and
sub-assemblies. The other assemblies and sub-assemblies mentioned in the product
structure are shown from Figure A-4 to Figure A-7.
261
Appendix A: Bill of Material (BOM)
Figure A-1 Hammer drill
A.2 BOM for Hammer Drill
Product structure is a result of design specification and is a helpful tool in identifying
assemblies, sub-assemblies and parts & components. The information when presented in
tabular form with parts and assemblies names and their respective material quantities
constitute BOM. BOM presents product design details by exploding assemblies and sub-
262
assemblies to part and component level and identifying the required materials with their
respective quantities as shown below for the selected hammer drill in Table A-1.
263
Figure A-2 Product structure (Hammer Drill)
264
Figure A-3 Dismantled drill with assemblies and sub-assemblies
265
Figure A-4 Parts and components in product structure
Figure A-5 Winding (Stator and rotor) and drive assembly
266
Figure A-6 Driven assembly (with gear and shock absorber) and Drill/Hammer switch
Figure A-7 Trigger assembly
267
The BOM for the drill at product level comprise parts & components and the drill
assembly. The drill assembly can be considered a phantom assembly. Phantom
assembly represents a group of components or parts that cannot be assembled together
unless some other parts and/or assemblies are added to it. This type of imaginary
assembly adds to no operation/processing cost. Respective material and material
quantities are also shown in Table A-1 for each component used. BOM explodes the
drill assembly into electrical and mechanical assemblies with further explosions later
on. Material quantities at the part and component levels add up to sub-assemblies and
assemblies level and eventually the overall product level.
Table A-1 Hammer Drill (product level) 1.6 Kg
Quantity per item

Component No of items Material
(total quantity) gm
Depth gauge 1 Plastic 10 (10)
Chuck key 1 Steel 40 (40)
Chuck 1 Steel 120 (120)
Shell 2 Plastic 50 (100)
Front support handle 1 Thermoplastic 150 (150)
Chuck screws 2 Steel 0.5 (1)
Body screws 6 Steel 0.5 (3)
Drill Assembly 1 - 1176 (1176)
Drill Assembly (phantom assembly) 1176gm
268
Electrical Assembly 1 - 629 (629)
Mechanical Assembly 1 - 547 (547)
Electrical Assembly (phantom assembly) 629gm
Winding 1 - 325 (325)
Connections 1 - 304 (304)
Winding (assembly level) 325gm
Stator 1 - 207 (207)
Rotor 1 - 118 (118)
Stator (sub-assembly level) 207gm
Shell 1 Electrical steel 87 (87)
Stator winding 1 Copper 52 (52)
Insulators 8 Plastic 3 (24)
Pegs 4 Thermoplastic 2 (8)
Holding strip 2 Steel 12 (24)
Cables 4 Copper 3 (12)
Rotor (sub-assembly level) 118gm
Shell 1 Electrical Steel 28 (28)
Core 1 Plastic 12 (12)
Copper plate 1 Copper 15 (15)
Rotor winding 1 Copper 63 (63)
Connections (phantom assembly) 304gm
269
Rotor Connection 1 - 28 (28)
Fuse 1 - 5 (5)
Trigger Assembly 1 - 31 (31)
Mains 1 - 240 (240)
Rotor Connection (phantom assembly) 28gm
Brush Assembly 1 - 14 (14)
Plastic cover 2 Plastic 2 (4)
Plastic end 2 Plastic 2 (4)
Wire holder 2 Copper 3 (6)
Brush Assembly (sub-assembly level) 14gm
Spring 2 Steel 2 (4)
Copper Wire 2 Copper 1 (2)
Copper plate 2 Copper 1 (2)
Brush 2 Coal 3 (6)
The Trigger Assembly (assembly level) 31gm
Screw 4 Steel 0.5 (2)
Clamp 2 Steel 1.5 (3)
Connector type I 2 Copper 1.5 (3)
Connector type II 2 Copper 1.5 (3)
270
Plastic Mould 1 Plastic 5 (5)
Trigger 1 - 10.5 (10.5)
Lock assembly 1 - 3.5 (3.5)
Trigger (sub-assembly level) 10.5gm
Spring 2 Steel 1.5 (3)
Plastic shell 1 Plastic 3 (3)
Clamp 1 Plastic 2 (2)
Trigger frame 1 Plastic 2.5 (2.5)
Lock assembly (sub-assembly level) 3.5gm
Spring 1 Steel 1.5 (1.5)
Plastic Press 1 Plastic 2 (2)
Mains 240gm
Cable 1 Insulated Cu Wire 155 (155)
Plug 1 - 85 (85)
Mechanical Assembly (phantom assembly) 547gm
Drive Assembly 1 - 281 (281)
Driven Assembly 1 - 167 (167)
Drill/Hammer Switch 1 - 99 (99)
Driven Assembly (assembly level) 167gm
Driven shaft 1 Carbon Steel 55 (55)
271
Driven gear 1 Carbon steel 23 (23)
Fastening spring 1 Steel 12 (12)
Shock absorber 1 - 77 (77)
Shock Absorber (sub-assembly level) 77gm
Shell 1 Alloy steel 62 (62)
Drive Assembly (assembly level) 281gm
Driving shaft 1 Steel 86 (86)
Fan 1 Plastic 35 (35)
Journal bearing 1 Stainless steel 32 (32)
Ball bearing 1 Stainless steel 26 (26)
Gear holder assembly 1 - 102 (102)
Gear Holder Assembly (sub-assembly level) 102gm
Static gear 1 Steel 24 (24)
Gear ring 1 Brass 24 (24)
Holding shell 1 Alloy steel 54 (54)
Drill-Hammer Switch (sub-assembly level) 99gm
Steel strip 1 Steel 35 (35)
Sphere 1 Steel 38 (38)
Plastic Mould 2 Plastic 13 (26)
272
From MRP point of view, BOMs are maintained at the cumulative level for a product
and represent combined quantities of similar materials at the lowest level as shown in
Table A-2. These quantities are helpful in procuring the material based on the product
manufacturing quantities.
Table A-2 Cumulative material quantities at the lowest level for the hammer drill
S. No. Material Quantity (gm)
1 Plastic 229.5
2 Thermoplastic 158
3 Steel 406.5
4 Electrical steel 115
5 Carbon steel 78
6 Alloy steel 116
7 Stainless steel 58
8 Copper 158
9 Brass 24
10 Insulated copper wire 155
11 Coal 6
12 Screws 12 Pcs
273
Appendix B Deviation Indices
The appendix B is aimed at deriving the equations for cost deviation indices. The indices
form part of the PCE Hybrid Model.
B.1 Material cost deviation index
Material cost deviation index value refers to the amount of deviation in the cost of a
material from its original cost. For example, material cost deviation index for the pth
product in the nth year represented by φ pn can be given in terms of material cost for that
n −1
product in the nth year ( Cmp
n
) and the (n–1)th year ( C mp ) as follows:
n −1
n
C mp − C mp n
C mp
φ pn = n −1
= n −1
−1 (B-1)
C mp C mp
Any deviation in the costs is a result of not just inflation but other factors. If I n is the
deviation index due to inflation in the nth year and J pn represents the index due to other
factors for the pth product in the same year, then φ pn can also be given as follows:
274
Appendix B: Deviation Indices
φ pn = I n + J pn ⇒ J pn = φ pn − I n (B-2)
The above equation (B-2) can be converted in a similar way for the (n+1)th year.
φ pn +1 = I n +1 + J pn +1 (B-3)
Where, φ pn+1 is the material cost deviation index for the pth product in the (n+1)th year, I n+1
is the deviation index due to inflation in the (n+1)th year and J pn+1 represents the index due
to factors other than inflation for the pth product in the (n+1)th year. The value of J pn+1 is
based on J pn and can be obtained by taking into account I n+1 i.e.
J pn +1 = J pn (1 + I n +1 ) (B-4)
Therefore, the value of J pn+1 can be replaced in the equation (B-3) as follows:
φ pn +1 = I n +1 + J pn (1 + I n +1 ) ; Now replacing the value of J pn from equation (B-2):
φ pn +1 = I n+1 + (φ pn − I n )(1 + I n +1 ) ; and now replacing the value of φ pn from equation (B-1):
275
 C mp
n

φ pn +1 = I n +1 +  n −1
− 1 − I n (1 + I n+1 )

(B-5)
C mp 
Equation (B-5) can be used to determine material cost indices for number of ‘p’ products in
the (n+1)th year using their respective material costs in the nth and the (n–1)th years and the
inflation values in the nth and the (n+1)th years. Therefore, the known cost data can be used
to predict the indices for the product range.
B.2 Labour cost deviation index
Let average per month wages of non–skilled, semi–skilled and skilled labour in the nth year
be W0n , W1n and W2n respectively and for the (n–1)th be W0n −1 , W1n −1 and W2n −1
respectively. If labour cost deviation index for non–skilled, semi–skilled and skilled labour
in the (n+1)th year are represented by ε 0n +1 , ε 1n +1 and ε 2n +1 respectively, equation (B-5) can
be used in a similar way to give these indices as follows:
 W0n 
ε 0n +1 = I n +1 +  n −1
− 1 − I n (1 + I n +1 ) (B-6)
 W0 
 W1n 
ε 1n +1 = I n +1 +  n −1
− 1 − I n (1 + I n +1 ) (B-7)
 W1 
276
 W2n 
ε 2n +1 = I n +1 +  n −1
− 1 − I n (1 + I n +1 ) (B-8)
 W2 
In order to determine the estimated labour rate value for the (n+1)th year, an average labour
cost deviation index value ( ε n +1 ) can be determined.
B.3 Processing cost deviation index
The aggregate shop floor wide processing rate in the nth and the (n–1)th years represented as
n n −1
RMA and R MA respectively can be used in a similar way to predict the processing cost
deviation index value for the (n+1)th year ( µ n +1 ) as follows:
 Rn 
µ n +1 = I n +1 +  MAn −1
− 1 − I n (1 + I n +1 ) (B-9)
 R MA 
B.4 MDC deviation index
If MDC fraction values ( C md / C mt ) in the nth and the (n–1)th years are represented as RMDC
n
n −1
and R MDC respectively, the above method can be used in a similar way to predict the MDC
deviation index value for the (n+1)th year ( ρ n +1 ) as follows:
277
 Rn 
ρ n +1 = I n +1 +  MDC
n −1
− 1 − I n (1 + I n +1 ) (B-10)
 R MDC 
B.5 Tool cost deviation index
Machine tool cost deviation index value for the (n+1)th year (ψ n +1 ) can be presented using
n −1
actual machine tool rates in the nth and the (n–1)th years (i.e. RMT
n
and RMT respectively) as
follows:
 Rn 
ψ n +1 = I n +1 +  MT
n −1
− 1 − I n (1 + I n +1 ) (B-11)
 RMT 
Similarly, labour tool cost deviation index value for the (n+1)th year ( σ n +1 ) can be given by
n −1
using the actual labour tool rates in the nth and the (n–1)th years (i.e. R LT
n
and R LT
respectively) as follows:
 Rn 
σ n +1 = I n +1 +  nLT−1 − 1 − I n (1 + I n +1 ) (B-12)
 R LT 
278
B.6 Building cost deviation index
Again an index value for building cost deviation in the (n+1)th year ( δ n +1 ) can be obtained
by using the actual building space rates in the nth and the (n–1)th years (i.e. RBn and R Bn−1
respectively) in the following way:
 Rn 
δ n +1 = I n +1 +  nB−1 − 1 − I n (1 + I n +1 ) (B-13)
 RB 
B.7 PO deviation index
If production overhead fraction values ( Ot / C Gt ) in the nth and the (n–1)th years are
n −1
n
represented as R PO and R PO respectively, the PO deviation index value for the (n+1)th
year (τ n+1 ) can be given as follows:
 Rn 
τ n +1 = I n +1 +  POn −1
− 1 − I n (1 + I n +1 ) (B-14)
 R PO 
279

Product Cost Estimation: Theoretical Development and Industrial Validation

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Product Cost Estimation: Theoretical Development and Industrial Validation

Caricato da

Copyright:

Formati disponibili

Product and Manufacturing Cost Estimation:

Theoretical Development and Industrial Validation

This thesis is submitted to the Department of Mechanical Engineering,

My Mum and Dad

List of Notations ......................................................................................................... 16

List of Acronyms ........................................................................................................ 22

C hapt e r 1 Introduction ......................................................................................... 27

1.2 Problems identification................................................................................ 29

1.3 Aims and objectives .................................................................................... 33

1.4 Thesis structure ........................................................................................... 35

C hapt e r 2 Background and Related Work ............................................................ 39

2.2 Manufacturing and cost control ................................................................... 41

2.3 Cost estimation............................................................................................ 45

2.4 Cost estimation in the early design stages .................................................... 47

2.5 Cost estimation for specific applications...................................................... 48

2.5.1 Cost estimation for a specific segment in a production cycle................ 49

2.5.2 Cost estimation for specific machining and manufacturing processes... 50

2.5.3 Cost forecasting for specific parts and products ................................... 50

2.5.4 Cost estimation for generic systems ..................................................... 51

C hapt e r 3 PCE Technique Classification System ................................................. 55

3.2 Development of hierarchical classification system (HCS)............................ 57

3.3 Intuitive cost estimation techniques ............................................................. 60

3.3.1 Case-based methodology ..................................................................... 60

3.3.2 Decision support systems (DSS) .......................................................... 63

Rule-based systems ......................................................................................... 65

Fuzzy logic approach....................................................................................... 68

Expert systems ................................................................................................ 69

3.4 Analogical cost estimation techniques ......................................................... 70

3.4.1 Regression analysis models.................................................................. 70

3.4.2 Back-propagation neural network (BPNN) models............................... 70

3.5 Parametric cost estimation techniques.......................................................... 71

3.6 Analytical cost estimation techniques .......................................................... 73

3.6.1 Operation based approach .................................................................... 73

3.6.2 Breakdown approach ........................................................................... 75

3.6.3 Tolerance-based cost models ............................................................... 76

3.6.4 Feature-based cost estimation .............................................................. 78

3.6.5 Activity-based costing (ABC) system .................................................. 79

C hapt e r 4 MRO and TRO Estimation Methods .................................................... 87

4.2 Cost estimation methodology at the selected company................................. 91

4.2.1 Material cost estimation ....................................................................... 92

4.2.2 Direct labour costs ............................................................................... 93

4.2.3 Overheads estimation........................................................................... 93

4.3 Proposed methodology for overheads estimation ......................................... 95

4.3.1 MRO estimation model........................................................................ 95

4.3.2 TRO estimation model......................................................................... 98

4.4 Model implementation and validation........................................................ 101

4.5 Conclusions............................................................................................... 112

C hapt e r 5 PCE Hybrid Model ............................................................................ 114

5.1 Introduction............................................................................................... 114

5.2 Product cost and modelling approach......................................................... 116

5.3 Direct cost elements .................................................................................. 122

5.3.1 Direct material costs .......................................................................... 123

5.3.2 Direct labour...................................................................................... 126

Labour units .................................................................................................. 128

Labour rate.................................................................................................... 130

5.4 Indirect cost elements ................................................................................ 132

5.4.1 Processing cost .................................................................................. 132

Processing units............................................................................................. 133

Processing rate .............................................................................................. 134

5.4.2 Material dependent cost ..................................................................... 135

5.4.3 Tooling cost....................................................................................... 136

Machine tool rate........................................................................................... 137