ARTICLE IN PRESS

Engineering Applications of Articial Intelligence 22 (2009) 929938

Contents lists available at ScienceDirect

Engineering Applications of Articial Intelligence

journal homepage: www.elsevier.com/locate/engappai

Implementation of intelligent systems, enabling integration of SMEs to

high-value supply chain networks
Richard Barton a,, Andrew Thomas b
a
Cardiff Business School, Cardiff University, MEC, Queens Buildings, The Parade Newport Road, Cardiff CF24 3AA, UK
b
Newport Business School, Newport University, UK

a r t i c l e in fo abstract

Article history: In the face of increasing competition from low-cost economies, European manufacturing companies are
Received 30 April 2008 focusing on optimisation of operational activities to remain competitive. Previous work has identied
Received in revised form how companies can optimise implementation of specialist technology to improve production
26 September 2008
capability; however increasing demands in service requirements such as customisability and exibility
Accepted 15 October 2008
are often negating the localised gains in capability.
Available online 9 January 2009
Supply chain management has become an increasingly important aspect of operations improvement
Keywords: to ensure support throughout the product realisation process. The key to creating a supply chain capable
Supply chain management of this rapid response and high level of adaptability is integration of intelligent systems and
Agility
management capabilities.
Integration
A site-visit-based survey and characterisation of small and medium enterprises (SMEs), comprising
Intelligent systems
SME actual or potential supply chain components, reveals that even those with well developed capabilities
Survey and attitudes to adopting production technologies are largely not proactive with technology adoption
targeting these needs.
A review of requirements for SMEs to achieve such competitive supply chain capabilities reveals a
hierarchy of technical expertise to be developed. This is presented as an implementation strategy for
staged introduction of these tools and techniques with a view to establishing high-value supply chains
capable of withstanding business pressures from developing economies.
& 2008 Elsevier Ltd. All rights reserved.

1. Introduction (2008) and part of the Superman (ERDF funded project,

In the face of increasing competition from low-cost economies,
European manufacturing companies are focusing on optimisation
of operational activities. In addition there are operations that are
targeted as value adding opportunities such as service, customi-
sability and exibility, all of which require either signicant
resource allocation to be held in reserve at the company, or a
supply chain capable of rapid response and high levels of
adaptability to these increasing customer demands. Hence supply
chain optimisation has become an increasingly important aspect
of operations management, but there is also an increasingly
prevalent attitude of passing the customer issues directly down
the supply chain, causing instability and ineffectiveness in
meeting customer demands.
This work builds upon previous investigation into
advanced manufacturing technology (AMT) adoption in small
and medium enterprises (SMEs) described by Thomas et al.
 Corresponding author. Tel.: +44 2920 874517.
E-mail addresses: bartonr2@cf.ac.uk (R. Barton),
thomasaj3@cf.ac.uk (A. Thomas).
19992002) research ndings. Contact was maintained with
many of the original companies through both Superman2, and
Manufacturing Advisory Service (MAS), consultancy based pro-
jects. This promoted deeper understanding of the companys
business strategy and current capabilities, to the benet of
this work.
Academic work in similar themes such as Voss (1986), Wacker
(1989) and Zairi (1991) explore a range of individual case studies
and subsequent diffusion toward an AMT implementation
strategy, however these works are dedicated to achieving nite
outputs in manufacturing capability. The value of this work lies in
its comparison of these capabilities with a tandem development
of technologies to improve supply chain management, which is
previously unexplored. Following initial studies in AMT utilisation
in the 1980s, and subsequent strategies for implementation and
optimisation in the 1990s, this work advances local understanding
in SMEs of intelligent systems (IS) tools developed in the 2000s
and proposes how this can be further disseminated by use of a
staged progression toward wider integration to intelligent supply
chain systems.
Section 2 describes the aim to investigate what support-
ing technologies were being employed by SMEs to maintain

0952-1976/$ - see front matter & 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engappai.2008.10.016
 ARTICLE IN PRESS
930 R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
individual gains in production capability, and tackle the
customer service demands mentioned above, through supply
chain management techniques focusing on intelligent tools and
systems including
 planning/scheduling,
 communication,
 process optimisation and
 relationship management and negotiation.
Through the survey and results described in Sections 3 and 4 of
this paper, it was found that there were distinct categories of SME,
characterised by company strategy and level of use of intelligent
tools and systems. These companies also typically displayed
similar capabilities in the themes listed above.
Hence it was possible to accumulate some appropriate generic
rules, and an implementation plan of required improvements and
investments which would be required for companies to elevate
themselves to a status of competence allowing integration to the
high-value supply chains to which they aspire. This is described in
Section 5, along with many examples of barriers and complexities
in practical implementation, and how other developing academic
work is starting to mature and tackle these issues.
1.1. SME denition
Taken together, the UKs 3.7 million small and medium
enterprises (SMEs) account for approximately 40% of the UKs
GDP (Wheatley, 2008), but are individually experiencing difcult
economic pressures as the global marketplace becomes more
open. However, if they could work towards become a well
integrated extended enterprise network then they have a massive
potential which would threaten the security of larger single
entities. This work aims to contribute to a collective under-
standing in the sample of 150 companies surveyed. The sample is
based in Wales (UK), each company employees between 10 and
200 people and turnover is between 0.5 and 20 million pounds
sterling. This is a signicant spread but is representative of the
SME population which could benet from this work and
collaborate directly for mutual benet.
1.2. SCM introduction
In its most basic form, the supply chain is an extended network
of suppliers, factories, warehouses, distribution centres, and
retailers through which raw materials are acquired, transformed,
and delivered to customers. Traditionally the focal point is an
Original Equipment Manufacturer (OEM), however in a competi-
tive global market, the OEM is frequently being eroded in terms of
its production capability. Subsequent outsourcing is leading to
more and more complex industrial organisations, in forms of
extended, or even more extremely virtual enterprise. As this
occurs, Ounnar et al. (2007) explain that network complexity is
subsequently greatly increased due to the fact that numerous
decision centres are required to interact. The contribution and
participation of each of the partners are thus fundamental to
ensure the supply chain achieves its production potential, which
often proves to be a weakness for SMEs (Arend and Wisner, 2005).
To minimise this effect, reducing geographical distance
between the partners can become a theme for supply chain
development (Thomas and Barton, 2007), but the development of
a Local Area Supply Chain Network (LASCaN) is only as good as the
skills, organisational limitations, resource constraints and tech-
nological capabilities of the companies within that network
which, again, traditionally isolates SMEs as the weak point in
the supply chain (Neupert et al., 2006). Recent studies have
pointed to the lack of capability amongst local suppliers and
hence many OEMs are forced to move towards Global Supplier
Networks (GASCaN) to meet demand, reduced component and/or
tooling costs, improved product quality and greater product
exibility. However, these advantages can be sometimes negated
through logistical issues which make the total acquisition costs
(TAC) of a product far greater than rst imagined (Grossman and
Jones, 2002).
Competition in the future will not be between indivi-
dual organizations but between competing supply chains
(Christopher, 1992).
Hence the challenge becomes one of reducing complexity,
increasing exibility and improving speed of response to unique
order specications without incurring poor quality or high costs. A
balance is subsequently required between the leanness of the
system components and the agility of the overall network.
To achieve this, all supply chain components must employ
intelligent tools and techniques to develop intelligent systems
capable of adjusting to these changing difcult demands.
1.3. Intelligent systems introduction
Intelligent systems (IS) can be dened as systems which
process input signals to actuate an output action, the form of
which will depend on rules based on previous experiences where
the system learned which actions best let it reach its objectives.
Hence the degree of intelligence in a system relates to the
systems level of performance in reaching its own objectives.
A procedural characterisation of an intelligent system is given
by Antsaklis (1994); intelligence is the property of the system
that emerges when the procedures of focusing attention,
combinatorial search, and generalization are applied to the input
information in order to produce the output.
Key issues of the domain highlighted by Lemmon (1994), cited
from Meystel and Messina (2000) include
 A desirable property of intelligent systems is that they are
adaptivey
 Intelligence is an internal property of the system, not a
behaviory
 A pragmatic reason for focusing on intelligent control systems
is that they endow the controlled system with enhanced
autonomyy
Articially intelligent systems (AIS) incorporate additional
functionality, often through intermediary agents, to simulate,
decide and control the output signal or action. An AIS must be
interoperable with other components, such as common sense
knowledge-bases, in order to create larger, broader and more
capable A.I. systems.
The paper goes on to demonstrate how these themes actually
appear in SMEs management strategy as follows; Section 2
presents the aim of the study, Section 3 illustrates the methodol-
ogy of the survey, Section 4 outlines the survey results and
ensuing discussion, Section 5 presents a framework of integration
to a supply chain and nally Section 6 summarises the conclu-
sions drawn.
2. The need for a supply chain development roadmap
There are many benets of utilising AIS; reducing variability,
lead-time and manual intervention; and increasing reliability,
 ARTICLE IN PRESS
R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
repeatability and speed of reaction to variable demand are just a
few which are inherently difcult to replicate in systems which
try to control complex systems but rely on manual intervention
and traditional modes of communication and data exchange.
However these are all relatively short-term objectives, aimed
at improving company performance. So, the key research question
is; whether companies realise the full potential of such systems,
and how are they currently placed to maximise the impact of a full
implementation?
There are equally numerous, if not more, gains to be had from
developing a strategic plan for network wide system intelligence.
Greater product exibility, an ability to correct quality and
technical problems quickly and effectively and, the ability to
develop long-term, sustainable relationships capable of adapting
to changes in broader market demands are a few examples (Hines,
1994). However these are not gains which can be achieved
through a single development phase. Once a long-term strategic
aim has been identied, and a development plan constructed to
integrate a supply chain network, a series of phases are required
to establish: what component capabilities are currently in place,
what system capabilities are required and what skill base is in
place to develop tools which will link the two.
If manufacturing SMEs are to become sustainable then the
correct selection, purchase and implementation of technology is
essential (Beatty, 1990), never more so than in this case where it
will drive improvement in company operations. Firstly, the
technology must meet the companys overall strategic vision
and direction. Failure to do this often results in technology being
under-utilised due to the lack of demand for its capabilities or,
skills shortages inherent within the company fail to exploit its full
capabilities. Likewise, the technology can fail to extend the
companys manufacturing capabilities in that it is has too little
technical capacity thus providing only modest improvements in
company performance. Also, the SME must have the drive and top
down management commitment towards technology implemen-
tation (TI) so that they are able to maximise the opportunity
available to them.
So far, these elements are applicable to every scope of
company, but only once these basic planning elements are in
place can a company take a decision on what systems to develop,
where to apply them and how to implement them. This
implementation, as mentioned, is likely to constitute a completely
bespoke development program with generic waypoints to ensure
adherence to the planned schedule and subsequent achievement
of the strategic outcome. Hence a longer term purpose of this
research is to establish how many companies have undertaken, or
are undertaking, this exercise with a view to intelligent system
implementation, what the ndings were (is it viable for
companies to use articial intelligence in its current level of
maturity?), how they are developing such technologies and what
the key success criteria are for others to follow and realise the
potential benets highlighted by their planning phase.
The initial phase, then, is to understand the current level
of competence in the industrial sector, especially SMEs who
hold the greatest latent potential to be realised by intelligent
systems management.
3. Methodology
From the MAS Wales database, condential to MAS providers
but demonstrated by ndings of the BERR (2008), of nearly 1000
companies visited, 150 manufacturing based SMEs were selected
for a detailed survey into their technological systems. The
companies were selected from a range of industrial sectors
and all were registered as Small to Medium Manufacturing
931
Enterprises. In an effort to consider the effects of the above
complexity issues, key characteristics were chosen to select the
companies; they were in one geographical region (Wales, UK),
they are all involved with a supply-chain network at or around 1st
tier level (i.e. were not often OEMs or raw material suppliers), and
all autonomous in operations management, i.e. free of over-riding
external inuence. Acquaintance, and usually existing relation-
ships with the companies and MAS/Cardiff University, helped
ensure that appointments were made and surveys completed for
all 150 companies.
A questionnaire was devised that identied accurately the
current technological platform of the SMEs as well as dening
their aspirations towards developing their operations, company
infrastructure, nancial strength, skills base etc. It was decided
that, each SME would be assessed by a researcher rather than rely
purely upon questionnaire feedback since this allowed for a more
realistic analysis of the companys operations. The position, and
often natural disposition towards the subject of technology varied
according to interviewee, however a staff member from materials
control/purchasing often proved to have the most appropriate
systems knowledge. In a small company this may be a director
level, while a large company would frequently have a manager or
co-ordinator who would have additional business pressures to
consider when attempting to implement intelligent systems
controls. While this initial bench-marking was not signicantly
affected by this variable at this stage, it may prove to be a
signicant factor for future work across such a cross-section.
Given the diversity in scope of SMEs operating in the UK, a
common set of characteristics had to be targeted to draw
comparative conclusions which were not distorted by one
common theme. In this case Business Type, Business Strategy
and level of use of intelligent tools/systems were selected as
objects of interest to the survey. Considering three main sets of
characteristics allowed the survey to demonstrate the correlation
between the nature of the TI and how it is developed or hindered
by the commercial approach, in a range of company types. Data
gathered included relative nancial performance (prot/loss),
Likert scales of implementation of select actions/initiatives
reecting the domain of the operational management of the
company (these were usually agreed between the researcher and
the interviewee) and also higher level interpretation of the
management strategy. Because these were relative to the
company scale and scope, there was little discreet data to plot
comparatively, hence the qualication of both the qualitative and
quantitative data for presentation in the format of the summary
table shown in Fig. 1.
Regarding the level of competence in intelligent systems,
particular themes of; planning/scheduling, communication, pro-
cess optimisation and relationship management were identied
as demonstrators for prevalence toward systems intelligence and
management. Scales of presence, maturity and depth of imple-
mentation within these themes were considered for the examples
of best practice seen by the researcher during the site visit.
Correlation between the company characteristics and the scale
and scope of individual best practices were plotted on an
individual basis at this stage, hence no summary table was
produced, but at the relatively high-level specication of the
themes listed above it became increasingly evident that the
business strategy and technology platform were good indicators
of the level of AIS utilisation as described in Section 4.2.
4. Results
Initially, survey results demonstrated differentiation between
levels of company, dened primarily by business type, levels of
 ARTICLE IN PRESS

932 R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938

Business Type Business Strategy Level of Use of

Intelligent
Tools/Systems
Category 1 Single-process The company is fighting Little / no understanding of
companies based business to maintain its market the benefits to be drawn
SMEs with using conventional share, concentrating on from AI techniques. Some
stand alone production traditional QCD awareness is present but no
capabilities. techniques which measures of understanding, hence little
(66% of are unchanging. performance and internal motivation toward
companies Has the single processes to service an implementation.
surveyed) advantage of being existing customer base
flexible in small
batch sizes.

Category 2 Higher-performance Believes integration to In addition to awareness,

companies business which external process chains there is some understanding
SMEs integral implements control is essential to of AI techniques, but this is
to a systems / development of company restricted to individual(s)
successful technology without but do not tend towards experience and not
process really managing or forward planning or an supported by the business.
chain. understanding it. ability to match the Often new, hi-tech
(31% of Technology can be resource employed to companies with particular
companies low cost but is company needs. These strengths, these companies
surveyed) under-utilised due companies tend to be are likely to invest in new
to limited taken by 'technology technology/techniques and
knowledge and lack fashion'. Results in are therefore likely to
of workforce technology being under- develop tools themselves
development. utilised. over time, although this
could be hugely improved
with closer collaboration.

Category 3 Business marked by These companies have Intelligent systems are

companies its strong capacity tried and tested solutions supported by teams of
SME for technological proposed by large, hi- people, proving benefits to
developing innovation and tech customers (or management and hence
bespoke tools enterprise, offering suppliers) and have supported for further
/techniques new products or experienced significant implementation into the
to optimise innovative improvement in supply business strategy.
integration processes, creating chain relationships which
into a global new markets and have yielded
supply chain. customers. On the improvements in product
(3% of fringe of losing SME value, hence profitability.
companies status.
surveyed)

Fig. 1. SME Categorisation.

implementation of intelligent tools and also degrees of integration
of these tools/techniques into an intelligent system.
The nature of company surveyed was seen to vary, despite
relatively similar positions in a supply chain, from single process
capability companies through to companies who had established
a small network themselves encompassing a range of capabilities
and technologies. This was subsequently reected by the
companys business approach in terms of attitude to customers
and supplier development.
Secondly, there was seen to be a variable scale of intelligent
tools/systems application, according to what activity was being
targeted by the company, hence this also proved to be a reection
of the business strategy and allowed for cross referencing of
company type against tools employed. Considering the scale of
activities seen at the various levels, it was subsequently possible
to move forward and develop the models for development and
integration toward the end of this paper.
4.1. Categorisation of SME capabilities
The initial analysis of the survey data clearly showed that the
SMEs surveyed could be split into three distinct categories each
having specic characteristics when measured against current
technological platform and capabilities. Fig. 1, below, demon-
strates this categorisation.
4.1.1. Category 1 companies
These were companies that were happy with their current
customer base and had no signicant aspirations to develop their
companies through the use of intelligent systems. These compa-
nies tended to be skeptical about the benets of AIS and the
benets that it could bring by way of increased customer base,
improved technology quality, reduced product cost and improved
delivery performance, etc. It can be argued that the development
and implementation of AIS into such companies may not simply
be relevant to the markets they operate in (66% of companies were
considered to fall into this category).
4.1.2. Category 2 companies
These were companies who constantly try to improve and
develop their company operations. These companies in general
could see the long-term benets of using AIS in developing their
product or process, but did not have the in company skills base,
nancial resources and knowledge on how to introduce AMT into
 ARTICLE IN PRESS
R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
their companies (31% of companies were considered to fall into
this category).
4.1.3. Category 3 companies
These were companies who had considerable experience of
introducing advanced technologies into their company and as
such had a nancial model which returned a large amount of their
yearly prots into the purchasing of technology to continually
develop their manufacturing operations. On the basis of good
experience, such investment was replicated in AIS to improve
company performance. Subsequently, these companies are achiev-
ing a greater market share and have contracts with larger
manufacturing industries producing products in high-value
markets (only 3% of companies were considered to fall into
this category).
It is no co-incidence that the Category 3 companies were
signicantly more protable and operated in the high-value
market sectors (automotive, aerospace, medical). These compa-
nies did not necessarily employ more staff but in many cases the
AIS allowed the company to achieve higher outputs with less staff
hence contributing to greater protability.
Category 2 companies demonstrate a keen intention to
achieve the results seen in level 3 companies but lack the
capabilities to do so. These companies are valuable from an
academic viewpoint as they could act as industrial partners; an
important resource to continue to develop intelligent tools and
techniques for testing and application. However in many cases the
creation of such a partnership would be predominantly down to
the academic partner.
Category 1 companies, of which the largest proportion of SMEs
was considered to fall into, is an area that requires further
development from a practical viewpoint. It is here that the
greatest need and demand for a suitable TI strategy exists. The
companies generally tended to be less receptive to advanced
technologies generally, and the lack of in-company skills along
with the fear of the unknown prevented these companies from
making the initial commitment and investment to a strategy of
technological advancement. Most concerning was the air of
indifference which makes these companies vulnerable to loss
of custom to overseas companies who may continue to advance
technologically, little sort of crisis will awaken these companies to
the scale of the problem and the depth of change required to
achieve long-term growth from this poor position in the market.
As mentioned in the introduction, the categorisation of
companies into the above coincided with a categorisation of
technical capability in the prescribed areas of:
 planning/scheduling
 communication
 process optimisation
 relationship management and negotiation
which were all themes concentrated on at the survey stage.
4.2. Categorisation of intelligent systems application
There are seen to be numerous isolated applications of control
systems in companies, which are seen to be closely linked to some
of the distinct areas of academic work, hence it could be suggested
that some of the factors affecting academic output in the eld
are directly effecting application in industry. This is a develop-
ing eld, and there are a number of small studies/developments
in particular areas which have been drawn out of academic
expertise including
933
a. engineering and technical research, such as cutting path
optimisation, tool management, etc.,
b. business-school environment, such as uncertainty manage-
ment, logistics, etc., and
c. systems engineering such as maintenance, interoperability,
etcy
This scale of inspiration for the academic work is reected in the
scale of implementation in industry. Fig. 2, below shows how
individual examples are seen in application, and how companies
perceive them to t into the business development framework in
place. This is augmented with research by the authors and
presented here as a composite picture of how AIS are being
incorporated to operations management.
4.2.1. Process planning and control
The most frequently seen application of process control, and
basis for intelligent systems is at an individual machine or
machine centre level. Advances in off-the-shelf hardware,
rmware and software have allowed companies, such as precision
engineers, to adopt sensors and supporting systems to recognise
non-standard operating conditions in product or process and
either act accordingly or report errors as appropriate. Particular
advances in computer technology, visual recognition software,
robot integration and tools to reduce complexity of operation have
brought this level of operation into the spotlight for the levels
of intelligence being more widely disseminated from earlier
academic work. The types of company looking toward this are
the Category 1 companies exploring new opportunities and also
the Category 2 companies that have realized a competitive
advantage through strength in a particular production technology
or process chain.
4.2.2. Communication and knowledge/data transfer
The next level follows on, gathering data and monitoring
trends from continued operation, and translating this into
knowledge and rule bases to move towards autonomy of control
of the particular machine centres and explore the possibilities of
communicating with processes before and after that particular
operation in the process chain.
An alternative at this level was seen to be operating at a higher
and separate level of communication and data transfer, predomi-
nantly with regard to logistics management. Collaboration
requires individual participants to adopt simplied, standardized
solutions based on common architectures and data models
(Horvath, 2001) and work undertaken to develop Virtual
Enterprise (VE) models demonstrates the far-reaching benets of
reducing interaction complexity and increasing reaction speed.
The model builds on individual advances in communication
technology and information technology such as Radio-Frequency
Identication (RFID), Business-to-Business (B2B) e-commerce and
supply chain coordination solution; sharing data with customers
and suppliers (Hanebeck et al., 2007).
This is the start of developing the framework from which to
hang the lower level process control tools.
4.2.3. Product/process /partner selection and qualication
Once systems are in place to manage stock levels, delivery
times, production batch sizes, and other such physical informa-
tion, coordination of inventory policies adopted by different
supply chain actors should be achieved (Giannoccaro and
Pontrandolfo, 2002), and the next step is to allocate the potential
supply agents to system inputs.
Case studies in intelligent decision support systems, knowl-
edge management and direct rule-based decision-making systems
 ARTICLE IN PRESS
934 R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
Fig. 2. Intelligent tools (scope) categorisation and systems maturity (in practice).
describe how this can be achieved from a variety of signals. In a
network case, the communication system is translating the actual
customer demands into suitable signals for this level of proces-
sing, which is subsequently lling the gap between the process
control and the information control. Because these previous
levels have often been developed in isolation, this can prove to
be a complex operation and beyond any company without
specialist knowledge partners and a comprehensive plan to meet
a strategic goal.
4.2.4. Demand/supply negotiation, balancing and actuation
All of the above tend to relate to the capabilities of the
particular company being observed. What is not seen is a robust
system which is not only spanning different levels but also
bridging them; OEMs want to be able to buy nished product in
a single transaction (reducing complexity), component manufac-
turers want to supply product to a single customer, and
component nishing operations at a separate supply chain entity
are isolated in-between unless there is a single route of commu-
nication ow being self-managed by the intelligent system.
By growing this concept, the truly intelligent system has been
realized which can accept information from a customer specica-
tion, initiate production at synchronized lead-times with all
suppliers in the supply chain to order a custom-made product, at
minimum production time, to a standard cost model. Hence
this is the ultimate level of ne-tuning which requires advanced
ideals such as evolutionary algorithms to develop priorities and
multi-level decision making. If this were to be realized, then a new
category of company could be established.
The curve in Fig. 2 is demonstrating the fact that there is a
denite increase in effort and investment required in reaching the
next level of integration to a higher level intelligent system. This is
due mainly to the level of complexity incurred by enlarging the
network being considered, but also reects the nature of published
work available for training and awareness; case studies inspire
replication in particular aspects/environments in the system, hence
general education and awareness is much higher at the bottom end
of system control. However a lack of genuine understanding in
industry, especially demonstrated here in SMEs, is shown by a lack
of application of existing tools (or maybe a lack of imagination as to
where they can be applied) in a wider context. Companies in large
cited the lack specialist knowledge and programming expertise
required to modify the more generic systems described as the
reason for a slow uptake. This is coupled with a compound wave in
the curve of implementation, which represents a function of the
required depth of penetration into an extended process chain as the
scale of implementation increases. When a company reaches a
level it has to ne-tune, consolidate and prepare in great detail
before attempting to move up to the next level.
4.3. Benets of realising a holistic capability supply chain
SMEs are slowly coming to realise that their growth, competi-
tiveness, and success depends on successful introduction of new
and advanced manufacturing technologies to the market in a
timely fashion. In this era of mass customisation, a reduction in
the period of time between design and delivery of the product to
the customer can lead to improved protability, higher levels of
 ARTICLE IN PRESS
R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
customer satisfaction and increased market share (Chen and
Kleiner, 2001). In order to achieve this time reduction, new
technology that is accurately matched to company requirements
can greatly assist companies in achieving this goal. Support for
this view is provided by Ettlie (1990) who nds that rms must be
more innovative in new processing technologies and management
practices, in order to improve their competitive positions and
ensure survival. However, TI is one of the riskiest, yet most
important endeavours of the modern company.
It is essential then, that new manufacturing technologies are
targeted by supporters of AIS for provision of uniformity of
information, of hardware which promotes interoperability and of
an interface which does not exclude all but the academic creators
of the system. The ideal would be to integrate development of AIS
structures in accordance with development of new manufacturing
hardware, but, as described above, AIS require detailed planning,
specication and vital ne-tuning phases before implementation
is possible, which will not even begin until the manufacturing
technology is already in development. So components of a system
must be put in place before the detail is known, which requires a
generic framework to be put in place now. To enable this, the
generic structure of supply chain development is now examined
in terms of how the described company and technology categories
can be assimilated into a supply chain network.
4.4. Barriers to implementation
The company management team invariably considers basic
aspects of performance, to which any new system must contribute
in order to secure capital expenditure at board level.
 Compress their design cycle so that new product introduction
and current product re-design is undertaken on a more
frequent basis.
 Employ lean manufacturing strategies in order to reduce the
cost of manufacture of their products. This includes the
development of single unit ow approaches and the achieve-
ment of signicant reductions in inventory levels.
 To be more responsive to customer demands.
 Improve quality levels.
 To respond to these ever increasing demands for a customised
product which suits the need of the customer (whether it be
colour, material spec, delivery times, etc.) and so.
 Flexibility of service and quicker response times are essential.
While, as previously described, the advantages of close integra-
tion to a supplier are:
 A greater understanding of the companys manufacturing and
service requirements (Youngdahl and Loomba, 2000) and the
ability to see how other suppliers meet these targets.
 A greater appreciation of the technological levels required to
compete in a global manufacturing environment.
 The forging of a strong supply chain relationship with the
customer over a period of time which will incrementally
develop the supplier to meet the companys immediate and
future needs.
Further complexity is added according to the nature of the
product, market and physical production environment. Confound-
ing supply chain related factors include:
 the number of parts in a particular product,
 the quality of these parts for either/both functionality and
durability required,
935
 the expected delivery time to the end-user,
 a wide range of sub-groups for each product family,
 each product group requires a unique process chain,
 primary suppliers of parts for the different product families are
based in different countries,
 many components require different nishing/secondary opera-
tions and as such, require secondary suppliers to undertake
this work,
 the level of customisability offered, which creates supply chain
system complexity.
All of these points demonstrate the difculty in logistics,
and the complex nature of the total supply chain for both
individual products, and the exibility required to cope with a
wide variety of mixed volume parts. These issues directly point to
the benets of replacing manual or semi-automatic systems
with intelligent tools, but also highlight that these have to
be unique to every network, and often bespoke to a value chain
within it.
This raises the question of whether it is economically viable to
bring in expertise and investment in hardware, software, training
and time in implementation. Potential monetary savings on
various components can vary signicantly according to the level
of capability of the existing supplier, the cost of material and
geographical proximity to the point of use of the product. On the
one hand, investment in intelligent systems will yield the benets
described above. While on the other hand, the narrow focus on
individual product cost takes no account of the other costs that
will have an effect on the company through the development of an
extended supply chain system and these savings can be slowly
eroded over time by specied order quantities and the stock
holding required to replicate a similar level of exibility. What
was once a worthwhile venture can quickly become a less than
lucrative arrangement. The key then is to complete a comprehen-
sive total acquisition cost (TAC) analysis before embarking on the
journey of supplier integration vs. re-negotiation.
One particular aspect inhibiting close integration, and limiting
the inuence of AIS, is proximity of the supplier. Particular
barriers restricting international interoperability include lan-
guage, law, information structure and business processes which
dilute the benets of a potentially slick control system. The
problem is that many traditional suppliers are unable to cope with
the demands of potential customers. Work by Thomas and Webb
(2002) support this statement and highlight the problem that
many UK supply companies do not have the technological
capability and human resource capacity to cope with the ever
increasing demands from customers. Thomas and Barton (2007)
identify the fact that due to many companies deciding to
outsource through global supply networks, local suppliers have
been unable to compete often deciding to operate in lower, less
technologically advanced manufacturing areas. The problem now
comes that when customer turn their attention back to the local
suppliers, the supply companies do not have the technology or
supply logistic capacity to cope.
As mentioned, the availability of knowledge is currently
perceived to be low, and production and intelligent systems skill
bases are mostly developed in isolation. This paper itself is
attempting to close this gap by highlighting the benets of
AIS to a business management audience and proposing that the
synergy of best practice development would deliver tangible
benets to companies.
Another particular threat to the sustainability of AIS is also
related to company culture, but demonstrated in the lack of
production information; do SMEs actually record information? For
example, maintenance of SAP is seen as a major drain on resource,
beyond that of an SME, but for those that embrace the control
 ARTICLE IN PRESS

936 R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938

system and have the knowledge to adapt it, and adapt to it, the
benet is a signicant overall cost saving.
As a rst attempt to close the gap between the current
company capabilities identied in this study, and the scenario that
would deliver the benets described in Section 4.3, a structured
implementation plan at top level, with a lower level plan for
project initiation is proposed below.
5. Framework of integration to a supply chain
The key to the development of an integrated supply
chain system is to initially identify the level of integration
the suppliers will be expected to achieve (Gattorna, 1998).
Baramichai et al. (2007) promote the relation of business
changes with appropriate approaches for supply chain congura-
tion and supplierbuyer relationship establishment to deter-
mine the business processes and the infrastructures needed to
support agility.
However, opposed to this is the traditional tendency toward
leanness to control costs, inhibiting system exibility by reducing
fat on the bone but maintaining liquidity of the supply chain
components. There should be a clear integrated strategy that
removes the competing issues of lean and agility (Pham and
Thomas, 2005).
With these points in mind, the initiation of an intelligent
system must rely on a robust existing platform which is not
heavily biased toward lean or agility and identify appropriate
suppliers and trading partners to form a supply network on which
they depend (Choy et al., 2004).
Customisable parts primarily due to their low volume will
carry a higher premium and therefore the TAC will make the
development of a local supply chain network highly desirable.
Alternatively, a global network operating at a high level of
intelligence will intuitively overcome many of the issues prevalent
in wide networks today, so the ideal is to identify parts which can
be produced in a medium range by the most able suppliers
capable of developing AIS in line with OEM growth.

Accurately Calculate TAC for each part Draw up preliminary list of

within the manufacture of the product appropriate suppliers within agreed

Identification of critical/valuable parts for manufacture by

integrated suppliers. Generally the low volume products where
the TAC can be reduced through use of intelligent system
management

Select suitable suppliers that could be developed to

manufacture product base

Technical
Process Engineer allocated to work with supplier. Product
Development
identified and product standard is analysed. Data specification
Stage
and standardisation requirements are set.
-
process control
Action plan developed, process capabilities are analysed, and level
process controls put in place accordingly.
Prototype production undertaken.

Process settings reviewed after successful pilot. Process control Capacity

systems put in place. Development
Stage
-
Current system capabilities are established. Higher volume Data transfer &
production runs developed. communication
Control systems are integrated with production systems

Developmental work and adjustment to manufacturing system

to meet variable demand.

Logistics and
Production / Process control model to replicate non standard Continuous
specification fine tuning of system to highlight priorities and Development
potential selection/decision points Stage
-
Demand / supply
Development of continuous improvement plan year on year. Set negotiation
process indices and throughput analysis to measure continuous
improvement.

Fig. 3. Intelligent systems implementation strategy to a specic value chain.

 ARTICLE IN PRESS

R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938 937

Company Operations
Analysis

Financial Analysis Project Analysis and

Parameter
Specification
Company Structure
and Management
Process Modelling
Technology Platform Customer Analysis
and Future and Data Alignment
Strategies
Process Control

Workforce Skills Technology /

Analysis Performance Analyse Technology
Wider Alignment
Implementation Plan
Recruitment of
knowledge / System Structure
Modelling & Test Analyse Product
research / academic
partner Adjust Product and
Information / Data Identify areas Process Features
Transfer requiring
improvement & use Realise Benefits of
system to select re-design through
Risk and Failure
process / product Active System
Analysis
parameters Negotiation

Verify success of
adjustments

Company Analysis and System Planning System and structure

Systems / Process Engineering Stage
Planning Stage Stage specification Stage

Fig. 4. Implementation strategy for integration to a supply chain network.

The actual prediction of what scale of company, and what cost
of product is ideal is not yet obvious and will require continued
measurement of networks being developed.
Fig. 3, below, demonstrates the planning and development
phases of developing such suppliers, and works toward a wider
implementation plan, shown in Fig. 4. This, in turn, incorporates
the required technology phases described in Fig. 2, which
essentially draws on the selection of Category 2 or Category 3
companies as described in Fig. 1.
Hence an overall plan is constructed for identifying suitable
products, suitable suppliers and potential progression toward an
integrated supply chain. While this is not prescriptive route for a
company to follow, undertaking these activities will establish
whether an SME is well advised to start progressing toward an
AIS, what its long-term aim is and what components, suppliers
and technology bases will be appropriate for achieving the
required results.
The following short section details some specic tools and
approaches an SME could consider as part of the approach to
establish an AIS, when it is new to the company technology
(Cooper, 1993), with reference to recent applied case studies to
inspire further reading.
A good understanding of supply chain ows is a prerequisite for
intelligent supply chain management. These ows form the basis
for the structure of the supply chain elements and their
interconnections. Measures from network theory Srinivasan
(2007) are proposed as viable tools for understanding and
designing supply chain structures. Symeonidis et al. (2008)
proposes that subsequent uctuations in ow and demand can
be derived from adoption of specic data-mining techniques, by
selecting the appropriate metrics for agent performance evaluation.
And Choy et al. (2007) follow up with an argument that a hybrid
on-line analytical processing (OLAP)/articial neural networks
(ANNs)/case-based reasoning (GBR) approach can be used to
predict future customer demands and allocate suitable suppliers
during the order fulllment process. While ne-tuning using
intelligent tools can potentially yield a collaborative process with
an intelligent tool to optimise negotiation (Caridi et al., 2005).
The structure of the supply chain itself can be considered one
of many states, but for the benet of completion, an ideal situation
is to model a closed-loop supply chain (Ji, 2008). As such, it can be
modeled as a knowledge management system with comprehen-
sible startmiddleend chronology.
Within the structure of an integrated AIS the supply chain can
be considered to compose of a set of intelligent software agents
(Fox et al., 2000). The network itself relies on the use of an agent
building shell, providing generic, reusable components which
collect and manage explicit knowledge of formal decision events.
Sanin and Szczerbicki (2007) proposes that set of experience
knowledge structure (SOE) has been shown as a tool that would
be a useful technology within many different intelligent systems
and platforms.
For complex dynamics and distributed supply chains, to
congurations viz. the make-to-stock (MTS) and the assemble-
to-order (ATO) have been analysed by a novel high intelligent time
(HIT) petri net, as described by Jain et al. (2006) to compare
relative merits, which may vary according to the nature of the
suppliers an subsequent process chain set up.
 ARTICLE IN PRESS
938 R. Barton, A. Thomas / Engineering Applications of Articial Intelligence 22 (2009) 929938
6. Conclusions and further work
From this work, it is evident that SMEs do not currently utilise
intelligent systems to manage supply chain relationships, and that
more importantly it can also be said that the majority of SMEs do
not have sufcient capability, or experience, to embark on an AIS
implementation with any condence in success. There are examples
of proactive management using intelligent tools, but integration to a
higher, complete, level eludes even many knowledgeable compa-
nies. Academic interest in furthering practical implementation lies
predominantly in this area of systematic implementation.
The majority of SMEs surveyed have only a basic comprehen-
sion of what is possible, and require partnership with a knowledge
partner to start the process.
Hence there is an immediate gap apparent in academic
concentration and industrial requirement. This paper aims to
describe what activities can be undertaken prior to engagement
with a knowledge-base partner, that will put an SME in a position
of understanding of what can be achieved provide a suitable basis
for implementation.
Finally, a number of academic works are referenced, not as a
literature review as such but as an idea of the scope of application
of intelligent tools available today.
Such is the bewildering array of studies, models, suppositions
and proposals that it is hardly surprising that an SME with little
knowledge of AIS tools and techniques does not know where to start
or what could be achieved. It is hoped that this paper serves to
simplify the process at least at the front end by proposing where an
SME can categorise itself, what subsequent levels of technology lie
ahead on the road to system implementation and nally what
constituent activities can be undertaken to anticipate any difculties.
Future work can potentially expand on the implementation
model proposed, which requires rigorous validation in context of
SME improvements. Such a study could be based on a number of
test bed companies that are typical of the existing category-one
family of companies, hence aim to undertake the journey toward
supply chain integration using the model as a strategic guideline.
Subsequently, additional detail can be developed for each stage of
development highlighting specic tools, techniques and imple-
mentation practices which have ensured success.
Further to this, the categorisation of companies and intelligent
systems can be rened to identify particular stagepoints in the
process beyond the basic 123. This will allow companies to
audit themselves against required capabilities and identify a
suitable starting point to achieve improvement as quickly and
efciently as possible.
Success is the key, and as this area of research continues to
grow rapidly, association of companies to knowledge, research or
academic expertise is essential until the depth of understanding
in the wider business community increases to a level of self-
perpetuation which the likes of lean manufacturing is enjoying
today. To this end it is also important that collaborations such as
this special edition are emulated in order to continue to integrate
research centres operating in a variety of environments.
References
Antsaklis, P., June 1994. Dening intelligent control, Report of the task force on
intelligent control. IEEE Control Systems, 0272-1708/94, pp. 15 and 5860.
Arend, R.J., Wisner, J.D., 2005. Small business and supply chain management: is
there a t? Journal of Business Venturing 20, 403436.
Baramichai, M., Zimmers Jr., Emory W., Marangos, C., 2007. Agile supply chain
transformation matrix: a QFD-based tool for improving enterprise agility.
International Journal of Value Chain Management 1 (3), 281303.
Beatty, C., 1990. Implementing advanced manufacturing technology. Business
Quarterly 55 (2), 4650.
BERR, 2008. Manufacturing Advisory Service Activity Report 20072008, seen at
/http://www.mas.berr.gov.uk/media/media/activity_report2007-2008.pdfS.
Caridi, M., Cigolini, R., De Marco, D., 2005. Improving supply-chain collaboration by
linking intelligent agents to CPFR. International Journal of Production Research
43 (20), 41914218.
Chen, Y., Kleiner, N., 2001. New developments in creating cycle time reduction.
Management Research News 24 (3/4), 1721.
Choy, K.L., Lee, W.B., Lau, H., Lu, D., Lo, V., 2004. Design of an intelligent supplier
relationship management system for new product development. International
Journal of Computer Integrated Manufacturing 17 (8), 692715.
Choy, K., Tan, K., Chan, F., 2007. Design of an intelligent supplier knowledge
management systeman integrative approach. In: Proceedings of the I MECH
E Part B Journal of Engineering Manufacture. Vol. 221(2), pp. 195211.
Christopher, M., 1992. Logistics and Supply Chain Management: Strategies for
Reducing Costs and Improving Services. Pitman, London.
Cooper, R.G., 1993. Winning at New Products: Accelerating the Process from Idea to
Launch, second ed. Addison Wesley, Reading.
Ettlie, J.E., 1990. Intrarm mobility and manufacturing modernization. Journal of
Engineering and Technology Management 6, 281302.
Fox, M.S., Barbuceanu, M., Teigen, R., 2000. Agent-oriented supply-chain manage-
ment. International Journal of Flexible Manufacturing Systems 12 (23), 165188.
Gattorna, J.L., 1998. Strategic Supply Chain Management: Best Practice in Supply
Chain Management. Gower Publishing, Aldershot.
Giannoccaro, I., Pontrandolfo, P., 2002. Inventory management in supply chains:
a reinforcement learning approach. International Journal of Production
Economics 78 (2), 153161.
Grossman, S., Jones, M., 2002. Overseas outsourcingmaking it work. IEE
Manufacturing Engineer Journal 82 (2), 211217.
Hanebeck, Hanns-Christian, Raisinghani, M., 2007. Delivering on the promise of
auto-ID through intelligent agent technologies. International Journal of
Innovation and Learning 4 (4), 411424.
Hines, P., 1994. Internationalization and localization of the kyoryoku kai: the
spread of best practice supplier development. The International Journal of
Logistics Management 5 (1), 6772.
Horvath, L., 2001. Collaboration: the key to value creation in supply chain
management. Supply Chain Management: An International Journal 6 (5),
205207.
Jain, V., Wadhwa, S., Deshmukh, S.G., 2006. Modelling and analysis of supply chain
dynamics: a high intelligent time (HIT) petri net based approach. International
Journal of Industrial and Systems Engineering 1 (12), 5986.
Ji, G., 2008. Performance evaluation of complaint management and virtual
enterprise in closed-loop supply chains by using exergoeconomics and
extenics. International Journal of Services and Operations Management 4 (3),
368397.
Lemmon, M.D., June 1994. On intelligence and learning, report of the task force on
intelligent control. IEEE Control Systems, 0272-1708/94, p. 63.
Meystel, A., Messina, E., 1719 July 2000. The Challenge of intelligent systems.
In: Proceedings of the 15th IEEE International Symposium on Intelligent
Control (ISIC 2000), Rio, Patras, GREECE.
Neupert, K.E., Baughn, C.C., Dao, T.T.L., 2006. SME exporting challenges in
transitional and developed economies. Journal of Small Business and
Enterprise Development 14 (4), 535545.
Ounnar, F., Pujo, P., Mekaouche, L., Giambiasi, N., 2007. Customersupplier
relationship management in an intelligent supply chain network. Production
Planning and Control 18 (5), 377387.
Pham, D.T., Thomas, A., 2005. Fighting t factories: making industry lean, agile and
sustainable. Manufacturing Engineer 84 (2), 2429.
Sanin, C., Szczerbicki, E., 2007. Genetic algorithms for decisional dna: solving sets
of experience knowledge structure. Cybernetics and Systems 38 (56),
475494.
Srinivasan, R.S., 2007. Application of network methods in intelligent supply chain
management. International Journal of Services Operations and Informatics 2
(3), 278299.
Symeonidis, A., Nikolaidou, V., Mitkas, P., 2008. Sketching a methodology for
efcient supply chain management agents enhanced through data mining.
International Journal of Intelligent Information and Database Systems 2 (1),
4968.
Thomas, A., Barton, R., 2007. Integrating local suppliers in a global supply network.
Journal of Manufacturing Technology Management 18 (5), 490512.
Thomas, A.J., Webb, D., 2002. The College Business Partnership Scheme and its
Impact on the Technical Development of SMEs in South Wales. Conference
paper of the 18th National Conference on Manufacturing Research, Leeds
Metropolitan University.
Thomas, A., Barton, R., John, E., 2008. Advanced manufacturing technology
implementation: a review of benets and a model for change. International
Journal of Productivity and Performance Management 57 (2), 156176.
Voss, C.A., 1986. Managing advanced technology management. International
Journal of Operations and Production Management 6 (5), 47.
Wacker, J.G., 1989. Technological competance, technology aversion and strategic
manufacturing competitiveness. International Journal of Operations and
Production Management 9 (3), 3947.
Wheatley, M., 25th September 2008. The Importance of Project Management.
seen at /http://www.knowledgetrain.co.uk/project-management-training-
course-article004.phpS.
Youngdahl, E., Loomba, P.S., 2000. Service driven supply chains. International
Journal of Service Industry Management 11 (4), 329347.
Zairi, M., 1991. Competing in manufacturing with AMT. Journal of Integrated
Manufacturing Systems 2 (3), 412.