International journal of computer integrated manufacturing , tập 23, số 7, 2010

The paper begins by reviewing various paradigms for manufacturing systems improvement includingdesign/redesign-, maintenance-, operator-, process-, product- and quality-led initiatives..

The barriers to realising sustainable process improvement: A root cause analysis

of paradigms for manufacturing systems improvement

B.J Hicks* and J MatthewsInnovative Design and Manufacturing Research Centre, Department of Mechanical Engineering, University of Bath, UK

(Received 2 February 2010; final version received 8 April 2010)

To become world-class, manufacturing organisations employ an array of tools and methods to realise processimprovement However, many of these fail to meet expectations and/or bring about new less well understoodproblems Hence, prior to developing further tools and methods it is first necessary to understand the reasons whysuch initiatives fail This paper seeks to elicit the root causes of failed implementations and consider how these may

be overcome The paper begins by reviewing various paradigms for manufacturing systems improvement includingdesign/redesign-, maintenance-, operator-, process-, product- and quality-led initiatives In addition to examiningthe knowledge requirements of these approaches, the barriers to realising improvement are examined throughconsideration and review of literature from the fields of manufacturing, management and information systems.These fields are selected because of the considerable work that deals with process improvement, changemanagement, information systems implementation and production systems The review reveals the importance offundamental understanding and highlights the lack of current methods for generating such understanding Toaddress this issue, the concept of machine-material interaction is introduced and a set of requirements for asupportive methodology to generate the fundamental understanding necessary to realise sustainable processimprovement is developed

Keywords: manufacturing improvement; tools and methods; knowledge requirements; generating understanding

1 Introduction

In today’s highly competitive global markets product

quality and cost, and manufacturing efficiency and

flexibility are critical factors in an organisation’s

commercial success (Roth and Miller 1992,

Manarro-Viseras et al 2005, Swink et al 2005) The dimensions

associated with production and in particular quality,

efficiency and flexibility ultimately define the unit cost

of the finished product, and are therefore a central

focus of any organisation’s business plan and

perfor-mance monitoring However, the three factors of

quality, efficiency and flexibility are heavily

inter-related and attempts to optimise one factor can have a

potentially detrimental effect on the other It is

therefore important to consider the collective effect of

these dimensions on the organisation’s manufacturing

capability (cf Figure 1(a))

Within a manufacturing context, quality refers to

the perception of the degree to which the product or

service meets the customer’s expectations For any

manufacturing process to be capable it must be able to

produce a quality product As the customer

require-ments for quality increase the manufacturing

capabil-ity must also evolve Manufacturing efficiency is

effectively a measure of the profit or return realisedfrom the manufacturing system or process (Hansen2005) At the manufacturing system level this canequate to the time it takes to complete a given task orthe number of staff members needed to facilitate theproduction of a particular item The aim of flexibility

in a manufacturing system is to change the mix,volume and timing of its output and essentiallydescribes the ability to process variant products(Matthews et al 2006) When considering the overallmanufacturing capability, flexibility has the twodimensions, range and response The range flexibilitystates what a manufacturing system can adopt in terms

of number of different products and output levels –termed product flexibility and volume flexibility; theresponse flexibility describes the ease with which asystem can be adapted from one state to another –termed delivery and mix in Slack (2005) This responseflexibility must be considered in terms of time, cost andorganisational disruption In general flexibility offersthe manufacturer some degrees of freedom to takeadvantage of demand opportunities and simulta-neously provide an ability to reduce losses (Bengtsson2001)

*Corresponding author Email: b.j.hicks@bath.ac.uk

Vol 23, No 7, July 2010, 585–602

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/0951192X.2010.485754

While attempts to improve particular aspects of,

for example, the product design or the manufacturing

process can lead to improvements in the areas of either

quality, efficiency or flexibility, it is ultimately the sum

of all systems, actors and inputs associated with the

realisation of the product that determine levels of

quality, efficiency and flexibility Hence,

manufactur-ing capability is dependent upon an organisation’s

people, its processes, its products and its practices (cf

Figure 1(b)) Achieving a high level of manufacturing

capability and the attainment of high levels of

performance within each of the these areas is

frequently associated with the notion of ‘World-Class

Manufacturing’ (Maskell 1991) While at a given point

in time an organisation may be performing at a high

capability level it is the ability to sustain an optimal or

near optimal level that is the characteristic of a truly

class organisation Hence, the notion of

world-class manufacturing and ‘world-world-class’ organisations is

more about the ability of an organisation; its people,

processes, products and practices (cf Figure 1(b)), to

adapt, improve and evolve within the context of the

changing business environment (cf Figure 1(c)) (Riek

et al 2006) This ability to respond and adapt isbecoming of increasing importance as product com-plexity increases (Sommer 2003); customer demand forproduct variety increases (Jiao and Tang 1999);product lifecycles shorten (Christopher and Peck2003); legislation concerning areas such as materials(European packaging and packaging waste directive2004/12/EC), emissions (Ambient air quality assess-ment EC Directive 96/62/EC) and Health and Safety(European Machine Safety 98/37/EC) tighten; supplychains and customers become global (Gelderman andSemeijn 2006)

As a consequence of the influence of people,products, processes and practices on an organisation’smanufacturing capability there exists a wide variety oftools, methods and approaches to deliver targetedimprovements in a particular area However, in manycases the improvement projects fail to meet expecta-tions and in extreme cases can fail to deliver anyimprovement or bring about new less well understoodproblems (Hicks et al 2002) Furthermore, of those

that do deliver improvements many are short-term

(Keating et al 1999) and the improvements are lost

when there is, for example, a change of staff, variation

in materials or process inputs, altered practices, the

introduction of new equipment or yet another

initia-tive From an organisation’s perspective these

pro-grammes not only require an investment of many tens

or hundreds of thousands of pounds (Chapman et al.,

1997, Sterman et al 1997, Keating et al 1999) but in

the case of failed initiatives incur an indirect cost which

can represent a magnitude of cost and lost opportunity

which far exceeds the cost of the original improvement

programme For example, optimising set-up and

process parameters could make the manufacturing

system sensitive to variation in inputs, e.g materials,

and result in significant downtime

For these reasons and to ensure long-term success,

manufacturing organisations need to possess a

func-tional and holistic understanding of the production

systems and the variety of tools, methods and

approaches for improvement (cf Figure 1(d)) in order

that they may be successfully applied and reapplied

within the context of the changing business

environ-ment Furthermore, as previously stated, it is the

ability of an organisation; its people, processes,

products and practices to adapt, improve and evolve

within the context of the changing business that

enables it to be ‘World Class’ A prerequisite for

achieving this is the means or capability to generate the

fundamental understanding necessary to respond

appropriately It is the critical dimension of

under-standing and the creation of methods for generating

the necessary understanding that is addressed in this

paper

This paper first explores the motivations for

manufacturing improvement and examines in detail

the principles and underlying knowledge requirements

of a range of common improvement paradigms The

barriers to realising sustainable improvement are then

discussed and the importance of generating and

communicating a fundamental understanding is

high-lighted The need to support organisations in

reinfor-cing and extending their fundamental understanding is

further argued and the deficiencies in existing

suppor-tive techniques are described In order to overcome

these deficiencies the concept of machine-material

interaction is introduced and its relationship to

‘function’ and fundamental understanding is discussed

The paper concludes with the development of a set of

requirements for a new supportive methodology which

enables machine–material interactions to be

investi-gated, and the necessary fundamental understanding

to be developed and contextualised with respect to the

knowledge requirements of a range of common

improvement paradigms

2 Improvement paradigmsThere are a wide variety of approaches and philoso-phies associated with the improvement of manufac-turing and production systems These higher levelparadigms generally involve a range of tools andmethods to target, plan and implement an improve-ment programme For the purpose of consideringthese various philosophies and their correspondingtools and methods (Brassard and Ritter 1994), theapproaches and the methods can be grouped underthe seven areas of: equipment design/redesign, main-tenance, operator-led, process-control, product mod-ification and new product introduction, quality, andtooling design and changeover The various manu-facturing paradigms and the corresponding tools andmethods that can be associated with each of theseseven areas are illustrated in Figure 2 and described

in detail in Tables 1 and 2 Of particular interest inthis work are the underlying knowledge requirementsnecessary to successfully apply the various tools andmethods These requirements are developed in Tables

1 and 2 from an analysis of the aims and underlyingprinciples of the various tools and methods, which arenow summarised

(1) Process control As levels of automationincrease and in particular, the automation ofchangeover and machine set-up, so does theneed to possess the understanding necessary toexplicitly define set-up rules and parameters.Intelligent monitoring and control has beensuccessfully applied in Component manufac-ture (Uraikul et al 2000, Murdock and Hayes-Roth 1991) and Machining processes (Hou

et al 2003, Liang et al 2004) but requires depth knowledge of the relationship betweenproduct variation and process variation - bothupstream and downstream Central to thesuccess of these methods is the need to under-stand and describe the acceptable variation inproduct attributes during all stages ofproduction

in-(2) Operator-led One of the key elements to theeffective operation and improvement of aproduction system is the successful training ofthe operating staff (Woodcock 1972) Training

is imperative to ensure changes to workingpractices and operating procedures are effec-tively taken-up For effective training to bedelivered the trainer needs to possess an in-depth understanding of the content (Davis andDavis 1998), which in the case of manufactur-ing improvement concerns both the tools andmethods for improvement and the production

system(s) Further, the content and learning

outcomes of the training have to reflect

good-practice or at least improved good-practices, which

must be determined in advance Central to the

success of the training is the need to develop a

common and shared understanding across all

the trainees in order to generate the same

intended learning outcome(s) This is

neces-sary to ensure consistent practices and in

particular, consistent operation of equipment,

control of materials and the adoption of

appropriate machine settings to maintain

quality and avoid excessive wear (Adebanjoand Kehoe 2001)

(3) Maintenance The ability to keep a turing process efficient depends heavily upongood work practices and effective maintenance.This is particularly important in today’s just-in-time production environment, where as aconsequence of reduced stock level minorbreakdowns are even more likely to stop orinhibit production (Eti et al 2006a) and reduceoverall equipment effectiveness (efficiency).There are two approaches for achieving this

The first is preventive maintenance which aims

to reduce the probability of failure in the timeperiod after maintenance has been applied Thesecond is corrective maintenance, which strives

to reduce the severity of equipment failuresonce they occur (Loftsen 2000) As noted byWaeyenbergh and Pintelon (2004) industrialsystems evolve rapidly so maintenance initia-tives will also have to be reviewed periodically

in order to take into account the changingsystems and the changing environment Thiscalls not only for a structured maintenanceconcept, but also one that is flexible There are

a variety of maintenance improvement methodsincluding Design for Service (DfS) (Dewhurstand Abbatiello 1996), Total Productive Main-tenance (TPM) (Willmott 1997) and ReliabilityCentred Maintenance (RCM) (Smith 2005)which arguably focus on the design, theoperator and the engineering function respec-tively These various approaches depend onboth the management and the operatorspossessing an understanding of: the function

of the process, the influence of machine settings

on process performance, the impact of wear onthe process, and the effect of operating condi-tions (production rate and environmentalconditions)

(4) Quality In a similar manner to maintenancethere are a variety of methods and initiativesthat support quality control, improvement andassurance These include Quality FunctionDeployment (QFD) (Govers 2001), TotalQuality Management (TQM) (Oakland 2003)and aspects of Six Sigma (Adams et al 2003).These various approaches require an under-standing of function and its relationship toquality, and an understanding of the interac-tion between the process and product, whichare essential for directing the measurement,analysis, improvement and control of processand process inputs (materials and staff) (Tho-mas and Webb ( 2003) and Antony (2007a,2007b))

(5) Tooling design and changeover The ultimateaim of improving tooling design is to improveproduction performance and in particularflexibility, without compromising efficiency.Key to achieving this is to determine the mostappropriate design or configuration of toolingand, if appropriate, the most efficient methodsfor changeover between tooling configurations(i.e minimising changeovers and/or changeovertime) This includes both the physical geometry(size, profile and number of) and control of the

tooling (kinematics – motion, velocity and

acceleration, timing and clearances) (Hicks

et al 2001) Central to the success of the

Single-Minute Exchange of Die (SMED)

(Shin-go 1985) or Design for Changeover (DFC)

(McIntosh et al 2001) activities is the need to

be able to understand and specify in advance

the machine settings (set-up point) and range of

variation (run-up adjustment) necessary for the

successful processing of each product variant

(6) Equipment redesign, modification and

replace-ment Where an increase in manufacturing

capability is sought that exceeds the existing

equipment or process capability, it is necessary

to either modify or replace the equipment In

cases where the process and the design

princi-ples which underlie the equipment are identified

to be close to their limits then a process and

equipment redesign may be necessary (Hicks

et al 2002) In either case – modification,

replacement or redesign – it is a prerequisite

that both capability and functional

require-ments are determined Central to determining

these requirements is the need to understand the

limitations of the existing equipment (Matthews

et al 2007, Ding et al 2009) The factors that

limit the capability can be inverted in order to

define the rules which are necessary for

success-ful processing This understanding is central to

realising redesigned or new equipment that

overcomes the limitations of existing equipment

and ultimately improves performance (quality,

efficiency and/or flexibility and capability) The

rules also provide a series of objective measures

for the evaluation and assessment of new

equipment (Matthews et al 2008)

(7) Product modification and new product

intro-duction In today’s dynamic global markets,

goods manufacturers are frequently faced with

the task of processing new or altered products –

such as new sizes, new materials and modified

configurations (Matthews et al 2009) Central

to achieving this, is the need to determine an

appropriate set of machine settings that enable

the product to be successfully processed No

matter whether it is the determination of

settings for a new product or the improvement

in process capability through product

modifi-cation, it is necessary to understand the

capability of the production process and its

relationship with the properties and

character-istics of the product (Frey et al 2000)

(8) Other manufacturing philosophises In

addi-tion to these seven areas of manufacturing

improvement there exist a number of

philosophies to support improvements in ufacturing and management These includelean thinking and Business Process Reengineer-ing The term ‘lean’ was coined by Womack

man-et al (1990) to describe the main aim of thephilosophy – the reduction of waste throughout

a company’s value stream However, for somelean promoters it is not just a set of tools for thereduction of waste (Bicheno 2003), but a way ofthinking which puts the customer first Oncethis way of thinking is adopted, lean tools areavailable to reduce waste and improve benefitsfor the customer For the successful adoption

of a lean approach a functional perspective ofthe production system is required in order forvalue streams to be identified and mapped, and

to ensure that value streams flow In amanufacturing context, function is the onlymeans to add value to the product Althoughnot all functions may add value In contrast tolean, business process reengineering or businessprocess redesign (BPR) focuses on improvingthe efficiency and effectiveness of the overallbusiness processes that exist within and across

an organisation This is achieved by ing the processes and assigning responsibilityfor those processes to dedicated teams and,where appropriate, systems (Hammer andChampy 1993) In order to maintain andimprove processes an understanding of thefunctions and processes and the value of eachfunction must be elicited

establish-The previous sections have discussed the variousmanufacturing improvement paradigms and corre-sponding tools and methods with respect to theirunderlying principles and the knowledge and under-standing that underpin their use Further examination

of the knowledge requirements reveals six fundamentalknowledge concepts relating to the improvement ofmanufacturing systems These include:

(1) An understanding of the relationship betweenthe properties and characteristics of the pro-duct, and the machine and process settings.(2) An understanding of the relationship betweenproduct variation and process variation, andtheir influence on quality and efficiency.(3) An understanding of the influence ofoperator procedures on quality, efficiency, andflexibility

(4) An understanding of the impact of wearand operating conditions (production rateand environmental conditions) on quality andefficiency

(5) An understanding of the limitations of the

existing equipment (quality, efficiency,

flexibil-ity and capabilflexibil-ity)

(6) A functional perspective of the production

system that contextualises the process and its

operations with respect to the final product

It is arguable that these six knowledge concepts are

critical for effective implementation of improvement

programmes and that they are hence a prerequisite for

realising sustainable improvement In order to explore

this further the barriers and root causes of failed or

partially successful organisational improvement

pro-grammes are reviewed

3 Barriers to realising manufacturing improvement

While there exists a plethora of publications presenting

the successful implementation of different

manufactur-ing improvement strategies (Brown et al 1994, Sohal

et al 1998, Bamber 1999, Henderson and Evans 2000,

Antony and Banuelas 2002, Apte and Goh 2004, Chan

et al 2005) the experiences of the authors and those of

the practitioners we have worked with are that many

initiatives fail to meet expectations and can fail to

deliver any improvement at all Furthermore, in

extreme cases these initiatives can have a detrimental

impact on capability or bring about new less well

understood problems This can result in an indirect

cost to an organisation that represents a magnitude of

cost and lost opportunity that far exceeds the level ofinvestment in the original improvement programme.The existence of only partially successful and failedinitiatives is supported by past and contemporaryliterature, an example of this being Redman andGrieves (1999), who noted that between 70–90% ofTQM programmes implemented have failed

In order to provide some insight into the commoncauses of partially successful and failed initiatives –and what can be thought of as the barriers to successfulimplementation – literature from the fields of manu-facturing, management and information systems arecritically reviewed These fields are selected because ofthe considerable bodies of work that deal with processimprovement, change management, information sys-tems implementation and production systems Anappraisal of the literature reveals six core areas: lack

of commitment, reactive organisations, layered tives, incomplete implementations, incorrect imple-mentations and resistance to change These sixdimensions are shown in Figure 3 and discussed inthe following sections

initia-3.1 Lack of commitment from the organisationOne of the most common causes for organisationalimprovement programmes to fail is the lack ofcommitment from the organisation (Sterman et al

1997, Olivia et al 1998, Mellor et al 2002, Tari andSabaner 2004) This can lead to inadequate support

infrastructure or training in improvement techniques,

thereby limiting the potential for successful

implemen-tation (Keating et al 1999) Top-down organisational

commitment is imperative to successful improvement

programmes, although, McIntosh et al ( 2001) argue

that the focus is often heavily concentrated on

organisational-led improvement and that the benefits

of product/ process design amendments are often

considerably under-exploited If those responsible for

the allocation of resources are not well informed about

the pros and cons of the implementation programme, it

is highly likely they will underestimate the effort, in

terms of time and cost, needed for the successful

completion of the project (Wilkinson et al 1998, Tari

and Sabanter 2004) In the field of business

transfor-mation and Enterprise Resource Planning (ERP) a

lack of commitment is also highlighted as a common

cause of failure This includes both lip service from

senior staff and a lack of engagement from middle

management (Buckhout et al 1999, Whittaker 1999)

3.2 Reactive approaches

In the dynamic business environments of today where

resources are already stretched it is common for

organisations to adopt a reactive approach, always

‘fire fighting’ issues such as quality and efficiency

Research by Olivia et al (1998) showed that such a

reactive approach not only assisted the failure of

specific initiates but caused profound effects on other

functions in the organisation such as product

devel-opment, pricing and human resources Overzealous

application of quality tools has led to declining

effectiveness and a backlash that damages even the

effective programmes in many companies (Keating

et al 1999) Eti et al (2006b) show that chemical plants

employing reactive strategies of maintenance are

incurring maintenance cost of 5% per annum of the

asset-replacement cost, in lost productivity i.e wastage

of $30,000 per $M of asset value, this in comparison to

companies employing proactive strategies who are

seeing 25% savings on these values Furthermore, with

increased adoption of Total Quality Management

approaches and reduced stock level owing to

just-in-time work practices minor breakdowns are even more

likely to stop or inhibit production (Eti et al 2006a)

Because of this, reactive maintenance approaches such

as run-to-fail or breakdown are becoming less

com-mon, and are only employed in areas that do not result

in increased expenditure (Mostafa 2004) It therefore

follows that initiatives, such as those involving quality

can rarely be implemented in isolation Rather, they

need to be implemented as part of an overall

improvement programme, which in the

aforemen-tioned case of quality also includes reliability

3.3 Layered initiatives one on top of anotherThe reactive approach discussed in Section 3.2 can alsocontribute to organisations implementing multipleimprovement initiatives concurrently This makes thelifecycle of the implementation difficult to identify(Irani and Love 2001) and the tasks of planning,implementation and monitoring difficult Althoughresearch has shown that quality and productivityimprovements need to occur together for organisations

to maintain or improve their competitive position(Chapman and Hyland 1997), particular initiativesneed to be completed so that their effect can beunderstood (Bessant et al 2001), and concurrentinitiatives need to be carefully coordinated In thefield of manufacturing, Wilkinson et al (1998) identifythat a lack of understanding and structure whenimplementing multiple quality improvements leads tosituations that are considered ‘indigestible’ for those

on the receiving end of management In essence,employees struggle to differentiate between improve-ment initiatives, so tend to have cursory ‘buy-in’ to theprocess, or implement initiatives incorrectly

3.4 Incomplete implementations

A common cause of underperforming improvementinitiatives can be attributed to incomplete implementa-tions This includes partial implementation of aninitiative, implementations which have not been fullyimplemented across the entire organisation and im-plementations which have not been integrated withinthe business strategy and processes The consequences

of this are that either little or no measurableperformance improvements can be identified andorganisations need to maintain their existing systemsand processes – effectively maintaining two parallelprocesses (Hicks et al 2006) These issues are furtherfrustrated by the fact that there is normally deteriora-tion in performance measures when such programmesare up-and-running (Carroll et al 1998) This againcauses managers to lose faith in the programme andwithdraw to the existing working practices Haley andCross (1993) noted how some managers saw theirimplementation of quality improvement paradigms as

a ‘fashion statement’ Redman and Grieves (1999) alsoreviewed multiple sources of TQM failure through the1990s and identified that incomplete implementationwas the most common cause

3.5 Resistance to changeResistance to change has been widely reported as one

of the key barriers to successful implementation ofbusiness process transformation and improvement

programmes (Hill 1991, Rees 1991, Marchington et al.

1992) While senior managers appear to be committed

to quality improvement strategies, it was the middle

and junior managers that were resistant to such

programmes Middle management see the

implementa-tion of such programmes as both labour and resource

demanding (Wilkinson et al 1992), whereas junior

management thought it would ‘reduce their discretion’

in the current job roles From the shop floor viewpoint,

almost every book, and academic publication presents

the issues of operator ‘buy-in’ If the members of the

shop floor, who are to be the hands-on users of such

processes, do not understand them or the benefits to

themselves, the implementation is bound to falter

(Schaffer and Thompson 1992) In addition to this,

previous research highlights shop floor suspicion as a

barrier when using performance measures as indicators

of success of implementation (Ukko et al 2007) The

perception being that the implementation of such

programmes only benefits management, and have little

impact on the welfare of the shop floor staff

3.6 Incorrect implementation

The most commonly reported reason for unsuccessful

implementation is that of incorrect implementation

(Miller and Congemi 1993, Regle et al 1994, Taylor

1997, Redman and Grieves 1999, Nwabueze 2001) This

can include the inappropriate adoption of a particular

tool, method or process given the industry sector of the

organisation and its existing business processes (Beer

et al 1990), and the incorrect tailoring of the tool,

method or process to the business; its processes, people,

procedures and products For example, where ERP

systems are considered the alignment of fit to an

organisation is critical for success (Bingi et al 1999,

Holland and Light 1999) this includes both the level of

business process reengineering necessary and the

amount of customisation (tailoring) of the system that

is necessary Where quality programmes are considered,

Guptara (1994) highlighted how quality guru’s can raise

awareness of quality issues; however they rarely provide

the tailored mechanisms to integrate improvement

programmes within the organisation and this can

eventually lead to incorrect implementation

3.7 The root cause of failed implementations

When considering the causes and consequences of the

six areas detailed previously, it becomes apparent that

many arise as either a result of a lack of understanding,

an inability to communicate understanding, or an

inability to generate the necessary understanding

Where this understanding relates to the system, its

intrinsic processes, external interactions, the wider

environment and the product of the process itself Forexample, in the case of resistance to change theprimary causes are a lack of understanding, a lack ofcommunication and lack of inclusion – which ulti-mately leads to lack of shared understanding In thecase of layered initiatives, the consequences are aninability to elicit the core understanding and difficulties

in performance measurement – which ultimatelyinfluences understanding

Given the aforementioned argumentation it followsthat in the context of manufacturing improvement, theunderlying root cause of failed and suboptimalinitiatives can be largely attributed to the level ofunderstanding of the relationship between the produc-tion system, its constituent processes, raw materials andthe product As previously stated, it is this under-standing that is necessary for effectively implementingimprovement initiatives and determining the optimummix of tools and methods to generate the maximumbenefit The importance of understanding has beenrecognised implicitly by researchers; however, addres-sing this deficiency has been largely overlooked Forexample, a weakness of Reliability-Centred Mainte-nance is that it is not always as analytically rigorous asall reliability-based analysis and hence is not developedupon a fundamental understanding but rather asimplified or Bayesian approach (Sivia 1996) Wherequality initiatives are considered there is a tendency tohire Total Quality Management (TQM) consultants tovisit for a half-day or so to start the process This putsincredible pressure on managers since they have littleongoing access to the expert help they need to make thiswork Also, some activities that are part of TQM arebest carried out by ‘outsiders’ who bring a differentkind of objectivity to the process and may help indeveloping the necessary understanding

4 Generating a fundamental understanding

In the previous sections it has been shown that themajority of manufacturing improvement approachesand tools require a fundamental understanding of theproduction system – including its constituent pro-cesses, raw materials and the product – and that thebarriers to successful implementation can be consid-ered to relate to either a lack of understanding, aninability to generate understanding or an inability tocommunicate understanding Furthermore, in today’sdynamic business environments where products, ma-terials, processes and staff continually change, organi-sations must continually reinforce and extend theirunderstanding The ability to increase and evolveunderstanding depends heavily upon tools and meth-ods which support the generation of understanding.For these reasons, it can be argued that a prerequisite

for realising sustainable process improvement is

fundamental understanding and in particular, an

ability to generate understanding

In the context of manufacturing improvement there

exist a variety of tools and methods which can be

con-sidered to support the development of understanding

These include methods such as Root Cause Analysis

(RCA) (Ammerman 1998) Fault Tree Analysis (FTA)

(Vesely et al 1981), Failure Mode Effect and Critical

Analysis FMECA (Stamatis 1981) and Value Stream

Mapping (VSM) (Rother and Shook 1999) FMECA

and FTA are based on the investigation of errors and

their causes, and are employed in the product lifecycle’s

idea identification, development and manufacturing

phases (Pisano 1997) However, their scope is limited as

they are only generally applied to investigate observed

failure and its impact, not why it has been observed

Although this is partially addressed by Root Cause

Analysis, where there is investigation into why the

failure happened, neither method adopts a functional

view that contextualises the failure with respect to the

intended function and the final product In contrast to

these failure driven approaches, customer focused

techniques such as Value Stream Mapping do adopt a

more functional perspective and attempt to identify

what action adds value to the product (Rother and

Shook 1999) However, this is also limited as it does not

consider how to assure value levels and whether the

levels of value are maintained, only that it flows

From a manufacturing organisation’s perspective it

is necessary to have an in-depth understanding of the

production system, its constituent processes, raw

materials and the product This perspective must be

interdisciplinary (maintenance, operators, quality,

materials etc) not just a single perspective such as

engineering Furthermore, the developed

understand-ing needs to be contextualised with respect to the

overall production system, product and function The

organisation needs to focus on observed failure

(reactive) and possible failure (proactive) this includes

the various dimensions of quality and efficiency and

their relationship to the production system, its

processes, materials and the product

5 Interaction, the key to fundamental understanding

In the context of manufacturing systems the

relation-ship between the various factors of machine, products,

process and materials is defined at the interface during

physical interactions (MMI) between the machine and

materials These machine-material interactions occur

where a machine component physically interacts with,

or influences, the product and any of its constituent

elements This includes the entire product lifecycle

from the processing of raw materials to the assembly of

the product, packaging operations and materials,collation and product handling, and eventually dis-posal and recycling One specific factor that is evidentfrom the review in Section 3 is that before anorganisation can begin to make targeted improve-ments, implement change or identify the limitations ofexisting systems, it is first necessary to possess thefundamental understanding of product, process andtheir combined interaction

This understanding will ultimately provide thestructure against which an organisation can reasonabout a system and thus, implicitly constrains the scope(potential) for realising improvements and for foresee-ing and overcoming particular problems and conflicts.More specifically, fundamental understanding is aprerequisite for developing a complete description ofthe system, its function(s) and performance, thedevelopment of common terminology (definitions)and a structured representation (diagram) of thesystem, its internal relations, inputs and externalinfluences These elements provide the basis forcommunication and reasoning about the system andalso provide a framework against which tools andmethods can be aligned and targeted, and their effectsmeasured The latter of which is essential for determin-ing the appropriate (optimal) mix of tools and methodswhich generate the maximum benefit for an organisa-tion It follows that there is a need to support theinvestigation of MMIs as not only a means to introduce

a specific improvement but to provide support in thegeneration of the fundamental understanding necessary

to best use the various tools and methods to bringabout successful improvement (change)

5.1 The requirements for a supportive methodologyThe previous section outlines the need to create astructured approach (method) that supports practi-tioners in investigating machine-material interactionand contextualising the understanding generated withrespect to the production system, its constituentprocesses, raw materials and the product Morespecifically, such an approach needs to:

Support the generation of the understanding andknowledge requirements that underpin commonimprovement paradigms (section 2.0)

Address the barriers to realising sustainableimprovement, and in particular the inability tocommunicate understanding (section 3.0) Overcome the limitations of current techniquesfor generating understanding and in particularthe lack of a proactive approach and the inability

to contextualise failure with respect to function(section 4.0)

Through consideration of these areas eight core

requirements can be elaborated for a new supportive

methodology

(1) To provide a scalable and extensible method

that provides the generation of a

comprehen-sive and fundamental understanding of the

entire production system, its operations,

func-tions and interacfunc-tions

(2) To support the development of common

terminology (definitions) for machinery,

opera-tions and funcopera-tions that is agreed by

represen-tatives from production, engineering, quality

and operations and shared across an

organisation

(3) To enable a formalisation of the understanding

and the unification of appropriate

interdisci-plinary knowledge including materials,

machin-ery and environmental conditions This would

provide an objective view of the process which

integrates materials and machinery knowledge

providing a means for different departments

and groups to undertake objective discussion

rather than adopting the cross-department

blame culture

(4) To provide a more complete description of

process efficacy (efficiency and effectiveness)

including measures of performance, quality and

process failure (including observed and possible

modes of failure) across the entire production

system

(5) To enable the identification of the factors

(including the properties, characteristics and

settings of machinery, product and packaging)

that affect process efficacy and to elicit the

important relationships

(6) To provide a structured representation

(stan-dardised diagram) of the system, its internal

relations, inputs and external influences, which

can be used to communicate and ensure all

stakeholders have a common, shared

understanding

(7) To enable the generation of qualitative and

quantitative rules that govern the efficacy of

functions (interactions) and define the

proper-ties and characteristics of the product, the

machine and settings necessary to achieve

desired levels of process efficacy These rules

may include for example limiting values,

suitable ranges of settings and/or optimal

settings for given products and/or materials

(8) To provide direction for the targeting of tools

and methods for manufacturing improvement

in order to deliver targets and sustainable

improvements and maximise benefits

It is has been argued that these requirements and asupportive methodology which meets these require-ments would generate the understanding and knowl-edge necessary to effectively implement targetedimprovements in the areas of process control, training,maintenance, quality, tooling design and changeover,redesign and replacement of machinery and newproduct introduction To illustrate the importanceand potential of a new supportive method the relation-ships between various common improvement ap-proaches and the requirements (1–7) are shown inFigure 4 In particular, Figure 4 highlights theimportance of holistic understanding, adopting afunctional perspective, determining a complete descrip-tion of process efficacy and identifying the factorswhich affect it It also highlights the importance of

‘rules’ for maintenance and design-led methods andtheir benefit to quality based methods

While the approach presented in this paperconcerns manufacturing systems, the requirementsand argumentation have been developed from a variety

of fields including manufacturing, management andinformation systems, leading to a more generalised set

of issues Similarly, the proposed requirements of asupportive methodology are arguably of wider applic-ability than manufacturing systems alone In particu-lar, the interaction-centred approach could be applied

to any systems that can be decomposed into operationsand functions that interact or manipulate the product.This could include, for example, manual tasks, dataprocessing and work flows In fact, interactiondiagrams have been developed within the UMLframework to describe interactions among the different

manufacturing improvement approaches

elements of a model This interactive behaviour is

represented in UML by two diagrams known as the

Sequence diagram and Collaboration diagram

(Abdurazik and Offutt 2000, Bauer et al 2001) The

sequence diagram emphasises on time sequence of

messages and the collaboration diagram emphasises on

the structural organisation of the objects that send and

receive messages While this form of diagram has been

applied predominantly to software systems there may

be opportunities for its application to production

systems

6 Conclusions

This paper deals with the area of manufacturing

(production) systems improvement and considers the

issues surrounding the realisation of sustainable

process improvement within the context of today’s

dynamic business environments In particular, the

motivations for manufacturing improvement have

been discussed and the important relationship between

quality, efficiency, flexibility and capability are

described within the context of equipment design/

redesign, improved maintenance, operator-led

im-provement, process-control, product modification

and new product introduction, quality improvement,

and tooling design and changeover improvement

Within these seven areas of manufacturing

improve-ment the principles and underlying knowledge

require-ments of a range of common improvement paradigms

are examined and six fundamental knowledge concepts

are elaborated that can be considered to represent the

understanding necessary to implement the various

tools and methods In addition to examining the

knowledge requirements of improvement paradigms

the barriers to realising sustainable improvement are

also examined through consideration and review of the

literature from the fields of manufacturing,

manage-ment and information systems These fields are selected

because of the considerable bodies of work that deal

with process improvement, change management,

in-formation systems implementation and production

systems This review reveals the importance of

under-standing and highlights the issues of a lack of

understanding, an inability to generate understanding

and an inability to communicate understanding as the

root causes of failed and partially successful

imple-mentations The issue concerning understanding and

generating understanding are further examined

through consideration of existing techniques that

support the generation of understanding within the

context of manufacturing The limitations of these

approaches and in particular, the lack of a

‘proactive-ness’ and the inability to contextualise failure with

respect to function are highlighted

In order to overcome these deficiencies, within thecontext of manufacturing systems, the concept ofmachine-material interaction is introduced and itsrelationship to ‘function’ and fundamental under-standing is discussed Using the six fundamentalknowledge requirements of manufacturing improve-ment tools, the barriers to successful implementationand the limitations of existing techniques for generat-ing an understanding of manufacturing systems, a set

of eight requirements for a new supportive ogy are developed These requirements include theneed for a functional perspective, an interdisciplinaryunderstanding, common terminology, a completeunderstanding of process efficacy, identification ofkey relationships, a structured representation and thegeneration of qualitative and quantitative rules, andthe need to provide direction for targeting improve-ments To illustrate the importance and potential of anew supportive method that meets these requirements,the relationship between the various improvementparadigms and the individual requirements aredescribed

methodol-AcknowledgementsThe work reported in this paper has been supported byDepartment for Environment Food and Rural Affairs(DEFRA) and the Food Processing Faraday KnowledgeTransfer Network, involving a large number of industrialcollaborators In particular, current research is being under-taken as part of the EPSRC Innovative Design andManufacturing Research Centre at the University of Bath(reference GR/R67507/01)

ReferencesAdams, C.W., Gupta, P., and Wilson, C.E., 2003 Six sigmadeployment Burlington, MA: Butterworth-Heinemann.ISBN 0750675233

Abdurazik, A and Offutt, J., 2000 Using UML tion diagrams for static checking and test generation In:Proceedings of the Third International Conference on theUnified Modelling Language, York, UK, 383–395.Adebanjo, D and Kehoe, D., 2001 An evaluation of factorsinfluencing teamwork and customer satisfaction Mana-ging Service Quality, 1 (11), 49–56

collabora-Ammerman, M., 1998 The root cause analysis handbook: Asimplified approach to identifying, correcting, and report-

0527763268

Antony, J., 2000a Ten key ingredients for making SPCsuccessful in organisations Measuring Business Excel-lence, 4 (4), 7–10

Antony, J., 2000b Improving the manufacturing process andcapability using experimental design: a case study.International Journal of Production Research, 38 (12),2607–2618

Antony, J and Banuelas, B., 2002 Key ingredients for theeffective implementation of Six Sigma program Measur-ing Business Excellence, 6 (4), 20–27 ISSN: 1368–3047

Apte, U.M and Goh, C., 2004 Applying lean manufacturing

principles to information-intensive services International

Journal of Service Technology and Management Fall, 5

(5–6), 488–506

Akao, Y., 1990 Quality function deployment Cambridge,

MA: Productivity Press

Bengtsson, J., 2001 Manufacturing flexibility and real

options: A review international Journal of Production

Economics, 74, 213–224

Bessant, J., Caffyn, S., and Gallagher, M., 2001 An

evolutionary model of continuous improvement

beha-viour Technovation, 21, 67–77

Bicheno, J., 2003 The new lean toolbox – towards fast,

flexible flow Buckingham: PICSIE Books

Buckhout, S., Frey, E., and Nemel, J., 1999 Making ERP

succeed: turning fear into promise Strategy and Business,

Second Quarter, 15, 60–72

Bauer, B., Muller, J.P., and Odell, J., 2001 Agent UML: A

formalism for specifying multiagent software systems

International Journal of Software Engineering and

Knowl-edge Engineering, 11 (3), 207–230

Bingi, P., Sharma, M.K., and Golda, J.K., 1999 Critical

issues affecting an ERP implementation Information

Systems Management, 16 (3), 7–14

Boothroyd, G., Dewhurst, P., and Knight, Q.A., 2001

Product design for manufacture and assembly CRC Press

ISBN 9780824705848

Brown, M.G., Hitchcock, D.E., and Willard, M.L., 1994

why tqm fails and what to do about it Business One Irwin

publishing ISBN-13: 9780786301409

Brassard, M and Ritter, D., 1994 The memory joggerII: A

guide of tools for continuous improvement and effects

planning Methlen, MA: GOAL/QPC

Bamber, C., Sharp, J., and Hides, M., 1999 Factors affecting

successful implementation of total productive

mainte-nance: UK manufacturing case study Journal of Quality

in Maintenance Engineering, 5 (3), 162–181

Carroll, J., Sterman, J., and Marcus, A., 1998 Playing the

maintenance game: How mental models drive

organiza-tional decisions In: J Halpen and R Stern, eds Debating

rationality: Nonrational elements of organizational decision

making Ithaca, NY: Cornell University Press, 99–121

Chan, F.T.S., Lau, H.C.W., Ip, R.W.L., Chan, H.K., and

Kong, S., 2005 Implementation of total productive

maintenance: a case study International Journal

Produc-tion Economics, 95 (1), 71–94

Chapman, R.L and Hyman, P.W., 1997 Continuous

improvement strategies across selected manufacturing

sectors Benchmarking for Quality Management and

Technology, 4 (3), 175–188

Christopher, M and Peck, H., 2003 Marketing logistics

Oxford: Butterworth-Heinemann ISBN 0750652241

Cohen, L., 1995 Quality function deployment: How to make

QFD work for you Cambridge, MA: Addison-Wesley

Publishing Company

Ding, L., Matthews, J., McMahon, C.A., and Mullineux, G.,

2009 An information support approach for machine

design and build companies Concurrent Engineering:

Research and Applications, 17 (2), 382–389

Davis, J.R and Davis, A.B., 1998 Effective training

strategies: A comprehensive guide to maximizing learning

in organizations San Francisco: Berrett-Koehler

Publish-ers ISBN 1576 75037X

Dewhurst, P and Abbatiello, N., 1996 Design for service In:

G.Q Huang, ed Design for X: concurrent engineering

imperatives London: Chapman & Hall

Deleryd, M., 1998 On the gap between theory and practice

of process capability studies International Journal ofQuality Reliability Management, 15, 178–191

Eti, M.C., Ogaji, S.O.T., and Probert, S.D., 2006a Strategicmaintenance management in Nigerian industries AppliedEnergy, 83, 211–227

Eti, M.C., Ogaji, S.O.T., and Probert, S.D., 2006b Reducingthe cost of preventative maintenance (PM) throughadopting a proactive reliability-focused culture AppliedEnergy, 83, 1235–1248

European Union Directive European packaging and ging waste directive 2004/12/EC of 11 February 2004amending 94/62/EC11, 2004

packa-European Union Directive Machine Safety 98/37/EC of theEuropean Parliament and of the Council, 22 June 1998.Frey, D.D., Otto, K.N., and Wysockij, A., 2000 Evaluatingprocess capability during the design of manufacturingsystems Journal of Manufacturing Science and Engineer-ing, 122 (3), 513–519

Gelderman, C and Semeijn, J., 2006 Managing the globalsupply base through purchasing portfolio management.Journal of Purchasing and Supply Management, 12, 209–217

Giess, M.D and Culley, S.J., 2003 Investigating turing data for use within design In: Proceedings of 14thInternational Conference on Engineering Design, Stock-holm, Sweden, 18–22 (10 pages on CD)

manufac-Govers, C.P.M., 2001 QFD not just a tool but a way ofquality management International Journal of ProductionEconomics, 69 (2), 151–159

Guptara, P., 1994 Lesson of experience- learning fromothers mistakes In: D Lock, ed Gower handbook ofquality management 2nd ed Gover, Brookfield VT.ISBN: 978-0-566-07451-6

Haley, J and Cross, P., 1993 Discussion in Niven, D., Whentimes get tough, what happens to TQM? HarvardBusiness Review, 14 (1), 45–61

Hammer, M and Champy, J., 1993 Reengineering the ation: A manifesto for business revolution Harper Business.Hansen, R.C., 2005 Overall equipment effectiveness (OEE).Industrial Press, ISBN-13 978-0-8311-3237-8

corpor-Hauge, B.S and Johnston, D.C., 2001 Reliability centredmaintenance and risk assessment In: Proceedings of theIEEE Reliability and Maintainability Symposium, NewYork, USA, 36–40 ISBN: 0-7803-5848-1

Henderson, K and Evans, J., 2000 Successful tion of six sigma: benchmarking General ElectricCompany Benchmarking and International Journal, 7(4), 260–281

implementa-Hicks, B.J., Bowler, C., Medland, A.J., and Mullineux, G.,

2002 A redesign methodology for analysing andimproving the performance capability of packagingmachinery International Journal of Industrial Engineer-ing, 9 (4), 389–398

Hicks, B.J., Medland, A.J., and Mullineux, G., 2001 Aconstraint based approach to the modelling and analysis

of packaging machinery Packaging Technology andScience, 14 (5), 209–225

Hicks, B.J., Culley, S.J., and McMahon, C.A., 2006 Astudy of issues relating to Information Management

Information Management, 26 (4), 267–289 ISSN 4012

0268-Hill, S., 1991 How do you manage a flexible firm? The totalquality model, Work Employment and Society, 5 (3),397–415

Holland, C.P and Light, B., 1999 A critical success

factors model for ERP implementation IEEE Software,

30–36

Hou, T., Liu, W.-L., and Lin, L., 2003 Intelligent remote

monitoring and diagnosis of manufacturing processes

using an integrated approach of neural networks and

rough sets Journal of Intelligent Manufacturing, 14, 239–

253

Irani, Z and Love, P.E.D., 2001 Information systems

evaluation: past, present and future European Journal

of Information Systems, 10, 183–188

Jones, R.B., 1997 Risk-based management Jerico Publishing,

Mumbai India

Jiao, J and Tseng, M.M., 1999 An information modeling

framework for product families to support mass

custo-mization manufacturing CIRP Annals, 48 (1), 93–98

Juran, J.M., 1992 Quality by design The Free Press ISBN

0-02-916683-7

Keating, E.K., Olivia, R., Repenning, N.P., Rockart, S.,

and Sterman, J.D., 1999 Overcoming the improvement

paradox European Management Journal, 17 (2), 120–

134

Limanond, S., Si, J., and Tsakalis, K., 1998 Monitoring and

control of semiconductor manufacturing process Control

Systems Magazine IEEE, 18 (6), 46–58

Liang, S.Y., Hecker, R.L., and Landers, G.R., 2004

Machining process monitoring and control: the state of

the art Journal of Manufacturing Science Engineering,

128 (2), 297–310

Lofsten, H., 2000 Measuring maintenance performance: in

search for a maintenance productivity index

Interna-tional Journal Production Economics, 63, 47–58

Manarro-Viseras, E., Baines, T., and Sweeney, M., 2005 Key

success factors when implementing strategic

manufactur-ing initiatives International Journal of Operations and

Production Management, 25 (2), 151–179

Marchington, M., Goodman, J., Wilkinson, A., and Ackers,

P., 1992 New developments in emloyee involvement

Employment Department Research Series No 2 London,

HMSO

Maskell, B.H., 1991 Performance measurement for world

class manufacturing: A model for american companies

New York: Productivity Press ISBN 0915299992

Matthews, J., Singh, B., Mullineux, G., and Medland, A.J.,

the ‘process flexibility’ of food processing equipment

Computers and Industrial Engineering, 51 (4), 809–820

Matthews, J., Singh, B., Mullineux, G., and Medland, A.J.,

approach to investigate manufacturing machine design

capability Journal of Mechanical Engineering, 53 (7–8),

462–477

Matthews, J., Singh, B., Mullineux, G., Ding, L., and

Medland, A.J., 2008 Modelling to reduce the

configura-tion phase time of machine design Advanced Materials

AMR.44-46.659

Matthews, J., Singh, B., Deslandes, A., Mullineux, G., and

Medland, A.J., 2009 Modelling approach for dedicated

machinery to support the effects of mass customization

International Journal of Computer Integrated

Manufac-ture, 22 (11), 1000–1011

McIntosh, R.I., Culley, S.J., Mileham, A.R., and Owen,

G.W., 2001 Improving changeover performance: a

strat-egy for becoming a lean, responsive manufacturer

Butter-worth-Heinemann

Mellor, R., Hyland, P., and Karayan, J., 2001 ContinuousImprovement tools and support mechanisms in Austra-lian manufacturing Journal of Interdisciplinary Studies,

14, 189–201

Mileham, A.R., Culley, S.J., Owen, G.W., Newnes, L.B.,Giess, M.D., and Bramley, A.N., 2004 The impact ofrun-up in ensuring Rapid Changeover Annals of CIRP,

53 (1), 407–410

Miller, R.L and Cangemi, J.P., 1993 Why Total qualitymanagement fails: perspective of top management.Journal of Management Development, 12 (7), 40–50.Mitchell, L., 2002 Preventative maintenance and RCM II.Manufacturing Engineer, August, 153–155

Mostafa, S., 2004 Implementation of proactive maintenance

in an Egyptian glass company Journal of QualityMaintenance Engineering, 10 (2), 107–122

Moubray, J., 1997 Reliability-cantered maintenance NewYork: Industrial Press ISBN 0-8311-3146-2

Murdock, J.L and Hayes-Roth, B., 1991 Intelligent itoring and control of semiconductor manufacturing.IEEE Expert, 6, 19–31

mon-Nwabueze, U., 2001 An industry betrayed: the case of totalquality management in manufacturing TQM Magazine,

13 (6), 400–408

Oakland, J.S., 2003 Total quality management: Text with

0750657405

Olivia, R., Rockart, S., and Sterman, J., 1998 Managingmultiple improvement efforts: lesson from a semiconduc-tor manufacturing site In: D Fedor and S Gosh, eds.Advances in management of organisational quality Vol 3.Greenwich, CT: JAI press, 1–55

Pahl, G and Beltz, W., 1996 In: Ken Wallace, ed.Engineering design: a systematic approach 2nd ed.London: Springer-Verlag

Pisano, G.P., 1997 The development factory: Unlocking thepotential of process innovation Harvard Business Press,USA ISBN 0875846505

Redman, T and Grieves, J., 1999 Managing strategicchange through TQM: learning from failure Newtechnology Work and Employment, 1999, 14 (1), 44–61.Reger, R.K., Gustafson, L.T., Demarie, S.M., and Mullane,J.V., 1994 Reframing the organisation: why implement-ing total quality is easier said than done Academy ofManagement Review, 19 (3), 565–584

Reik, M.P., McIntosh, R.I., Owen, G.W., Culley, S.J., andMileham, A.R., 2006 A formal design for changeover(DFC) methodology Part 1: theory and background.Proceedings of the IMechE, Part B: Journal of Engineer-ing Manufacture, 22 (B8), 1225–1235

in manufacturing Business Horizons, Elsevier, 35 (4),73–81

Rother, M and Shook, J., 1999 Learning to see: Value-streammapping to create value and eliminate muda LeanEnterprises Institute, Incorporated US ISBN 0966784308.Scholtes, P.R., Joiner, B.L., and Streibel, B., 2003 The TeamHandbook ISBN:1884731260 Madison, WI: Oriel Inc.Sivia, D., 1996 Data analysis: A Bayesian tutorial Oxford:Clarendon Press, 7–8 ISBN 0-19-851889-7

Smith, A.M., 2005 RCM: gateway to world class tenance Madison, WI: Oriel Inc Elsevier Butterworth-Heinemann ISBN: 9780750674614

main-Schaffer, R and Thomson, H., 1992 Successful changeprograms begin with results Harvard Business Review,80–89

Shingo, S., 1985 A Revolution in manufacturing: The SMED

system Cambridge, MA: Productivity Press

Slack, N.D.C., 2005 The flexibility of manufacturing

systems International Journal of Operations and

Produc-tion Management, 25, 1190–1200

Sommer, R.A., 2003 Business process flexibility: a driver for

outsourcing Industrial Management and Data Systems,

103 (3), 177–182

Sohal, A.S., Samson, D., and Ramsay, L., 1998

Require-ments for successful implementation of total quality

management International Journal of Technology

Man-agement, 16 (4–6), 505–519

Stamatis, D.H., 2003 Failure mode and effect analysis:

FMEA from theory to execution American Society for

Quality ISBN 0873895983

Swink, M., Narasimhan, R., and Kim, S.W., 2005

Manu-facturing practices and strategy integration: Effects on

cost efficiency, flexibility, and market-based performance

Decision Sciences, 36 (3), 427–457

Sterman, J., Repenning, N., and Kofman, A., 1997

Unanticipated side effects of successful quality programs:

Exploring a paradox of organizational improvement

Management Science, 43 (4), 503–521

Tari, J.J and Sabater, V., 2004 Quality tools and techniques:

Are they necessary for quality? International Journal

Production Economics, 92, 67–280

Taylor, W.A., 1997 Leadership challenges for smaller

organisations: self perceptions of TQM implementation

Omega, 25 (5), 567–579

Thomas, A.J and Webb, D., 2003 Quality systems

implementation in Welsh small- to medium-sized

en-terprises: a global comparison and a model for change

Proceedings of IMECHE, Journal of Engineering

Manu-facture, 217 (4), 573–579

Ukko, J., Tenhunen, J., and Rantanen, H., 2007 mance measurement impacts on management and leader-

International Journal of Production Economics, 110, 39–51

Uraikul, V., Chan, W.C., and Tontiwachwuthikul, P., 2000.Development of an expert system for optimizing naturalgas pipeline operations Expert System with Applications,

18, 271–282

Vesely, W.E., Goldberg, F.F., Roberts, N.H., and Haasl,D.F., 1981 Fault tree handbook NUREG-0492, U.S.Nuclear Regulatory Commission, Jan

Waeyenbergh, G and Pintelon, L., 2004 Maintenanceconcept development: A case study International Journal

of Production Economics, 89, 395–405

Willmott, P., 1997 Total productive maintenance: the westernway Butterworth-Heinemann ISBN 0 7506 1925 2.Woodcock, J.A.D., 1972 Cost reduction through operatortraining and retraining Kogan Page (Associates) ISBN:0850380111

Womack, J.P., Jones, D.T., and Ross, D., 1990 The machinethat change the world New York: Rawson Associates.Wheeler, D.J., 2000 Normality and the process-behaviourchart SPC Press ISBN 0-945320-56-6

Wilkinson, A., Marchington, M., Goodmann, J., andAckers, P., 1992 Total quality management and employ-

ee involvement Human Resource Management Journal, 2(4), 1–20

Wilkinson, A., Redman, T., Snape, E., and Marchington,M., 1998 Managing through total quality management.Macmillan Press London

Whittaker, B., 1999 What went wrong? UnsuccessfulInformation Technology projects Information Manage-ment and Computer Security, 7 (1), 23–29

A new approach for conceptual design of product and maintenance

Zaifang Zhang and Xuening Chu*

School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, 200240 PR China

(Received 1 July 2009; final version received 20 February 2010)Services such as maintenance are increasingly important for a manufactured product, which can improve customersatisfaction and promote sustainable consumption A trend has appeared that manufacturers change their attentionsfrom providing only a product to offering both product and its services as a whole However, preliminary literaturereview shows that few studies focus on integrated design of product and service In the paper, a design approach isproposed for supporting conceptual design of product and maintenance (P&M) In the layout process, the approachuses an improved quality function deployment (QFD) tool to translate customer requirements into conceptspecifications An information exchange mechanism is developed to exploit the interrelationships between P&M Inthe mechanism, a failure mode and effects analysis (FMEA) tool is used to identify and analyse failure modes andtheir effects on the product concept Then maintenance concepts are generated based on the results of QFD andFMEA The proposed approach is applied in a conceptual design case of a horizontal directional drilling machinewith its maintenance Furthermore, the paper also addresses the management and improvement of P&M concepts.Keywords: product-service system; maintenance; conceptual design; quality function deployment; failure mode andeffects analysis

1 Introduction

With the increasing need of sustainable production and

consumption, service activities, as a source of core

value, are becoming more and more important for a

manufacturing enterprise (Aurich et al 2006)

Custo-mer requirements (CRs) are shifting from purchasing

just a physical product to acquiring a result or a

function supported by the product combined with

related services (Mont 2002, Maussang et al 2005,

Baines et al 2007) Enterprises are then shifting their

business focus from designing physical products only,

to designing the offering of product and related

services which are jointly capable of fulfilling specific

CRs (Manzini and Vezzoli 2003) A product-service

system (PSS) is proposed and implemented to deal with

this issue With the assistance of PSS, appropriate

integrated concepts can be generated and sustainable

production and consumption can be promoted

The PSS was defined as ‘a marketable set of

products and services capable of jointly fulfilling a

user’s need The product and service ratio in this set

can vary, either in terms of function fulfilment or

economic value’ (Goedkoop et al 1999) Three main

categories of PSS have been identified:

product-oriented, use-oriented and result-oriented PSS Each

category has different emphasis to deliver the required

function, i.e the reliance emphasis changes from on

the product to on the service (Tukker 2004) Thepotential benefits of PSS can be summarised according

to the studies of Mont (2002) and Aurich et al (2006):PSS can improve product core competences, meetindividual CRs, change traditional stakeholder rela-tionships, and reduce environmental loads

Although product design and service design focus

on different aspects, both product and service should

be considered to satisfy CRs Maintenance is the mostefficient way to keep the function available during theproduct lifecycle (Takata et al 2004), which is selected

as the representative of services An integrated work of product & maintenance (P&M) development

frame-is proposed which can be considered as an initiatework for PSS development (future work will enlarge toall related services but not only maintenance) Gen-erally, conceptual design is one of the most importantstages because most lifecycle cost and critical perfor-mance of a product or service is determined in thisstage Pahl and Beitz (1996) also emphasised theimportance of conceptual design because it is verydifficult to correct the fundamental shortcomings in thelater embodiment and detail design phases Therefore,this study will address the conceptual design problems

of P&M, which begin with CRs and end with a set offeasible P&M concepts

Since information and activities in P&M tual design are mutually dependent, development of an

concep-*Corresponding author Email: xnchu@sjtu.edu.cn

Vol 23, No 7, July 2010, 603–618

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/09511921003736766

integrated environment to model the relations between

product design and maintenance design and to achieve

the optimal P&M concept considering both the

pro-duct quality and the maintenance quality is required

Presently, engineering design methodologies for P&M

conceptual design have not been discussed sufficiently

An integrated approach based on quality function

deployment (QFD) and failure mode and effects

analysis (FMEA) is proposed for the issue Translating

CRs is the first important step during conceptual

design (Dean et al 2009) QFD has been developed

into a proven successful methodology which can

facilitate the understanding and response to CRs in

the product/service development (Hauser and Clausing

1988, Akao, 1990, Pun et al 2000, Chan and Wu 2002,

Fung et al 2003, Bu¨yu¨ko¨zkan et al 2007, Kahraman

et al 2006, Luo et al 2008, Zhang and Chu 2009a)

An improved QFD is adopted to translate CRs into

Engineering Characteristics (ECs) of P&M, and then

product module characteristics and maintenance

stra-tegies FMEA is a systematic assessment tool for

product safety and reliability analysis in design or

other engineering fields, with the aim of preventing

potential failures (Stamatis 1995, Xu et al 2002, Teoh

and Case 2005) An information exchange mechanism

based on FMEA is proposed to tackle with the

interrelationships between product concept and

main-tenance concept Furthermore, fuzzy set theory is

incorporated with the integrated approach to capture

the vagueness and uncertainty of the designers’

judgments in P&M conceptual design stage

The paper is organised as follows The related work

is reviewed in the next section The third section

describes the framework of P&M concept design

The fourth section proposes the integrated approach

for P&M conceptual design In the fifth section the

proposed approach is applied in a real-world case of

horizontal directional drilling (HDD) machine

Con-clusions are then presented in the final section

2 Related work

2.1 PSS

Most of the early studies on PSS development were

primarily conducted from the viewpoint of marketing

and management Mont (2002) built a theoretical PSS

framework to improve core competitiveness and to

provide new business benefits Three main

uncertain-ties are analysed regarding the applicability and

feasibility of PSS: the readiness of companies to adopt

them, the readiness of consumers to accept them, and

their environmental implications Manzini and Vezzoli

(2003) described PSS as a framework of the new types

of stakeholder relationships which can produce new

convergence of economic interests and a potential

concomitant systemic resource optimisation Recently,engineering methods and tools have been developed

to support PSS development Aurich et al (2004 and2006) argued that technical services can influence theeconomic and ecologic performance of product andprovide new user benefits Based on process modular-isation, a systematic design method was proposed forlifecycle oriented technical PSS Morelli (2006) con-sidered the capability of PSS should belong to thedesign domain and proposed the tool of IDEF0(integration definition for function modelling, level 0)

to represent and blueprint a PSS In order to generateoptimal PSS concept, Sakao and Shimomura (2007)and Sakao et al (2009) introduced a new designdiscipline, Service Engineering (SE), which can allowdesigning services in parallel with products Theresearchers also developed a computer-aided designtool called Service Explorer to implement SE develop-ment In SE, a design model is composed of foursub-models: flow sub-model, scope sub-model, viewsub-model, and scenario sub-model The model wasproved to be effective through service redesign of anexisting hotel in Italy and business cases such as sellingwashing machines, providing pay-per-wash service andcleaning washing machines Extending the traditionalQFD from product into product and service, An et al.(2008) built a concrete integrated roadmap structureand provided a supporting approach for efficient road-mapping The approach divides the traditional House

of Quality (HoQ) into two main parts for product andservice Motivated by PSS, manufacturers have greateraccessibility to product lifecycle data (e.g in-serviceinformation and knowledge) which can improve thecompanies’ core design and engineering capabilities(Ward and Graves 2006) Yang et al (2009) proposed

a service enabler, an information management systemthat can receive product lifecycle data and managethem for the realisation of product-oriented and use-oriented PSS Goh and McMahon (2009) reportedtheir experiences and suggested two methods (i.e.statistical analysis and data mining) to improve reusing

of this in-service information capture and feedback

2.2 QFD and FMEA in maintenanceTakata et al (2004) analysed technical role change ofmaintenance in product lifecycle and then presented amaintenance framework that shows managementcycles of maintenance activities, including maintenanceplanning, maintenance task execution and lifecyclemaintenance management Waeyenbergh and Pintelon(2002, 2004) proposed a design framework to generatethe customised maintenance concept (i.e maintenancestrategy) which had been applied in a companyproducing cigars and cigarillos Within the framework,

the optimum maintenance concept can be selected

through identification of the most important systems

and most critical components and analysis of

main-tenance strategy decision tree (mainmain-tenance strategies

include failure based maintenance, design-out

main-tenance, detective based mainmain-tenance, condition based

maintenance, and use based maintenance)

QFD is a general tool for maintenance design

Pramod et al (2006 and 2008) proposed a maintenance

QFD model based on QFD and total productive

maintenance for enhancing maintenance quality of

product, and then checked the implementing feasibility

of the proposed method through an automobile service

station Lazreg and Gien (2009) proposed an

inte-grated model linking two popular approaches (i.e six

Sigma and maintenance excellence) to improve the

effectiveness of maintenance Coupled with QFD,

the model is used to deploy the design parameters in

order to reduce variations and time and eliminate the

occurrence errors in the maintenance process

It is important to perform appropriate maintenance

activities in maintenance management Kimura et al

(2002) developed a virtual maintenance system

using FMEA to perform appropriate maintenance

operations for manufacturing facility management

A computer-aided FMEA tool was developed for

maintenance planning considering the time-consuming

and tedious characteristics in the traditional FMEA

Park et al (2009) also developed a maintenance

sys-tem based on FMEA The syssys-tem implemented the

diagnosis item and cycle update for maintenance

of mail sorting machine The researchers described

FMEA deployment step, result and statistics analysis

in detail Echavarria et al (2007) developed a design

methodology to increase the availability of a product/

system by reconfiguring the system or subsystems The

methodology uses FMEA to identify connectivity and

criticality of components and then adopts functional

redundancies to keep the system reliability Bae et al

(2009) developed a web-based Reliability Centered

Maintenance (RCM) system for the Korean

auto-mated guideway transit train The system uses FMEA

to identify and prioritise possible failure modes of the

train The results can be updated through collecting

historical failure data and the reliability indexes

QFD and FMEA can also be ‘applied as the dual

role tools or within an engineering process

frame-work’ (Al-Mashari 2005) Ginn et al (1998) proposed

a methodology for interactions between QFD and

FMEA and then analysed their common value

throughout the product development An example of

Ford Motor Company was cited to illustrate the

advantages of the tools Chin et al (2005) developed an

integrated system framework for product design

optimisation in terms of cost, quality and reliability

considerations by using QFD, value engineering andFMEA Almannai et al (2008) proposed an integrateddecision tool based on both QFD and FMEA inaddressing technology, organisation and people atthe earliest stages of manufacturing decision-making.QFD and FMEA are adopted to identify the mostsuitable manufacturing automation alternative andthe associated risk in the manufacturing system designand implementation phases, respectively Korayemand Iravani (2008) applied FMEA and QFD duringthe robot design in order to improve its reliability andquality In their study, FMEA and QFD are used toidentify failure modes and key quality characteristics

of the robot, respectively And then corrective actionsare implemented for critical items In order to respond

to CRs effectively and efficiently, Wang and Chang(2007) developed an integrated approach for support-ing product conceptual design QFD coupled withtheory of inventive problem solving (TRIZ) assists thedesigners to find the suitable engineering parametersfor the product, and FMEA fulfils reliable analysis ofthe product

Based on the review, QFD and FMEA were usedeffectively for product or maintenance development.However, how to carry out concurrent design of P&Mshould be discussed adequately QFD and FMEAshould be improved to translate CRs into ECs ofP&M, and then concept characteristics in an integratedmanner Information exchange should be implementedfor P&M concepts Meanwhile, subjective uncertainty

is needed to tackle within the conceptual designprocess

3 The framework of P&M conceptual designProduct combining with its related maintenance, as thecore of PSS, is the main part for providing andensuring the expected function of customers Con-ceptual design is to generate the optimum P&Mconcept considering both product quality and main-tenance quality According to the study of Aurich et al.(2006), a new model for PSS concept is proposed

in Figure 1(a) Product concept is surrounded byrelated maintenance concept, which is constructed by ahierarchy product structure tree Each cell of the treerepresents a module concept Maintenance concept

is an attachment for a particular product or moduleconcept, which can be described by maintenancestrategies and maintenance actions Maintenance stra-tegies generally include failure based maintenance,condition based maintenance, use-based maintenance,and so on The parameters of maintenance strategiesalso should be determined, e.g maintenance frequency

or time interval Action module expresses a processwhich consisted of the related core activities for

implementing the service as shown in Figure 1(b) For

example, action module of part polishing may include

several activities: order assigning, part disassembling,

part polishing, part assembling and debugging

Determination of interrelationships between P&M

is a key activity in the P&M conceptual design As

shown in Figure 1(c), product risk information and

maintenance information can be seen as the key

interrelationships between product concept and

main-tenance concept A FMEA tool is used to identify the

potential failure models, analyse their effects,

deter-mine their priority and then implement appropriate

maintenance actions Each maintenance concept may

have different maintenance information which should

be returned to its corresponding product or module

concept According to the information, corresponding

improvements should be taken on the product concept

Although the mutual collaboration, the aggregating

mechanism of P&M can be established

The methodologies for product design have been

well developed and some of these methodologies can

be extended to support P&M concept development

Based on the domain theory in Axiomatic Design (Suh

1990), a three-domain framework of P&M conceptual

design is given at the top of Figure 2 It includes

requirement domain, function domain and concept

domain The requirement domain represents the total

CRs of target customers CRs can be described as a set

of features, {CRk}, where k is the total number of CRs

The function domain represents the function structure

and its ECs For a better translation, ECs are divided

into product-related-ECs and

maintenance-related-ECs {P–ECp, M–ECp}, where p1 and p2 are the

total numbers of ECs of product and maintenance,

respectively The concept domain represents product

concept and maintenance concept The product

con-cept is similar to the traditional definition, which

includes a product structure tree with module

con-cepts Considering the complexity of product,

module-level maintenance concept should be developed for

each module concept The holistic maintenance

con-cept is generated through the combinations of these

module-level maintenance concepts

4 The proposed approachThe three-domain framework is transformed into atangible quantitative developing process by incorpor-ating the improved QFD and FMEA tools as shown atthe bottom of Figure 2 The process can be dividedinto planning level and operation level

In the planning level, the QFD tool creates twointerlinked mappings The first mapping starts withCRs (inputs) and then translates these requirementsinto P-ECs and M-ECs, respectively The secondmapping follows through these ECs and translatesthem into the product module characteristics andmaintenance strategies (outputs) Compared with thetraditional QFD, several significant modificationsare done in the proposed approach First, similar tothe study of An et al (2008), two separate sets ofP-ECs and M-ECs are linked to CRs in the proposedHoQs This can be convenient for systematic analysis.The correlation matrix in the first HoQ is divided intofour matrices: two self-correlation matrices of P-ECsand M-ECs, and two correlation matrices betweenP-ECs and M-ECs The correlation matrices may beasymmetric because the correlation value of ECito ECjmay be not equal to that of ECj to ECi (Moskowitzand Kim 1997, Reich and Levy 2004) Similarly, thecorrelation matrix in the second HoQ is divided intothree matrices: two self-correlation matrices of productand maintenance, and one attaching matrix betweeneach module with its maintenance The attachingmatrix indicates the affiliations of the maintenancewith their modules In the second HoQ, the relation-ship matrix between M-ECs and maintenance strate-gies is used to express the evaluation values of eachmaintenance strategy with each M-EC Based onmultiplying them with the weights of M-ECs, main-tenance strategies are easily evaluated Generally, onlyone appropriate maintenance strategy is adopted for

a given product or module concept Considering thecomplexity, the product can be divided into severalmodules The total evaluation score for each main-tenance strategy of the product concept can bedetermined through summing the evaluation score

of each module And then the optimal one for the

product concept can be identified The outputs of the

second HoQ are also used to guide P&M conceptual

design and then feasible P&M concepts are generated

by designers

In the operation level, the main task is to determine

the maintenance actions and make the P&M concepts

match with each other An information exchange

mechanism is fulfilled by using the FMEA tool The

tool is to identify failure modes, their causes and effects

of the product concepts A module-level analysis

procedure is proposed for FMEA For a specificmodule concept, the designers use the FMEA tool toidentify and analyse the potential failure modes Forthese failure modes, a judgment should be used toestimate whether they can be eliminated If yes, thedesigners will give some suggestions for modifying themodule concept If not, other actions should be added

to prevent or monitor the failure modes For example,

if a failure mode has high risk but cannot be easilyeliminated, the action may be ‘‘sample inspecting

(0.017,0.069, 0.098,0.315)(0.007,0.037, 0.057,0.226)

(0.048,0.088, 0.107,0.178)(0.017,0.052, 0.063,0.120)

within a two-month interval’’ to prevent its occurrence.Based on the results of FMEA and maintenancestrategies, the maintenance actions can be determinedfor a specific product concept Furthermore, animproving information set can be sent back to themodule concept For example, improvements arerevising product design and then installing inspectionsensors corresponding to a maintenance action ofmonitoring the module status.

Subjective uncertainty always exists in the decisionmaking process of P&M conceptual design Fuzzy settheory has the capability for capturing uncertainty.Furthermore, the designers may give their ownjudgments to construct the HoQs in many waysdepending on their personality (Bu¨yu¨ko¨zkan et al.2007) A fuzzy group decision making method pre-viously developed by the authors is used to aggregatethe designers’ judgments (Zhang and Chu 2009b) Forconstructing the HoQs, linguistic term sets consist ofthe following nine terms, i.e very low (VL), very low tolow (VLL), low (L), medium low (ML), medium (M),medium high (MH), high (H), high to very high(HVH), and very high (VH) FMEA uses a riskpriority number (RPN), a product of the probability

of occurrence (O), the level of severity (S), and theprobability of detection (D) to assess risk values offailure modes The higher the RPN, the moreimportant and serious the failure could be Fuzzytheory has been incorporated with RPN calculation(Chang et al 1999, Braglia et al 2003, Wang et al.2009) In the study, fuzzy RPNs are still used A fuzzyten-point scale is adopted for evaluating O, S and D,i.e (1, 1, 2), (1, 2, 3), (2, 3, 4), (3, 4, 5), (4, 5, 6), (5, 6, 7),(6, 7, 8), (7, 8, 9), (8, 9, 10), (9, 10, 10) The fuzzy RPN

is the fuzzy product of O, S and D Assume two fuzzynumber (a, b, c) and (d, e, f), the product of the twonumbers is (ad, be, cf)

5 Application in P&M conceptual design of HDDmachine

HDD is a trenchless technique which has beensuccessfully applied in different areas (e.g telecommu-nication, natural gas, drainage lines, and electricinstallations) for the installation of pipelines andconduits at relatively shallow depths (Wang andSterling 2007) As the key equipment for trenchlessconstruction, HDD machine is a typical complexproduct which consists of several multidisciplinarysub-systems, e.g mechanism (including power head,drill pipe, anchor device, etc), hydraulic system, electricsystem and engine system A heavy industry enterpriseunder consideration is a large sized manufacturer ofHDD machine in China The manufacturer nowadopts PSS as its business strategic, i.e transform

(0.017,0.053, 0.064,0.129)(0.017,0.053, 0.064,0.129)

(0.011,0.045, 0.067,0.246)

(0.010,0.058, 0.091,0.350)(0.011,0.056, 0.085,0.326)(0.004,0.016, 0.024,0.098)(0.007,0.036, 0.056,0.221)

(0.017,0.076, 0.114,0.392)(0.010,0.055, 0.084,0.317)

from providing only products to offering product and

related services as a whole Under this strategy, the

conceptual design of P&M is fulfilled

5.1 Case study

Referring to the analysis of several opening

publica-tions (Bevilacqua and Braglia 2000, Waeyenbergh and

Pintelon 2002, Wang et al 2007), the alternative

maintenance strategies are given briefly as the

following:

Corrective Maintenance (CM) CM is only

applied when a system or product breaks down

Time-Based Maintenance (TBM) TBM is

planned and applied periodically to potential

failures based on module reliability

character-istics The maintenance can be divided into two

classical strategies: constant-interval

mainte-nance (constant time interval) and age-based

maintenance (age is the total operating time after

the previous maintenance action)

Condition-Based Maintenance (CBM)

Mainte-nance decision is performed based on the

measured data from an inspecting system

Predictive Maintenance (PM) Through data

analysis, PM can find a possible temporal trend

and predict controlling value And then staff can

plan the retrieval actions

Note that when using the latter three strategies,

failures cannot be totally avoided Therefore, CM as a

complementary means should be performed when

failures occur Obviously, no matter which

mainte-nance strategy is selected for a product, CM always

exists in the maintenance concept

According to a market investigation, the major

CRs of customers are identified as follows: CR1 higher

construction reliability, CR2 higher construction

ability, CR3 higher construction efficiency, CR4 good

energy saving and emission reduction level, CR5

comfort for driver, CR6 enough security, CR7

moderate cost and CR8 good technical ability for

maintenance Referring to these above requirements,

P-ECs and M-ECs are categorised respectively as

follows: P-EC1 core ability, P-EC2 module reliability,

P-EC3 controlling technology, P-EC4 security

protec-tion, P-EC5 working mode, P-EC6 energy saving and

emission reduction level, P-EC7 structure

perfor-mance, and P-EC8 slurry ability; M-EC1 reliability

and security, M-EC2 maintenance technology, M-EC3

maintenance cost, M-EC4 adding values, M-EC5

technical ability of maintenance and M-EC3 response

timeliness Six key modules of HDD machine are

specified: M1 engine module, M2 hydraulic module,

M3 electric module, M4 aiding module (including drillpipe, anchor equipment, etc.), M5 dynamic head andM6 slurry module Other modules may be importantfor implementing product function but are not as keysfor meeting CRs and then these are not considered inthe QFD tool

The two HoQs are constructed in Tables 1–2 byusing the approach of Zhang and Chu (2009b) Thetwo tables are given in partition modes because of theircomplexity Relative weights of CRs, Correlationmatrix and relationship matrix of the first HoQ arelisted in Tables 1(a), (b) and (c), respectively A scale ofseven linguistic variables, where are very low (VL), low(L), medium low (ML), medium (M), medium high(MH), high (H), and very high (VH), is adopted toexpress the importance level of ECs which may be noteasily be expressed quantitatively (e.g reliability andsecurity level) Correlation matrix of the second HoQ

is given in Table 2(a), which is just for engine module,hydraulic module and electric module Values amongother modules are equal to 0 or too low to be neededfor consideration Relationship matrix between P-ECsand product modules is given in Table 2(b) Table 2(c)gives the relationships matrix between M-ECs andmaintenance strategies of engine module (those ofother modules are similar)

In the first HoQ, the relative weights of P-ECs andM-ECs are calculated by the steps as follows: (1)multiplying the weights of CRs and relationship valuesfor P-ECs and the values for M-ECs, respectively; (2)multiplying the results of (1) and correlation (the self-correlations of P-ECs or M-ECs), respectively; (3)multiplying the results of (2) and correlation (thecorrelation between P-ECs and M-ECs), respectively;(4) calculating the relative weights of P-ECs and M-ECs based on the results of (3) Based on the relativeweights of ECs, the values of each EC can bedetermined by the designers as shown in Table 1 Inthe second HoQ, the relative weights of productmodules are calculated by the steps as follows: (1)multiplying the weights of P-ECs and relationshipvalues of modules; (2) multiplying the results of (1) andcorrelation (only the correlation of module); (3)calculating the relative weights of P-ECs based onthe results of (2) The characteristic values of modulescan be determined by designers according to therelative weights and correlated P-ECs, and the resultsare given in the second column of Table 3 In terms ofthe outputs of the second HoQ, module concepts can

be generated by the designers The relative weights ofmaintenance strategies for each module are calculated

by the steps as follows: (1) multiplying the weights ofM-ECs and relationship values of maintenance; (2)multiplying the results of (1) and the relative weights ofmodules; (3) calculating the relative weights of

maintenance by summing the results of (2) The step

(2) is aimed to calculate the absolute importance values

of maintenance strategies based on the importance of

modules For different modules of a given product

concept, there may be different optimum maintenance

strategies when only considering the characteristics of

the module itself However, adopting different

main-tenance strategies is generally improper for only one

product The optimum maintenance strategy should be

selected based on the evaluation scores of maintenance

strategies for each module Considering the

particu-larity of CM, the evaluation scores are determined by

summing the scores (of step (3)) of these modules

where CM values are not maximal

Through the FMEA tool, the designers can identify

the potential failure modes of each module concept In

a specific product concept (assume that its

mainte-nance strategy is CBM), an example of engine module

in Table 4 is given to explain how to use FMEA Only

these crucial failure modes and possible causes for the

specific concept are listed here The RPN can be

determined based on the scores of O, S and D and then

the corresponding actions, improving information can

be formed by the designers There may be different

maintenance intervals because of different O, S and D

for these modules These intervals should be

integer-proportion and then some maintenance actions can be

combined for facilitating implementation The

improv-ing measures can update the engine concept All the

results of FMEA are used to determine the

main-tenance actions of the engine concept Considering

the information which has some effects on module

concepts, the designers will rethink the module

concepts and improve some designs (e.g replace the

pressure pipe in order to enhance the bearing ability,

and add several sensors to monitor the status of

module) Based on the results of QFD and FMEA, the

feasible P&M concepts can be generated through the

combinations of different product concepts and

main-tenance concepts Three alternative P&M concepts are

given in Figure 3 The maintenance concepts are

concisely described in this figure In these concepts,

electrical module, aiding module and slurry module use

the same concepts Other concepts may be different inproduct modules or maintenance modules For exam-ple, maintenance concepts of engine module in P&M 1and P&M 2 are artificially inspecting and self-inspect-ing, respectively Dynamic head modules adopt (gearmanner and low-speed motor and gear box) and(chain manner and low-speed motor and gear box),respectively

5.2 Further analysis5.2.1 Concept selection and managementConcept decision making is the final step in the P&Mconceptual design Based on the evaluation anddecision of designers, the optimum alternative can beselected as the exploiting object The evaluation indicescan be built according to the CRs (e.g modulereliability, safety ensuring ability) and manufacturingrequirements (e.g module manufacturabiltity, modulecost) Considering uncertain conditions, fuzzy logic orother tools may be integrated for the decision-makingapproach (e.g muliti-criteria decision making, artificialintelligence techniques) (Bu¨yu¨ko¨zkan and Feyzioglu2004a, Bu¨yu¨ko¨zkan and Feyzioglu 2004b)

As argued by Gausemeier et al (2006), the ideas thatwere initially rejected should be saved and managed butnot rejected easily According to that study, the P&Mfeasible concept sets will be saved in product databaseand an analysis can be given as follows

With the changing of outer and inner conditions,some feasible concepts which have high (or low)performance may be turned into low-priority (orhigh-priority) Different from the traditional method

of discarding low grade concepts, the remaindersshould be stored in the design database after selectingthe optimum alternatives from the feasible P&Mconcepts For example, considering the customersmay prefer the performance or cost in differentapplications, which means different weights may beassigned for performance and cost, the evaluationvalues are changed and then the final optimum conceptalternative may be different The outer conditions,

proportional controlling technology, Perf.-H

M2-1,2

current protection, pressure protection, remote control technology, Perf.-H

M3-1

Figure 3 The alternative P&M concepts.

such as the upgrading of environment protection level,

some selected alternatives must be discarded but some

unselected alternatives (these alternatives may have

high cost because they require high technology

devotion to decrease emission) may be deemed as the

optimum alternatives

5.2.2 Concept improvement

The success guarantee of each product design project

can be from two main parts: designers (inner factor) and

markets (outer factor) For each design activity, the

inner process (from requirement analysis to product

manufacturing) is mainly driven by designers The

abilities of designers, e.g professional skills,

experi-ences, etc., mostly determine the success of the ongoing

developing project After the product leaves the factory,

markets will become the main body of verifying the

success of the product (outer process) The market

information should be considered as an input for

improving product and then a close-loop is formed to

ensure the project success Constrained by current

enterprises’ environments, e.g technical strength,

per-sonal skills, financing devotion level, the launched

project may be not fully met by the CRs Therefore,

feedback analysis should be taken to find the

deficien-cies which can be as the bases of improve design in the

future work by technical progress or others

SERVQUAL, developed by Parasuraman et al

(1988), is a diagnostic technique for identifying service

quality strengths and weaknesses for an enterprise.The service quality can be defined as a function of thegap between expected service and perceived service.Furthermore, Pun et al (2000) proposed a radar chart

of relative scale to implement gap analysis In thisstudy, an analytical method using radar structurebased on SERVQUAL is proposed to address the gapbetween the current performance level and expectedlevel for the delivery binding of P&M from manufac-tures to customers After the binding comes on themarket, two radar charts of absolute and weight scalesare given to analyse the gap between their current anddesired future performance level of P&M according tocustomer questionnaires as shown in Figure 4 In theproposed method, the desired P&M levels for CRsare simply assumed as a 1–5 scale, respectively Inorder to facilitate the gap analysis, the absolute values

in Figure 4(a) should be translated into weight scales.The results are given in Figure 4(b) In the translatingprocess, the most important item is selected as thebenchmarking Based on the relative importance ofCRs, the performance levels of other CRs can becalculated In this case, the importance vector of CRs

is (0.16, 0.19, 0.05, 0.14, 0.05, 0.09, 0.16, 0.16)approximately (from Table 1(a)) and the mostimportant item is CR2 Through the gap analysis inthe weight chart, the gap between current and expectedlevel for each CR can be easily compared anddetermined Obviously, CR4 and CR7 are selected asthe main gap for redesign P&M

6 Conclusions

In order to ensure functions are available and meet

CRs, integrated P&M design has become a new trend

in the design field However, the related supporting

methods and tools still have not been fully developed

This study proposed a systematic approach based on

QFD and FMEA to support P&M conceptual design,

and also attempted to address the integrated

develop-ment process The approach uses an improved QFD

tool to translate customer requirements into concept

specifications An information exchange mechanism by

using FMEA is proposed to exploit the

interrelation-ships between P&M A real-world case of an HDD

machine illustrated the use of the proposed method

and highlighted the feasibility of the method

In the P&M development, cost estimation may be

an important issue according to the studies in similar

fields (e.g mass customisation) (Tu et al 2007) Future

work will consider other services (e.g equipment,

transport and training) in the design process and then

integrate them with service execution system Some

effective tools will be incorporated with the proposed

method, e.g service analysis tool to evaluate service

performance

Acknowledgements

The project was supported by the Shanghai Technology

Innovative Activity Planning Program (09dz1124600), the

National High-Tech R&D Program for CIMS, China (No

2007AA04Z140), the Research Fund for Doctoral Program

of Higher Education, China (No 20070248020) and

Shang-hai Leading Academic Discipline (No.Y0102) The authors

express sincere appreciation to the anonymous referees for

their helpful comments to improve the quality of the paper

References

Akao, Y., 1990 Quality function deployment: integrating

customer requirements into product design Cambridge,

MA: The Productivity Press

Almannai, B., Greenough, R., and Kay, J., 2008 A

decision support tool based on QFD and FMEA

for the selection of manufacturing automation

technol-ogies Robotics and Computer-Integrated Manufacturing,

24, 501–507

Al-Mashari, M., 2005 Key enablers for the effective

implementation of QFD: a critical analysis Industrial

Management & Data Systems, 105 (9), 1245–1260

An, Y.J., Lee, S.J., and Park, Y.T., 2008 Development of an

integrated product-service roadmap with QFD: a case

study on mobile communications International Journal

of Service Industry Management, 19 (5), 621–638

Aurich, J.C., Fuchs, C., and Devries, M.F., 2004 An

approach to life cycle oriented technical service design

Annals of the CIRP, 53 (1), 151–154

Aurich, J.C., Fuchs, C., and Wagenknecht, C., 2006

service systems Journal of Cleaner Production, 14,

1480–1494

Bae, C., et al., 2009 Development of a web-based RCMsystem for the driverless rubber-tired K-AGT system.Journal of Mechanical Science and Technology, 23 (4),1142–1156

Baines, T.S., et al., 2007 State-of-the-art in product-servicesystems Proc IMechE, Part B: J Engineering Manu-facture, 221, 1543–1552

Bevilacqua, M and Braglia, M., 2000 The analytichierarchy process applied to maintenance strategy selec-tion Reliability Engineering and System Safety, 70, 71–83

Braglia, M., Frosolini, M., and Montanari, R., 2003 FuzzyTOPSIS approach for failure mode, effects and criticalityanalysis Quality and Reliability Engineering Interna-tional, 19, 425–443

fuzzy-logic-based decision-making approach for new product opment International Journal of Production Economics,

devel-90, 27–45

approach based on soft computing to accelerate theselection of new product ideas Computers in Industry, 54,151–167

group decision-making to multiple preference formats inquality function deployment Computers in Industry, 58(5), 392–402

Chan, L.K and Wu, M.L., 2002 Quality function ment: a literature review European Journal of Opera-tional Research, 143, 463–497

deploy-Chang, C.L., Wei, C.C., and Lee, Y.H., 1999 Failure modeand effects analysis using fuzzy method and grey theory.Kybernetes, 28, 1072–1080

Chin, K.S., et al., 2005 A CIMOSA presentation of anintegrated product design review framework Interna-tional Journal of Computer Integrated Manufacturing, 84(4), 260–278

Dean, P.R., Xue, D., and Tu, Y.L., 2009 Prediction

of manufacturing resource requirements from customerdemands in mass-customisation production Interna-tional Journal of Production Research, 47 (5), 1245–1268

Echavarria, E., Tomiyama, T., and van Bussel, G.J.W., 2007.Fault diagnosis approach based on a model-basedreasoner and a functional designer for a wind turbine

An approach towards self-maintenance In: Journal of

August 2007) IOP Publishing, 75 012078

Fung, R.Y.K., et al., 2003 Modelling of quality functiondeployment planning with resource allocation Research

in Engineering Design, 14 (4), 247–255

Gausemeier, J., Berger, T., Liu, Y.Y., and Mao, Q.D., 2006.Ideas management: identifying future business ideas.Journal of Engineering Design, 13 (1), 1–7

Ginn, D.M., Jones, D.V., Rahnejat, H., et al., 1998 The

‘QFD/FMEA interface’ European Journal of InnovationManagement, 1 (1), 7–20

Goedkoop, M.J., Van Halen, C.J.G., Te Riele, H.R.M., andRommens, P.J.M., 1999 Product service systems, ecolo-gical and economic basis PricewaterhouseCoopers N.V

& Pi!MC, Storrm C.S., Pre consultants

Goh, Y.M and McMahon, C., 2009 Improving reuse of service information capture and feedback Journal ofManufacturing Technology Management, 20 (5), 626–639.Hauser, J.R and Clausing, D., 1988 The house of quality.Harvard Business Review, 66, 63–73

in-Kahraman, C., Ertay, T.J., and Bu¨yu¨ko¨zkan, G., 2006 A

fuzzy optimization model for QFD planning process

using analytic network approach European Journal of

Operational Research, 171, 390–411

Kimura, F., Hata, T., and Kobayashi, N., 2002

Reliability-centered maintenance planning based on computer-aided

FMEA The 35th CIRP-International Seminar on

Manu-facturing Systems, Seoul, Korea

Korayem, M.H and Iravani, A., 2008 Improvement of

3P and 6R mechanical robots reliability and quality

applying FMEA and QFD approaches Robotics and

Computer-Integrated Manufacturing, 24, 472–487

Lazreg, M and Gien, D., 2009 Integrating six sigma

and maintenance excellence with QFD International

Journal of Productivity and Quality management, 4 (5–6),

676–690

Luo, X.G., Tu, Y.L., Tang, J.F., and Kwong, C.K., 2008

Optimizing customer’s selection for configurable product

in B2C e-commerce application Computers in Industry,

59 (8), 767–776

Manzini, E and Vezzoli, C., 2003 A strategic design

approach to develop sustainable product service systems:

examples taken from the ‘environmentally friendly

innovation’ Italian prize Journal of Cleaner Production,

11, 851–857

Maussang, N., Zwolinski, P., and Brissaud, D., 2005 Design

of product-service systems The 10th ERSCP, Antwerp,

Belgium

Mont, O., 2002 Clarifying the concept of product-service

system Journal of Cleaner Production, 10, 237–245

Morelli, N., 2006 Developing new product service systems

(PSS): methodologies and operational tools Journal of

Cleaner Production, 14, 1495–1501

Moskowitz, H and Kim, K.J., 1997 QFD optimizer:

A novice friendly quality function deployment decision

support system for optimizing product designs

Com-puters and Industrial Engineering, 32 (2), 641–655

Pahl, G and Beitz, W., 1996 Engineering design A

systematic approach Berlin: Springer Verlag

Parasuraman, A., Berry, L.L., and Zeithaml, V.A., 1988

SERVQUAL: a multiple-item scale for measuring

con-sumer perceptions of service quality Journal of Retailing,

64 (1), 12–40

Park, J.H., Kim, H., and Park, J.H., 2009 FMEA (Failure

Mode Effect Analysis) for maintenance of mail sorting

2009, ed Communications and Networking Berlin,

Springer 555–562

Pramod, V.R., et al., 2008 MQFD and its receptivity

analysis in an Indian electronic switches manufacturing

company International Journal of Management Practice,

3 (1), 82–95

Pramod, V.R., et al., 2006 Integrating TPM and QFD for

improving quality in maintenance engineering Journal of

Quality in Maintenance Engineering, 12 (2), 150–171

Pun, K.F., Chin, K.S., Lau, and Henry, 2000 A QFD/hoshin

approach for service quality deployment: a case study

Managing Service Quality, 10 (3), 156–169

Reich, Y and Levy, E., 2004 Managing product design

quality under resource constraints International Journal

of Production Research, 42 (13), 2555–2572

Sakao, T and Shimomura, Y., 2007 Service engineering: a

novel engineering discipline for producers to increase

value combining service and product Journal of Cleaner

Production, 15, 590–604

Sakao, T., Shimomura, Y., Sundin, E., and Comstock, M.,

2009 Modeling design objects in CAD system forservice/product engineering Computer-Aided Design, 41(3), 197–213

Stamatis, D.H., 1995 Failure mode and effect analysis:FMEA from theory to execution Milwaukee, WI:ASQC Quality Press

Suh, N.P., 1990 The principles of design New York: OxfordUniversity Press

Takata, S., et al., 2004 Maintenance: changing role inlife cycle management Annals of the CIRP, 53 (2), 643–655

Teoh, P.C and Case, K., 2005 An evaluation of failuremodes and effects analysis generation method forconceptual design International Journal of ComputerIntegrated Manufacturing, 18 (4), 279–293

Tu, Y.L., Xie, S.Q., and Fung, R.Y.K., 2007 Productdevelopment cost estimation in mass customization.IEEE Transactions on Engineering Management, 54 (1),29–40

Tukker, A., 2004 Eight types of product-service system:eight ways to sustainability? Experiences from Sus-ProNet Business Strategy and the Environment, 13,246–260

Waeyenbergh, G and Pintelon, L., 2002 A framework formaintenance concept development International Journal

of Production Economics, 77, 299–313

Waeyenbergh, G and Pintelon, L., 2004 Maintenanceconcept development: a case study International Journal

of Production Economics, 89, 395–405

Wang, L., Chu, J., and Wu, J., 2007 Selection of optimum

hierarchy process International Journal of ProductionEconomics, 107, 151–163

Wang, C.S and Chang, T.R., 2007 Integrated QFD, TRIZand FMEA in conceptual design for product develop-ment process In: Proceedings of the 13th Asia PacificManagement Conference, Melbourne, Australia 1085–1095

Wang, X and Sterling, R.L., 2007 Stability analysis of

a borehole wall during horizontal directional drilling.Tunnelling and Underground Space Technology, 22 (5–6),620–632

Wang, Y.M., Chin, K.S., Poon, K.K.G., and Yang, J.B.,

2009 Risk evaluation in failure mode and effects analysisusing fuzzy weighted geometric mean Expert Systemswith Applications, 36 (2), 1195–1207

Ward, Y and Graves, A., 2006 UK Lean Aerospace InitiativeReport, University of Bath, Bath

Xu, K., Tang, T.C., Xie, M., Ho, S.L., and Zhu, M.L., 2002.Fuzzy assessment of FMEA for engine systems Relia-bility Engineering System Safety, 75, 17–29

Yang, X., Moore, P., Pu, J.S., and Wong, C.B., 2009 Apractical methodology for realizing product servicesystems for consumer products Computers and IndustrialEngineering, 56 (1), 224–235

Zhang, Z.F and Chu, X.N., 2009a A new integrateddecision-making approach for design alternative selec-tion for supporting complex product development.International Journal of Computer Integrated Manufac-turing, 22 (3), 179–198

Zhang, Z.F and Chu, X.N., 2009b Fuzzy group making for multi-format and multi-granularity linguisticjudgments in quality function deployment ExpertSystems with Applications, 36 (5), 9150–9158

decision-Customer’s behaviour modelling for manufacturing planning

S Makris and G Chryssolouris*

Laboratory for Manufacturing Systems and Automation, Department of Mechanical Engineering and Aeronautics,

University of Patras, 26500 Rio, Patras, Greece(Received 23 July 2009; final version received 12 March 2010)This paper deals with a customer-driven manufacturing planning approach Manufacturers have adopted moderncommunication technologies for the information flow related to customers’ orders However, there is still highuncertainty in the information provided This work introduces a model for estimating the probability that, once acustomer has received a potential delivery date for a product, whether they will actually place the order In thisinstance the manufacturing resources should be committed to this order The Bayesian networks method is adoptedand an automotive industrial case study is discussed

Keywords: probabilistic models; mass customisation; decision support systems

1 Introduction

This work discusses a method of evaluating the

probability that a customer, under a certain delivery

time and price and given a set of factors, submits an

order for a product This is based on the Bayesian

networks method (Schay 2007) and is applied to the

automotive industry With appropriate modification,

the method is applicable to industry in general

The automotive industry for a long time has been

following the ‘push’ model; build products, based on

one or more forecasts and eventually creating stocks of

these products The dealers’ role was important since

they were obliged to acquire a minimum quantity of

the vehicles produced, and therefore, push the product

to the market The customer’s element was actually not

the driving factor in the automotive production The

automotive industry is turning from the ‘push’ model

to a customer-driven model, which necessitates that a

method be developed to quantify the customers’ likely

responses to what the automotive is offering (Michalos

et al 2010)

Today’s research on mass customisation

(Chrysso-louris et al 2008; Mourtzis et al 2008) does consider

the fluctuating demand This fluctuating demand is

based on assumptions and forecasts the production of

many products’ variants apparently without meeting

the customers’ real needs Involving the customer in

the manufacturing process is more than just

forecast-ing his possible preferences, but it actually calls for his

precise requirements to be met In the mass

customisa-tion approach, which aims simultaneously to target

scores of individual customers by offering themproduct variations, tailored specifically to fit theirindividual needs, understanding customer behaviourunder growing market stratification conditions iscritical (Pasek et al 2009) A growing interest overthe past decade in the mass customisation approachunderscores the importance of the individual consumerchoices, in terms of both their structure and para-meters (Tseng and Piller 2003)

There is a risk involved in the order promisingprocess and the method proposed in this paper hasbeen developed in order to face it This risk is caused

by the fact that a dealer provides the customer with adue date for their order As soon as a due date is given

to the customer, the production plant needs to be inposition to accomplish the order in this date There-fore, when planning the resources usage for a period oftime, it is necessary to know if this order should beconsidered or not Consequently, it is essential to have

an estimation of what the possibility will be as towhether the customer will actually place the order orwithdraw it In the case of the order being withdrawn,the planning should be adjusted So, a more accurateestimation of a customer’s likely decision will be animportant aid for the preparation of a more accurateplan for the production that will be subject to aminimum amount of changes

In order for this risk to be eliminated, the paperdiscusses a method of quantifying the likelihood that acustomer will actually place his order and therefore,help the production planner to establish a more

*Corresponding author Email: xrisol@lms.mech.upatras.gr

Vol 23, No 7, July 2010, 619–629

ISSN 0951-192X print/ISSN 1362-3052 online

Ó 2010 Taylor & Francis

DOI: 10.1080/09511921003793809

accurate plan A typical case, in which such resource

locking situation occurs, is that of the automotive

when a customer, at dealership, asks for a delivery date

that would eventually lead to locking the plants and

supply chain’s resources in order for the delivery date

to be considered rather robust

The above are demonstrated in the example given

in Table 1, where two customers are competing for the

same delivery date, however, Customer B withdraws

their order since they think that the delivery date D2

offered is late On the other hand, Customer A gets an

early delivery date D1, however, for some reason, such

as a more competitive priced vehicle offered by a

competing brand in the same time frame, he withdraws

too The method discussed in this paper aims to

address the quantification of each of the customer’s

likelihood to submit or withdraw his order and assist

in the process of providing a suitable delivery date that

will increase the sale probability The paper addresses

the uncertainty for order submission utilising the

Bayesian networks approach Production research

has adopted a number of models for dealing with the

uncertainty in manufacturing

The majority of the research effort addresses the

issues of the information flow and the coordination of

manufacturing resources across the production

net-works The internet-based communication technology

can be adjusted to the needs of the customer driven

manufacturing (Poirier and Bauer 2001, Wiendahl and

Lutz 2002, Chryssolouris el al 2003, Chryssolouris

2006, Makris et al 2008, Mourtzis et al 2008) In a

research that evaluated similar problems to the current

paper, Tolio and Urgo have discussed the problem of

production planning approaches that are considering

the availability of complete information and usually

fail to deal with real manufacturing environments,

characterised by uncertainty affecting the time that the

manufacturing operations are executed, the routing ofthe parts, the requirement of materials and theresources Their research analysed the issue of negotia-tion and planning of external resource usage, in amanufacturing system, affected by uncertainty Inparticular, the need of resources is considered un-certain and it is modelled through a scenario basedformulation (Tolio and Urgo 2007)

A probabilistic model for decision makers dealingwith the uncertainty of equipment selection process(Manassero et al 2004) has been developed andverified in automotive manufacturing, ranking a set ofalternative and assigning a probability that theranking remains stable even in case of uncertainty

in assumptions In addition, a genetic algorithmbased dynamic scheduler and a distributed, agent-based shop floor control system have been imple-mented aiming at systems which can handle criticalcomplexity, reactivity, disturbance and optimalityissues at the same time (Monostori et al 1998).Monostori discussed the process of applying tomanufacturing, pattern recognition techniques, expertsystems, artificial neural networks, fuzzy systems andhybrid artificial intelligence (AI) techniques Inaddition, hybrid AI and multi-strategy machinelearning approaches were discussed Agent-based(holonic) systems were highlighted as promising toolsfor managing complexity, changes and disturbances inproduction systems The additional integration ofmore traditional AI and ML techniques, with theagent-based approach in the field of intelligentmachines, can be predicted resulting in systems withemergent behaviour (Monostori, 2003)

In the literature, customer behaviour modelling hasbeen discussed for identifying the way that the wealth

of data in databases can be used for the evaluation ofcustomers’ preferences According to Bounsaythip andRunsala’s report, the current methods of estimating acustomer’s behaviour are based on building theirdatabases with a significant amount of data, in orderfor them to adapt to the needs of each customer.Methods such as neural networks, K-means clustering,self-organising maps, and decision trees have beenproposed in the literature for performing data miningand data clustering for customer profiling Bounsaythipand Rinta-Runsala, 2001) Song et al have developed

a methodology that detects changes of customerbehaviour automatically from the customer profilesand the data of sales at different time snapshots (Song

et al 2001) Pasek et al (2009) attempted to quantifythe way that the combined effects of individualdecision making under abundant choice conditionsimpact the model defining optimal variety on the firm’slevel, in an effort to integrate the informationflow between the product design and marketing

during the order promising process

Receive delivery date D1,

D1 D2 is late,withdrawCustomer decides

withdraw the order

t5

(Pasek et al 2009) Kwan et al (2005) developed

constructs for measuring the online movement of

e-customers, and used a mental cognitive model to

identify the four important dimensions of the

e-customer behaviour, abstracted their behavioural

changes by developing a three-phase e-customer

behavioural graph, and tested the instrument via a

prototype that used an online analytical mining

(OLAM) methodology A prototype with an empirical

Web server log file was used for verifying the feasibility

of the methodology (Kwan et al 2005)

All the methods mentioned above are actually

performing customer segmentation and profiling,

based on a wealth of data with reference to the

customer The method proposed in the current paper

needs to evaluate a customer, without having his

historical data The added value of the method is to

utilise knowledge of the domain’s specialist and to

introduce the critical factors that make a model

quantify the buyer’s likely decision The specialist’s

knowledge comprises the specific factors that influence

the customer’s decision as well as the weight of each

factor on the total decision The Bayesian network

proposed models these factors and assigns conditional

probabilities for modelling the importance of each

factor

The factors influencing the customer’s behaviour

are discussed in section 2 and the Bayesian networks’

model is developed according to these factors The use

of this method is demonstrated in a typical automotive

industrial case study

2 Model analysis

2.1 Law of total probability and Bayes theorem

This work is based on two basic mathematical

principles which are outlined below, the Law of Total

Probabilities as well as the Bayes’ theorem

According to the Law of Total Probabilities, the

probability of an incident A1 is the sum of the

pro-bability of every incident Bn, multiplied with the

probability of the incident A given the Bn (Everitt

It is possible to combine incidents and represent

them graphically, by constructing a belief network, a

typical example of which is shown in Figure 1

The Bayes’ theorem relates the conditional andmarginal probabilities of stochastic events A and B(Everitt 2006, Schay 2007)

P(AjB) is the conditional probability of A, given

B It is also called the posterior probabilitybecause it derives from or depends upon thespecified value of B

P(BjA) is the conditional probability of B givenA

P(B) is the prior or marginal probability of B,and acts as a normalising constant

This paper discusses a set of factors for modelling acustomer’s likely decision about submitting an orderfor a highly customised product or not These factorscan be used for quantifying the customer’s decision.The paper demonstrates the means of possibly utilisingthe Bayesian networks method in order for thesefactors to be modelled

2.2 Bayesian network for calculation of probability

A factor identified to be having an impact on theprobability of sale in the context of this work and casestudy is the vehicle’s driving quality This relationship

is depicted by the Bayesian network in Figure 1 Thefactor GoodDrive models the quality of drive and hasthree potential states, Excellent, Moderate and Ade-quate Similarly, the probability of sale is modelled bythe node ProbabilityofSale and has three potentialstates: high, medium and low

According to the Bayesian networks’ principles, tocalculate the state of the ProbabilityofSale, with thehighest probability to occur, it is necessary that theConditional Probabilities Table, seen in Figure 2, to bedefined This table defines the probability of the stateHighoccuring as soon as the state Excellent, Moderate,

Adequateoccur in the node GoodDrive The table has a

size of 3 6 3, that is three factors from the node

GoodDriveand three from the node ProbabilityofSale

Filling in this table is done either manually by experts

or by utilising historical data (Rabiner 1989)

This paper has defined 19 factors that influence a

customer’s decision to proceed with submitting his

order The factors are discussed later in section

‘Modelling the customer’s likely decision’ However,

in case that a second factor, such as the

Competitive-Price factor is modelled with the previous approach,

then the Bayesian network of Figure 3 is obtained

To evaluate the impact of both factors, the

GoodDrive and the CompetitivePrice factors on the

ProbabilityofSale, it is necessary that the conditional

probabilities table (CPT) be filled in as shown in

Figure 4 The table size has grown dramatically to be

3 6 3 6 3 ¼ 27, combining all three potential states

of the three nodes

The size of the CPT depends on the number ofstates (s), the number of parents (p), and the number ofparent states (sp) in the following way (Gerssen andRothkrantz 2006):

of the CPT would be: size (CPT )¼ s 6 (sp)p¼ 3 6

319¼ 3.486.784.401 It is obviously impractical to fill

in a CPT that requires this number of entries,therefore, an alternative way of approaching it isnecessary

2.3 Reverse Bayesian network for calculation ofprobability of sale

This work suggests the reverse process for conductingBayesian inference A simple example is discussed toclarify the use of the Law of total probabilities and theBayes’ theorem in this work

In the following Bayesian network, the impact ofProbabilityofSale to the GoodDrive node is examined

as shown in Figure 5 It is assumed that the probabilitythat a customer will buy a vehicle is known Therefore,the most likely state to occur from the Probabilityof-Salenode is known In addition, it is known that highprobability of sale is most likely when the drive quality

of the vehicle is also high This is modelled by theconditional probabilities table that relates the Prob-abilityofSale with GoodDrive potential states and isshown in Figure 6 Then, with the use of the Law oftotal probability, it is possible for the probabilities ofthe GoodDrive states to be calculated

P Excellentð Þ; ¼ P ExcellentjHighð ÞP Highð Þ

and CompetitivePrice impact on ProbabilityofSale

and revenue using Bayesian networks

Substituting the probabilities, we obtain:

P Excellentð Þ ¼ 0; 957911þ 0; 040390; 00

þ 10200; 00

¼ 0; 95791 that is 95; 791% ! 96%:

Therefore, given that the probability of sale is high,

the customer would prefer a car that offers Excellent

drive by 96% This is a result that also follows

common sense

Based on the above, it is possible to reverse the

inference logic by considering the Bayes theorem and

utilising the prior probability that a customer will be

having particular preferences about the vehicle’s drive

quality This is represented by the GoodDrive node,

and then the likelihood that a state of the

Probabil-ityofSale node will occur will be calculated

This is demonstrated by the example in Figure 7, by

applying the Bayes Theorem, it is possible for the node

ProbabilityofSale from the node GoodDrive to be

derived

Thus, it is feasible to reverse the probabilities

calculation method by the adoption of the Bayes

theorem This reversing approach is very important

since it can be utilised to quantify the

Probabilityof-Sale, based on a number of other factors by first

structuring a Bayesian network of factors having been

influenced by the ProbabilityofSale and then by

reversing the calculation as shown above, the

Prob-abilityofSale is finally calculated

The benefit of this reverse process is that we can

calculate the ProbabilityofSale node by filling in a

3 6 3 CPT for the link of the ProbabilityofSale with

the node GoodDrive For every additional link that isadded to the Bayesian network, following this reverseapproach, a single CPT 3 6 3 in size has to be filled in;

it is therefore a manageable task This approach will begeneralised in the following section and will bedemonstrated in the form of a realistic model

3 Industrial case studyThis section demonstrates the application of thereverse decision making method in an automotiveindustry case Initially, the most important factors thatinfluence a customer’s decision are outlined andafterwards they are modelled with the help of theBayesian networks method

3.1 Modelling the customer’s likely decisionThe major factors that influence a customer’s decision

to place a vehicle order have been identified andanalysed after industry experts have been interviewed.These factors are presented shortly as follows:

These factors have been modelled as nodes in aBayesian network Each factor has three potentialstates that evaluate the customer’s perception of the

factor For example, the factor NeedTheVehicleEarly,has three potential states:

No requirements, if the customer has no specialrequirements about the time he receives thevehicle,

Insignificant requirements, if the customer hassome preference but of minor urgency,

Specific requirements, if the customer has stricttime requirements, e.g customer needs thevehicle as soon as possible for a family trip andtheir old car has broken down

ProbabilityOfSale and GoodDrive

PðExcellentjHighÞ  PðHighÞ þ PðExcellentjMediumÞ  PðMediumÞ þ PðExcellentjLowÞ  PðLowÞ

0; 330; 957910; 33þ 0; 040390; 33þ 0; 0191020¼ 0; 959 ! 96%

ProbabilityofSale