ARTICLE IN PRESS

Int. J. Production Economics 90 (2004) 129–149

Business process modelling: Review and framework

Ruth Sara Aguilar-Save! n*
.
Department of Production Economics, Linkoping .
Institute of Technology, SE 581 83 Linkoping, Sweden

Received 18 April 2002; accepted 9 April 2003

Abstract

A business process is the combination of a set of activities within an enterprise with a structure describing their logical
order and dependence whose objective is to produce a desired result. Business process modelling enables a common
understanding and analysis of a business process. A process model can provide a comprehensive understanding of a
process. An enterprise can be analysed and integrated through its business processes. Hence the importance of correctly
modelling its business processes.
Using the right model involves taking into account the purpose of the analysis and, knowledge of the available
process modelling techniques and tools. The number of references on business modelling is huge, thus making it very
time consuming to get an overview and understand many of the concepts and vocabulary involved. The primary
concern of this paper is to make that job easier, i.e. review business process modelling literature and describe the main
process modelling techniques. Also a framework for classifying business process-modelling techniques according to
their purpose is proposed and discussed.
r 2003 Elsevier B.V. All rights reserved.

Keywords: Business process modelling; Enterprise modelling; Modelling tools; Modelling techniques

1. Introduction an understanding of the business processes of an

organisation. Furthermore, business processes are
Value-adding processes have become more and a key factor when integrating an enterprise
more the principle of organising the business, (Aguilar-Save! n and Olhager, 2002). Conceptual
rather than a functional hierarchy perspective. modelling of business processes is deployed on a
Some business processes application examples can large scale to facilitate the development of soft-
be found in Swanson (2003), Artiba (2001), Guinet ware that supports the business processes, and to
(2001), Mart!ınez et al. (2001), Al-Mubarak (2003) permit the analysis and re-engineering or improve-
and Chan (2002) to mention a few. Hence the ment of them.
modelling of business processes is becoming Even though it was 1960 when Levitt first
increasingly popular. Both experts in the field of mentioned the importance of business processes
Information Technology and Business Engineering it was not until the last decade that processes
have concluded that successful systems start with have acquired a real importance in enterprise
design (Levitt, 1960). Authors such as Harrington
*Tel.: +46-13-285785; fax: +46-13-288975. (1991), Davenport (1993) and Hammer (1990),
E-mail address: [email protected] (R.S. Aguilar-Sav!en). among others, promoted the new perspective. The

0925-5273/03/$ - see front matter r 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0925-5273(03)00102-6
 ARTICLE IN PRESS

130 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

increasing popularity of business process orienta- published by ‘‘Information and Software Technol-
tion (Hammer and Champy, 1993) has yielded a ogy’’ (Macintosh, 1993).
rapidly growing number of methodologies, and The focus of the present review therefore is on
modelling techniques and tools to support it. The available techniques and tools explicitly aimed at
process of selecting the right technique and the modelling business processes. These last three
right tool has become more and more complex not words have been the key during the search.
only because of the huge range of approaches Nevertheless, the author is aware that other
available but also due to the lack of a guide that techniques might exist that are used or that might
explains and describes the concepts involved. For be used for modelling business processes, which
example, when searching the Internet for guides on are not referred in the literature. Such techniques
business process modelling many thousands of have not been identified in this review since all
references may be found. There is therefore a need modelling techniques applicable to processes in
to aid practitioners and academics alike in filtering general are applicable to business processes in
the wealth of data available on business process particular. However, whether all business process
modelling so that they do not spend excessive time modelling techniques and tools are applicable or
and effort in undertaking repetitive searches. not to process modelling is beyond the scope of
Instead, they can dedicate themselves to reviewing, this paper. The author of this paper confines
understanding and applying many of the asso- attention to business processes and the set of
ciated concepts and vocabulary. modelling techniques applicable to them. In the
The aim of this paper is two-fold. On the one remainder of this paper the terms ‘‘processes’’ and
hand, it is to review a number of business process ‘‘modelling techniques’’ are used in this more
modelling techniques and tools. On the other limited sense.
hand, it is to propose a framework to classify the This section presents the background of the
techniques according to their purpose as a guide present research, its aim, and the research
to practitioners and academics who may need to approach followed as well as its scope. Section 2
choose from these techniques. builds the proposal framework used in Section 5
As the research approach followed, the author based on the literature survey and highlights some
conducted a literature search on the state of relevant associated concepts. Section 3 describes
business process modelling techniques and tools briefly the main process modelling techniques
using as research sources scholarly and trade where some of their key factors are identified.
literature both in scientific journals and material Also some of their strengths and weaknesses are
on the web. Searching on the Internet yielded a discussed. Section 4 describes some of the generic
considerable amount of data although as an initial methodologies with their process modelling cap-
source of information it was difficult to filter. Web abilities. A classification of the techniques de-
sources are useful for further detailed and specific scribed in Sections 3 and 4 is proposed in Section 5
information on certain techniques or tools and as a framework to facilitate their selection. Finally,
especially in identifying tool availability and some conclusions and ideas for future research end
potential vendors. With the aid of a number of the paper.
databases, such as Cambridge Scientific Abstracts-
Internet Database Service, IEEE Xplore and
Compendex by Engineering Information Inc., more 2. Proposal framework and literature survey
than 7000 scientific journals and conference
proceedings since 1985 were reviewed. The key To study and understand systems, one con-
words used during the search were business process structs models according to particular view-
modelling and/or review or framework. Most of the points and using particular modelling technique.
papers found are published in journals or proceed- Kettinger et al. (1997a) present an important
ings related to Information Systems or Computer overview of methods, techniques, and tools used
Sciences. The oldest paper is dated 1993 and in Business Process Re-engineering (BPR). As part
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 131

of that study a list of some related business 1. initial—setting up of processes,

process modelling techniques and tools was 2. repeatable—repeatable processes,
published. This list does not give detailed descrip- 3. defined—documented processes standardised
tions of the techniques nor the tools. However, it throughout an organisation,
has been the starting point of the research 4. managed—measured and controlled processes,
presented in this paper, which gives a more and
thorough overview with detailed analysis of the 5. optimising—continuous process improvement.
mentioned techniques in Kettinger et al. (1997b)
It is easy to imagine that for each level different
and others.
models are needed. Levels 1–3 require models
It is important to identify the uses or purposes of
whose purposes are to describe the process and
the models when undertaking modelling of any
thus knowledge of the processes to be captured and
kind. It seems clear that in order to choose the
analysed. Levels 4 and 5 require models whose
right technique, the modeller must know the
purposes are decision support in order to monitor
purpose of the model to be constructed. Different
and control processes. Macintosh (1993) proposes
techniques are more suitable to certain purposes,
to define enriched representations of processes and
e.g. one thing is a model, which describes the
the use of knowledge-based approaches to design
process, and another a model to build a system to
new intelligent tools to model business processes.
control the process.
This permits representations of activities including
For instance, Phalp et al. (1999) distinguish
time, resources, causality and authority although
between two uses of business process models: one
no further details are given. According to Macin-
for traditional software development, and another
tosh business process models can be used explicitly
to restructure business processes. In Phalp (1998)
to provide decision support in: analysis, process/
the latter purpose is explained further arguing that
plan instantiation and re-engineering.
pragmatic approaches are mostly concerned with
Giaglis and Doukidis (1997) emphasise business
capturing and understanding processes, while
process models use for change management which
rigorous paradigms are typically used for analysis
may be considered in more general terms as the
of the process. Furthermore, it is argued that
need to learn, analyse, monitor and control the
‘ythis suggests a need for different notational
process and thus needing descriptive and decision
approaches, for different modelling purposes and
support models. The most popular of these
audiences’ (Phalp, 1998). For software develop-
approaches include: BPR (Hammer, 1990), Con-
ment process typically diagrammatic notation is
tinuous Process Improvement (CPI) (Harrington,
required, for capturing a legible and understand-
1991), Total Quality Management (TQM) (Oak-
able view of the business process. Typically the
land, 1993), and Organisational Transformation
user does not need to play or interact with the
(OT) (Adams, 1984). Other authors such as
model but rather just observe it. When analysing
Workman et al. (2000) claim too that many
the business process it is necessary to have more
different models may be needed to analyse
sophisticated mechanisms than qualitative analysis
business processes depending on the purpose.
of static diagrammatic models, models that present
Some of these references aim at defining aspects
both dynamic and functional aspects of the
for a business process model to be complete. e.g.
process. In these cases users might want a model,
Workman et al. (2000) present the historical
which permits him/her more interaction (e.g. it
development of enterprise organisation and in-
might be simulation) to analyse the question ‘what
formation technology distinguishing six phases:
if’. Finally, in presenting the business process,
approaches easy to understand are chosen, again a 1. the functional hierarchy,
readily understandable, typically diagrammatic 2. the functional hierarchy with function oriented
notation is suitable. automation,
Macintosh (1993) defines five levels of process 3. the functional hierarchy with shared database
maturity: on mainframes,
 ARTICLE IN PRESS

132 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

4. the process oriented enterprise, (1995) identify information requirements for

5. the supply chain oriented enterprise, and business re-engineering to build tools that can
6. the web-enabled agile enterprise. support business knowledge acquisition, busi-
ness process modelling, performance/quality ana-
For each historical development phase they lysis and analysis of alternative BPR solutions.
explain the main descriptive aspects arguing that Rajala et al. (1996) introduce a new requirement,
for each phase different models have been needed. the customer orientation, and propose a new
For phase 4 they emphasise the need of modelling modelling framework, which integrates cus-
business processes and they define what they called tomer orientation to business process simulation
business model architecture. Giaglis and Doukidis modelling.
(1997) examine the nature of business processes Hommes et al. (2000) give a more general
in the light of modern change management framework to define a business process technique.
approaches and propose a set of requirements They identify on the one hand four elements that
for their modelling as follows: constitute any individual model: notation, mean-
* Technical requirements: formal modelling, ing, concept relationship and modelling concept,
quantitative modelling, stochastic modelling, which are called the way of modelling (modelling
model documentation, model adaptability/reu- concepts). On the other hand there are three
sability and objective-driven modelling. elements that constitute the way of working:
* Political/social requirements: Feasibility of al- procedure relationship, activity relationship and
ternative designs, communication of models activity, which describe the procedures by which
and user friendliness. the models are constructed (modelling procedure).
Hommes et al. framework is focused on describing
These requirements are based on the thought the modelling technique. Whether the resultant
that businesses are essentially ‘‘socio-technical’’ model is adequate or not is another question.
systems. Giaglis and Doukidis (1997) show too Phalp (1998) uses a similar idea, which underlines
that simulation can be an invaluable tool for that notation and method are two important
Business Process Modelling (BPM) especially considerations when modelling business processes.
when modelling inter-organisational business pro- Both method and notation will depend on the
cesses. They add three additional requirements for desired model characteristics, which in turn will
inter-organisational business modelling: modular depend on the purpose.
model design, modular model analysis and model Hence, business processes can be described at
decomposition and integration. The idea is to different levels of detail depending on the abstrac-
show features that a business process model tion put into analysing the organisation, which
should have in order to be classified as performing depends in turn on the purpose of the analysis. As
successfully. All requirements identified by Giaglis a result of the literature review, it was identified
et al. are basically meant as guidelines for that business process models are mainly used
prospective users or developers of business process either to learn about the process, to make decisions
simulation models. Hence, we can state that to on the process or to develop business process
define model requirements that enable it to be software. Usually, these purposes relate to some
considered complete is function of the purpose of extent to some model characteristics. That is to
the model. In this sense, Phalp (1998) proposes say, some business process models are better suited
that models used to analyse business processes depending on the specific purpose. We suggest a
for developing software should include expert very basic classification of business process models
judgements and heuristics, measurements, form- giving attention to their level to allow and
ality and be executable. In the same context facilitate model changes in passive and active.
Toussaint et al. (1997) find three essential aspects Passive means not to allow changes whereas active
of processes that models should present, namely: means to allow changes. We combine these two
functional, static and dynamic. Jarzabek et al. dimensions to classify business process models.
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 133

sured sets of activities designed to produce a

specified output for a particular customer or
Model change permissiveness

market’’. There are so many other definitions but

in essence all are the same: processes are relation-
ships between inputs and outputs, where inputs are
transformed into outputs using a series of activ-
ities, which add value to the inputs.
It is beyond the scope of this paper to go into
further discussion on the difference between
business processes and processes in general. It
seems that some authors take them as synon-
ymous. For example, in contrast to Davenport,
Purpose of the model
Hammer (1990) defines businesses process as ‘‘a
Fig. 1. Classification framework for business process modelling collection of activities that takes one or more kinds
techniques. of input and creates an output that is of value to
the customer’’. However, an important distinction
This is represented in a matrix with two axes one for the author of this paper is that a business
for each dimension (see Fig. 1). process is related to enterprises, as they define the
Concerning the assessment of the quality of a way in which the goals of the enterprise are
business process modelling technique Hommes et al. achieved and thus they are a subset of the set of
(2000) present a set of quality properties and processes. According to ENV 12204 (1995) a
procedures to make an objective assessment of business process is a partially ordered set of
these properties possible called Quality-based Mod- Enterprise Activities which can be executed to
elling Evaluation (Q-ME) framework. They link the realise a given objective of an enterprise or a part
properties of the model to the way of modelling and of an enterprise to achieve some desired end-result.
working mentioned above. Based on the FRISCO There are many classifications of business
report (Falkenberg et al., 1996) there are three processes too. Often ‘core’ and ‘supportive’ busi-
quality properties: expressiveness, arbitrariness and ness processes are distinguished. A core (or
suitability. Other authors such as Barros and primary) process is initiated from outside an
Hofstede (1998) and Harmsen (1997) propose other organisation, e.g. the chain of activities that
properties such as: comprehensibility, coherence, realises the delivery of a product to a customer.
completeness, efficiency and effectiveness. A supportive (or secondary) process creates the
In Pandya (1995) a comparison of some conditions for the primary process to be carried
methodologies and some modelling tools for out. This last might be classified in its turn as
business processes is undertaken and a guide management processes that control the organisa-
showing their suitability in manufacturing man- tion’s overall strategies and objectives; and sup-
agement and business processes is proposed. port processes, which support the core processes
Nevertheless, there is neither a classification of by offering sufficient resources.
the methodologies and/or tools nor clear distinc- There are many business process modelling
tion between them. techniques, as were mentioned above (PROCESS,
2003). As a result of the literature review the
following were found as the most frequently used
3. Brief description of the main process modelling and therefore they are considered as the main
techniques techniques. Sections 3 and 4 are a summary of a
longer working paper. For further information,
Before describing any technique we define what details and references of those sections see Aguilar-
a business process is. According to Davenport Save! n (2001). The most important characteristics
(1993) processes are defined as ‘‘structured, mea- of each technique are discussed below.
 ARTICLE IN PRESS

134 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

3.1. Flow chart technique ability. The Flow Chart model is very easy to use.
It does not take a very long time to draw a sketch
A Flow Chart is defined as a formalised graphic of a process.
representation of a program logic sequence, work The weakness of the standard is that it is too
or manufacturing process, organisation chart, or flexible. The boundary of the process may not be
similar formalised structure (Lakin et al., 1996). It clear. Flow charts tend to be very big. Already in
is a graphical representation in which symbols are the evaluation model, the flowchart can be too
used to represent such things as operations, data, large. There is also no difference between main
flow direction, and equipment, for the definition, and sub-activities, which makes the chart hard to
analysis, or solution of a problem. The Flow Chart read. Since there are no sub-layers, it is hard to
modelling method uses flowcharts to represent navigate and it is difficult to find information in
processes. It uses a sequential flow of actions and the chart. Of course it is easier to follow the course
does not support a breakdown of the activities. of events, but the risk of getting lost is high.
The Flow Chart model is possibly the first process Visualising the process with a flowchart can
notation. It has frequently been used over many quickly help identify bottlenecks or inefficiencies
years although there is no exact date for its origin. where the process can be streamlined or improved.
Fig. 2 depicts an example of a simple process by The best use of Flow Chart technique is when it
using a flow chart. The process starts when a is used to deal with processes that need a high level
customer makes an order to a company. The of detail. On the contrary, it is not very good for
company’s Marketing Department receives the giving an overview. Usually, there is no natural
order, introduces the information in the informa- way of describing responsibilities or performers in
tion system at the company and sends the order the chart. This makes it hard to connect the
forward to the Distribution Centre. Distribution organisational functions, often referred to as
Centre verifies the availability of the required ‘departments’, to activities.
products and if they are available they ship the
products to the customer together with the invoice, 3.2. Data flow diagrams—Yourdon’s technique
otherwise they inform Marketing of the non-
availability so that Marketing may inform the Data flow diagrams (DFD) are diagrams that
customer. show the flow of data or information from one
The main characteristic of Flow Charts is their place to another. DFDs describe the processes
flexibility. A process can be described in a wide showing how these processes link together through
variety of ways. The standard just gives the data stores and how the processes relate to the
notation, but how the different building blocks users and the outside world. They are used to
are put together is up to the designer of the chart. record the processes analyses as a part of the
When we look at a flowchart representation, it is design documentation (http://panoramix.univ-par-
easy to recognise the processes it describes. The is1.fr/CRINFO/dmrg/MME/misop025/info.html)
real strength of the standard is the communication and (http://threesl.com/data flow diagrams.htm).
A DFD can be seen as a method of organising
data from its raw state. DFDs are the backbone of
Receive Order
Distribution yes
Print Invoice
structured analysis that was developed in the early
Centre Stock?
sixties by Yourdon. Fig. 3 depicts the same process
as in Fig. 2 although using DFD notation.
no By using DFD, the analyst will be able to specify
Shipping a process at the logical level. This means that he
will be able to describe what a process will do,
Advise Inform
Marketing customers rather than how it will be done. They are used in
discussions between analysts and users as they can
Fig. 2. Example of flow chart. be easily understood and verified, and are easy to
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 135

1 Marketing Order from

Customer Receive Customer order customer
order
Receive order [vendor] Marketing

Send order [vendor] Marketing

No available
2 Distrib.Cent Customer
order
Verify
availability
Verify availability [controller] Distribution Centre
or
Yes available
Not Print Invoice [distributor] Shipment
available
4 Shipment 3 Shipment
Shipping Print Contact customer [vendor] Marketing Shipping products [distributor] Shipment
products invoice
Fig. 4. Example of an ‘action diagram’.

Fig. 3. Example of a data flow diagram.

are often organisational functions. They also
draw and amend. Each process can be broken include software systems, customers and suppliers.
down into sub-processes at a lower level to show RADs provide a different perspective of the
more detail. They only show the flow of data, not process and are particularly useful in supporting
materials. DFD are used in the functional model communication. They are easy and intuitive to
to specify the meaning of operations and con- read and understand presenting a detailed view of
straints and show functional dependencies. It the process and permitting activities in parallel.
shows how information enters and leaves the With careful modelling, RADs might define the
process; what activities change the information; degrees of empowerment within the business and
where information is stored within the process, can also demonstrate how processes interact. It
and the organisational function to which the can even be used to describe how software systems
activity belongs. interact. RADs are, in fact, object state transition
‘Action Diagrams’ are a special case of DFD diagrams used in object-oriented models. They
.
with simpler notation (Goldkuhl and Rostlinger, describe how a role object changes state as a result
1988) and permit a contextual analysis. In this of the actions and interactions, which occur.
sense, ‘action diagrams’ represent an exception of Disadvantages are that the technique explicitly
DFD because they introduce data concerning the excludes business objects, which are manipulated
performer, may show both information and by the process, as machines or products. The
material flows, and distinguish between knowledge process is presented as a sequence of activities not
and information. Fig. 4 shows the action diagram letting a decomposition of the process, and thus it
of the same process of Fig. 2. makes an overview difficult.

3.4. Role interaction diagrams—RID

3.3. Role activity diagrams—RAD
Role interaction diagrams (RIDs) are a graph of
Role activity diagrams (RADs) are based a process resulting from the combination of RADs
around a graphic view of the process from the and Jacobson’s object interaction diagrams
perspective of individual roles, concentrating on (Boma, 1996). Activities are connected to roles in
the responsibility of roles and the interactions a type of matrix. Activities are shown vertically on
between them (Holt et al., 1983). Roles are the left axis and the roles are shown horizontally at
abstract notations of behaviour describing a the top. Text and symbols are used together in
desired behaviour within the organisation. They order to represent the process. Horizontal lines
 ARTICLE IN PRESS

136 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

CUSTOMER MARKETING DISTRIBUTION

ORDER
CUSTOMER ORDER
ORDER
RECEIVE ORDER

AVAILABILITY CHECK
REJECTED ORDER
COMMUNICATE NO AVAILABILITY

RESERVATION IN STOCK
PRODUCT PROCESSED
INVOICING

Fig. 5. Example of a RID.

show human interactions (Boma, 1996). Fig. 5 perform the activity, and the name of the person
shows the same process as in Fig. 2 using RID. assigned to the activity, followed by one column
Although slightly more complex than flow for each period in the project’s duration. Each
diagrams, RIDs are fairly intuitive to understand, period may be expressed in hours, days, weeks,
easy to read but they tend to be messy, with many months, and other time units. In some cases it may
arrows pointing left and right and are therefore be necessary to label the period columns as period
quite hard to build. Inputs to, and outputs from 1, period 2, and so on. Gantt charts relate a list of
the activities are not modelled. Therefore, impor- activities to a time scale, thus they might be used to
tant information is lost. When editing an existing represent a process graphically and control its
diagram, it can be hard to insert new activities or current situation of performance, although its use
roles. When a new activity or a new role is to be to analyse a process is limited. They are very
inserted, big parts of the diagram have to be simple graphic representations but they do not
moved to allow space. Since each activity is bound show clear dependencies between activities.
to a performer, the responsibilities are well defined
and thus the connection to the organisation is easy 3.6. IDEF
to make.
RIDs are not as flexible as flowcharts, for The Integrated Definition for Function Model-
example. They have quite rigid notation. But ling (IDEF) is a family of methods that supports a
compared with other modelling techniques, RIDs paradigm capable of addressing the modelling
are nevertheless flexible. Due to their notation and needs of an enterprise and its business areas
ability to break down activities, very complex (IDEF, 2003). IDEF’s roots began when the US
processes can be displayed. The best use of RIDs is Air Force, in response to the identification of the
in workflow design. RIDs are primarily used for need to improve manufacturing operations, estab-
processes that involve co-ordination of interrelated lished the Integrated Computer-Aided Manufac-
activities. turing (ICAM) program in the mid-1970s. The
requirement to model activities, data, and dynamic
3.5. Gantt Chart (behavioural) elements of the manufacturing
operations resulted in the initial selection of the
A Gantt Chart (Aguilar-Save! n, 2001) is a matrix Structured Analysis and Design Technique
that lists on the vertical axis all the tasks or (SADT). SADT is more than a technique. It is a
activities to be performed in a process. Each row whole methodology to be used as a regimented
contains a single activity identification, which approach to analysing an enterprise.
usually consists of a number and a name. The The IDEF family is used according to different
horizontal axis is headed by columns indicating applications. The most important parts are:
estimated activity duration, skill level needed to IDEF0, IDEF1, IDEF1X, IDEF2, IDEF3, IDEF4
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 137

and IDEF5. However, for business process mod- as representing a sequence of activities. The
elling, the most useful versions are IDEF0 and activities may be placed in a left to right sequence
IDEF3 and therefore they are explained further within decomposition and connected with the
below. flows. It is natural to order the activities left to
IDEF0 is a modelling technique used for right because, if one activity’s output is used as
developing structural graphical representations of input by another activity, drawing the activity
processes or complex systems as enterprises. It is boxes and concept connections is clearer. Thus,
used to specify function models, which are ‘‘what without intent, activity sequencing can be em-
do I do?’’ models. These show the high-level bedded in the IDEF0 model.
activities of a process indicating major activities IDEF1 is used for information modelling, which
and the input, control, output, and mechanisms captures conceptual views of the enterprise’s
associated with each major activity. The processes information. IDEF1X is used for data modelling,
can be further decomposed to show lower-level which captures the logical view of the enterprise’s
activities, but at some point the required view may data and is based on an entity relationship model.
require another notation to portray such things as It is a design method for logical database. IDEF2
branch control. These models are composed of Simulation Model Design method is used to
three types of information: graphical diagrams, represent time varying behaviour of resources in
text and glossary. These three types are cross- a manufacturing system. Various commercial
referenced to each other. The major component is products and notations have replaced it.
the graphical diagram, containing boxes, arrows, IDEF3 Process Description Capture method is
box-arrow interconnections and associated rela- used to capture behavioural aspects of a process. It
tionships. The IDEF0 format is shown in Fig. 6 to allows different views of how things work within
describe the same process as in Fig. 2. an organisation. Unlike IDEF0, IDEF3 has been
The IDEF0 standard is the most popular developed for explicitly describing processes. The
process-modelling on the market. The very strict former shows what is done within the organisation
rules in IDEF0 make it suitable for implementa- while the latter shows how things work with it.
tion as computer software. By working backwards From domain experts, descriptions are captured in
along the chain from output to inputs, much data which the precedence and causality relationships
and control can be defined. Thus it can be analysed between activities and events of the process are
and improved. The hierarchical structure facil- shown. IDEF3 consists of two modelling modes:
itates quick mapping at a high level. One weakness the process flow description (PFD), which descri-
is the tendency of IDEF0 models to be interpreted bes how things actually work in the organisation,

Check credit

Receive order
from customer
Order information

Receive order
from Marketing
Order information

Verify Invoice
availability yes
Finished
goods
no Notify
customers

Fig. 6. IDEF0 graphical diagram’s components.

 ARTICLE IN PRESS

138 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

and the object state transition description (OSTD), program execution) or by means of more formal
which summarises an object’s allowable transitions analysis methods (which are equivalent to pro-
in a particular process. It is suitable to model both gram verification).
simple and complex processes due to its decom- Petri nets were originally developed in the 1960s
position ability. and 1970s, and they were soon recognised as being
The basic notation of the IDEF3 method one of the most adequate and sound languages for
consists of a series of square and oblong boxes, description and analysis of synchronisation, com-
and circles and arcs which link them. Attached to munication and resource sharing between concur-
each icon is an elaboration form, which contains a rent processes. However, attempts to use Petri nets
description of that icon, reference label, etc., and a in practice revealed two serious drawbacks. First
detail of related objects, facts and constrains of all, there were no data concepts and hence the
acting upon it. IDEF3 is used in several areas models often became excessively large, because all
such as Business Process Engineering (BPE) and data manipulation had to be represented directly
Reengineering (BPR), software process definition in the net structure. Secondly, there were no
and improvement, and even in the software hierarchy concepts, and thus it was not possible
development and maintenance. to build a large model via a set of separate sub-
IDEF4 object-oriented design method was models with well-defined interfaces. CP-nets in-
developed to support the object-oriented para- corporate both data structuring and hierarchical
digm. It currently supports design to implement C decomposition without compromising the qualities
language applications. IDEF5 method provides a of the original Petri nets and thus removed these
theoretically and empirically well-grounded meth- two serious problems.
od specifically designed to assist in creating, Fig. 7 shows a description of an object’s
modifying and maintaining ontology. Ontology is behaviour states as an example of a CPN. The
a part of philosophy whose goal is to divide the object considered in the figure is ‘an acknowl-
‘‘world’’ into different objects. edgement’, which is affected by different activities
performed in a process: to send, to receive or to
3.7. Coloured Petri-net—CPN update. According to these activities and a series
of rules the object will change: waiting, passive,
Coloured Petri nets is a graphical oriented inactive, etc.
language for design, specification, simulation and
verification of systems. It is particularly well suited 3.8. Object oriented methods
for systems that consist of a number of processes,
which communicate and synchronise (http://www. The term object orientation (OO) has different
daimi.au.uk/PetriNets/tools/quick.html). Colour- meanings. The history of object oriented program-
ed nets are extended Petri nets in which symbols ming starts with the development of the discrete
are differentiated by ‘‘COLOURS’’. A CPN model event simulation language Simula by Dahl and
consists of a set of modules which each contain a Nygaard in Norway in 1967. Generally, it is used
network of places, transitions and arcs. The to describe a system that deals primarily with
graphical representation makes it easy to see the different types of objects, and where the actions
basic structure of a complex CPN model, i.e. to one can take depend on what type of object one is
understand how the individual processes interact manipulating. Thus, OO methods might be defined
with each other. CP-nets have a formal, mathe- as methods to model and programme a process
matical representation with a well-defined syntax described as objects, which are transformed by the
and semantics. This representation is the founda- activities along the process. The fundamental
tion for the definition of the different behavioural construct is the ‘‘object’’, which combines both
properties and the analysis methods. The beha- data structure (attributes) and behaviour (opera-
viour of a CPN model can be analysed, either by tions) in a single entity. Objects may represent real
means of simulation (which is equivalent to world applications (Rumbaugh et al., 1991).
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 139

Fig. 7. Example of a CPN.

Benefits of using OO methods are described in (3) increase the internal consistency across ana-
(Bruce, 1998). lysis, design and programming,
This method is based on three concepts: objects (4) explicitly represent commonality between
that represent a real-world entity. An object has a classes and objects,
state, i.e. one of the possible conditions in which (5) build specifications resilient to change,
the object may exist represented by the values of (6) reuse OO Analysis and OO Design results,
the properties (attributes). State changes are and
reflected by the behaviour, i.e. how an object acts (7) provide a consistent underlying representation
and reacts determined by the set of operations the for analysis, design and programming.
object can perform on itself, and also knowing its
interface, functions and methods. A set of similar One of the main advantages of OO method is
objects is called class. For example, the attributes the effectiveness of the process to identify and
for the class animal are having four legs and a tail. refine objects.
Its behaviours is sleeping and eating. Then OO is one of the main methods used for process
possible instances or objects of the class animal modelling, especially when the model needs to be
are cat, elephant, and horse. Finally, messages are enactable. There are many different techniques
requests for the receiver objects to carry out the based on OO. The main techniques used are:
indicated method or behaviour and return the
result of that action to the sender objects. States * Booch’s Object Oriented Design (OOD) Techni-
change through behaviour when the object que (Booch et al., 1999) and http://www.slac.-
receives a message. standford.edu/~marino/html/booch/method.html,
Coad and Yourdon (1991) mention seven key * Coad and Yourdon’s OOA/OOD technique
motivations and benefits in favour of OO methods (Coad and Yourdon, 1991; Coad et al., 1995),
instead of using other analysis methods. These * Rumbaught object modelling technique (OMT)
are (Rumbaugh et al., 1991),
* Shlaer-Mellor Technique, also known as the
(1) tackle more challenging problem domains, object-oriented systems analysis (OOSA) (http:
(2) improve analyst and problem domain expert //cheetah.sdd.sri.com/eliot/ads/shlaer-mellor.
interaction, html).
 ARTICLE IN PRESS

140 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

Differences among those techniques are basi- Use-case diagram, illustrates the relationships
cally related to their notation. Therefore, we will between use-cases. Each use-case, typically defined
explain only one called Unified Modelling Lan- in plain text, describes a part of the total system
guage (UML), which is considered the standard functionality. Component diagram, a special case
OO modelling language. Coad and Yourdon’s of class diagram used to describe components
method precedes UML. For further details within a software system. Deployment diagram,
on these techniques and OO methodology see a special case of class diagram used to describe
Aguilar-Save! n (2001). hardware within a software system.
Unified Modelling Language: UML is a language
for specifying, visualizing, constructing and doc- 3.9. Workflow technique
umenting the artefacts of software systems, as well
as for business modelling and other non-software In general terms, it is defined as the compu-
systems. UML uses OO methods for modelling. terised facilitation or automation of a business
The UML represents a collection of engineering process, in whole or in part, during which
practices that have proven successful in the documents, information or tasks are passed from
modelling of large and complex systems, see one participant to another for action, according to
UML (2003) and Booch et al. (1999) for further a set of procedural rules (Fischer, 1995). Workflow
information. is a flow of tasks between computer applications or
The UML covers conceptual things, such as people in an organisation. Two or more members
business processes and system functions, as well of a workgroup to reach a common goal can define
as concrete things, such as programming-language workflow as well as any task performed in series or
classes, database schemas, and reusable software in parallel. Workflow is more than a technique to
components. The Unified Modelling Language model a process. It is a method to analyse and
serves as a basis for representing most methods improve a process, including its modelling. A work
using a common set of modelling constructs and a management system is a system that defines, creates
common notation. It captures the concepts from and manages the execution of workflow through
the OMT, Booch, and OOSE methods, but they the use of software whose order of execution is
hope that other methodologists will adopt it also, driven by a computer representation of the work
so that users can understand models from any logic (Hollingsworth, 1995).
method without confusion. The UML can be The workflow development process uses work-
considered as the standard of the entire object flow models to capture the relevant information of
oriented community.1 the processes. This process comprises four stages:
The UML consists of nine different diagrams, Information Gathering, Business Process Model-
and each diagram shows a specific static or ling, Workflow Modelling, and Implementation,
dynamic aspect of a system: Class diagram, Verification and Execution. Fig. 8 shows the basic
describes the structure of a system. The structures concepts and terms used in workflow and their
are built from classes and relationships. Object relationships. Some advantages are: work not
diagram, expresses possible object combinations of forgotten, shorter learning time, data transfer,
a specific class diagram. Statechart diagram, process improvement, easier to make changes,
express possible states of a class (or a system). decentralisation, workflow can be used in combi-
Activity diagram, describes activities and actions nation with other systems. Disadvantages: lost
taking place in a system. Sequence diagram, shows human contact, lack of motivation, feeling
one or several sequences of messages sent among a controlled.
set of objects. Collaboration diagram, describes a There is no particular notation for workflow
complete collaboration among a set of objects. systems. This is due to the existence of a number of
workflow languages, which aim to describe and to
1
A comparison of Object Notations available at: http:// specify workflow. Each one of these languages uses
www.cs.queensu.ca/home/stlab/local/UML-paper.html. a specific notation, sometimes a graphical one, to
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 141

Business process
is defined in a is managed by

Process Workflow
Definition Management
Sub-processes used to create System
composed of and manage
controls
Activities Process Instances
which may be
Include one or more
or
during execution Activity Instances
Manual Automated are represented by
which
activity Activity
and/or
Invoked
Work Items
Applications

Fig. 8. Concepts of workflow.

A4 A5 not familiar with these concepts. For example,

A10 Open System Architecture for CIM (CIMOSA)
A3 and Integrated Enterprise Modelling (IEM) are
A8 A9 A11 A13 two methodologies dealing with enterprise integra-
A1 A2
tion. As a part of this task they model enterprise
A6 A12
business processes. CIMOSA does this by using its
A7
own constructs, which are designed to be easily
programmed based on OO methods. IEM model
Fig. 9. Example of workflow. business processes by using diagrams based on
IDEF0 and by using EXPRESS language that,
again, is based on OO methods. In other words,
describe the processes. Fig. 9, for example, shows a the techniques used within these methodologies to
workflow process description using a graphical model business processes are based on what has
notation where arrows represent information flow, already been explained although calling them
and squares are workflow activities. These lan- CIMOSA and IEM. This means that e.g. CIMO-
guages can be classified as: Graph-Based lan- SA and IEM are not different or new business
guages, Net-Based languages (based on Petri nets) process modelling techniques than those included
and Workflow Programming languages. in Section 3. The aim of this section is to clarify
and present some of these cases.

4. Generic methodologies 4.1. SSADM methodology

Sometimes when searching business process Structured systems analysis and design metho-
modelling techniques names come up that represent dology (SSADM) is a methodology used in the
more than a technique. They are generic meth- analysis and design stages of systems development.
odologies with process modelling capabilities. It is not considered as a particular technique for
Unfortunately those names are often used to point process modelling. It is considered as a set of
out either the methodology or the modelling procedural, technical and documentation stan-
technique involved, which may create confusion dards for systems development. SSADM adopts
among practitioners especially to those who are a prescriptive approach to information systems
 ARTICLE IN PRESS

142 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

development which specifies advance modules, being examined and illustrate issues that will be
stages and tasks that have to be carried out, the considered for analysis, reflection and change.
deliverables to be produced and furthermore the They include components such as clients, people
techniques used to produce the deliverables. involved, tasks performed and environment. This
SSADM adopts the Waterfall model of systems technique is very useful in understanding the
development, where each phase has to be com- interaction of different elements involved in the
pleted and signed off before subsequent phases can process and the interaction between processes,
begin according to Downs et al. (1992) and although it is not suitable for a structured analysis,
Nicholas (2003). or to report a description.
SSADM revolves around the use of three
techniques, namely Logical Data Modelling, Data 4.3. GRAI methodology
Flow Modelling and Entity/Event Modelling. The
structure of SSADM consists of five main mod- Graph with results and activities interrelated
ules, which are in turn broken down into a (GRAI) or sometimes called GRAI integrated
complex hierarchy of stages, steps, and tasks: (1) methodology (GIM) is a methodology developed
Feasibility Study; (2) Requirements Analysis; (3) to address production management decisions in
Requirements Specification; (4) Logical System manufacturing systems. As with SSADM, some
Specification; and (5) Physical Design. For the authors call GRAI a methodology and others call
analysis of processes SSADM uses Data Flow it a technique for process modelling. Actually,
Diagrams already explained in Section 3. strictly speaking GRAI is more than a technique
and also GIM, which is a whole methodology
4.2. Soft systems methodology (Doumeingts et al., 1996). From the beginning,
GIM meant GRAI-IDEF0-Merise. Merise is a
Soft Systems Methodology (SSM) is a metho- method developed in France for the analysis and
dology used to support and to structure thinking design of information systems. It is used for
about, and intervention in, complex organisational designing data models (static models) with En-
problems. Existing systems, or those yet to be tity/relationships as formalisms, and enactable
designed, are viewed as social systems, which are models (for instance simulation models) based on
derived from human activities. Human activity, Petri nets as formalisms.
behaviour and interaction are all factors that need Today the GRAI methodology uses four views,
to be reconstructed, see Soft Systems Methodol- namely the functional, physical, decisional and
ogy web page: http://members.tripod.com/SSM informational systems, to provide the analyst with
Delphi/ssm4.html. a generic description of the manufacturing system
The SSM process consists of seven stages. In while focusing on the control aspects of this
these stages, one uses relevant techniques and system. These views permit the building of partial
switches between the real world and the concep- models of the enterprise. Processes are seen from
tual modelling world where appropriate. These different viewpoints through the four views. For
seven stages are: (1) Define the problem situation: example, the role of the functional view is to create
unstructured; (2) Express the problem situation; a simplified representation of the entire system
(3) Formulate root definitions; (4) Build concep- showing the main functions (activities) within the
tual models; (5) Comparison of stages 2 and 4; (6) system as well as the interactions between them.
Define feasible and desirable changes; (7) Take GIM uses IDEF0 diagrams to represent the
action to improve the situation. This methodology functional view as well as the physical view. For
helps to understand and analyse a process from the the design of the informational system GIM uses
human perspective. One of the techniques used to MERISE that is based on Entity/relationships
describe a process is called Rich Pictures, which are diagrams and Petri nets. Besides GRAI-GIM
highly contextual representations of things. They (Roboam, 1993) provides for its users: a modelling
represent some of the richness of the situation framework, a set of modelling formalisms, and
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 143

a structured approach. The analysis of decisions model exactly the behaviour of a real system due
within processes at GRAI GIM is based on GRAI to the huge number of variables involved. In
grid and GRAI nets techniques to model processes general, simulation is not used when the system
with focus on the decisional flow. This is a unique can be modelled analytically because it requires
feature of GIM. Thus, GRAI grid and GRAI nets quite a big investment in time and resources.
permit building process models focusing on
responsibilities and decision-making processes
being more adequate to be used as a technique 5. Proposal framework: Classification of process
for the analysis and design of the process. GRAI modelling techniques
grid and GRAI nets techniques were based on
IDEF0 when invented. Practitioners and academics require simple and
clear guidelines in order to facilitate the task of
4.4. Simulation choosing the most appropriate technique. This
section proposes a classification of the techniques
According to Kelton et al. (1996) simulation is a according to their purposes and change model
collection of methods and applications to imitate permissiveness. Earlier sections aid in understand-
the behaviour of real systems. Simulation can be ing the positioning of each technique in this
classified, according to certain characteristics, in framework. As a result of the analysis carried
deterministic (input data is fixed) or stochastic out in these Sections 3 and 4, Table 1 was built to
(input data is randomised), static (system descrip- present a summary of the above techniques.
tions in a mathematical way where time has no In order to make it easier and faster to view a
role) or dynamic (time plays an essential role) and relationship between each technique and some of
continuous (systems change their state continu- its associated tools available in the marketplace,
ously) or discrete (events that occur at separated the table in the Appendix was developed. It is a
points of time). Processes seen as systems might be summary list of the tools in alphabetic order. The
modelled using simulation for the purpose either table is based on information from Aguilar-Save! n
of understanding the behaviour of the process or (2001) which in turn is based on the information
of evaluating various strategies for the operation presented by Kettinger et al. (1997a) and com-
of it either for decision-making or for learning pleted with information from vendors Web-based
purposes. However, simulation is based on other marketing material.
techniques such as Petri nets or OO methods when The techniques presented in Sections 3 and 4 can
modelling systems, so the technique used will be classified according to the two dimensions
depend on the simulating device selected i.e. the described in Section 2, namely purpose of the
tool. Today, there are many available simulation model and model change permissiveness. Fig. 10
tools that can model all kind of systems, no matter shows the resultant framework. The framework
its complexity. Nevertheless, from the user point of classification together with Table 1 is proposed to
view the models can be created using the simulat- be used to choose among the business process
ing device facilities without minding which tech- modelling techniques. The idea is to provide users
nique is behind it. Likewise, the simulating device with a framework that helps them to decide which,
facilities give the user the possibility to simulate a among the explained techniques, is the one they
process model made with one of the techniques should apply for a specific case.
presented in Section 3. This double interaction Process modelling techniques might be used
between simulation and the modelling techniques either to develop software that supports processes
makes difficult the classification of simulation. or to analyse the processes themselves. In both
However, the author of this paper considers it cases sometimes a model is required to describe the
worth mentioning anyway since simulation repre- process either as a data capture or a presentation
sents a great possibility to model a business exercise. This aids learning about the process.
process. As a disadvantage simulation cannot Sometimes models are needed to make decisions on
 144
Table 1
Business process modelling techniques

Technique Description Attributes Characteristics Strengths and Weakness

User perspective Modeller perspective

Strength Weakness Strength Weakness

Flow Graphic Flow of actions Not sub-layers Communication Can be too large Flexibility quick, No method
Chart representation Great details ability simple available

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

No overview Different
notations
DFD Descriptive Flow of data Explains Easy to Only flow of Easy to verify
diagrams for logical level understand data is shown and draw
structured analysis sub-layers
RAD Graphic view Flow of individual Detailed view Supports Not possible to Include business Different

ARTICLE IN PRESS
object state roles Degree of communication be decomposed objects notations
transition empowerment Intuitive to read
diagrams No overview
RID Matrix Flows of activities Inputs to and Intuitive to Important Rigid notation Difficult to edit an
representation of and roles outputs from understand information is Complex processes existing diagram
processes for are not not included can be displayed Hard to construct
co-ordination of modelled
activities Performers are
included
Gantt Matrix Flow of activities Relate Easy overview Not aid for Simple No clear
Chart representation and duration activities representation analysis or representation of
to time and control of design dependencies
performance
IDEF0 Structural Flows of activities, Based on Shows inputs, Trend to be Strict rules
graphical inputs, outputs, SADT outputs, control interpreted only Possible to build
representation, control and Sub-layers and mechanisms as a sequence of a software
text and glossary mechanisms The most overview and activities Quick mapping
popular details Roles are not
represented
IDEF3 Behavioural Precedence and Allows Easy to Many partial Strict rules and Need lot of data
aspects of a system causality different views understand diagrams to notation Possible to Time consuming
relationships Process flow dynamic aspects describe a build a software when modelling
between activities descriptions in a static way process complex systems
and object state
transition
description
diagrams
Sub-layers
Coloured Graphical oriented Network of places, Extended Petri nets Easy to Models are Formal Time consuming
Petri Nets language to transitions Token are understand how excessively large mathematical when modelling
design, specify, and arcs differentiated individual representation
simulate and by colours processes Well defined
verify systems Hierarchical interact with syntax and
decomposition each other semantics
Possible to build
a software
Data concepts
Object Describe a system Object’s structure Three concepts: Enactable model Model are Internal Need lot of data
Oriented with different type and behaviour objects, classes to control and excessively large consistency across Time consuming

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

Methods of objects and messages monitor and detailed design, analysis when modelling
There are many processes Fragmented and programming Complexity
modelling information Possible to build
techniques a software
based on OO
Workflow Computerised Flow of Flow of tasks Easy to analyse Possible build Lack of a

ARTICLE IN PRESS
facilitation or information, tasks between Shorter learning a software particular
automation of a and procedural computers and time Data transfer notation
business process rules people Easy to make Many languages
Decentralised changes
Rich Contextual Represent process Represent Support It is not Easy to illustrate Lack of a
Pictures representation human some of the communication structured components as particular notation
of things problematic richness of the and approach clients, people,
process being understanding of tasks and
examined the process environment
GRAI Descriptive Decision making Sub-layers Shows inputs, Many partial Strict rules Need lot of data
grid and diagrams of the process and flow Distinction outputs, control diagrams to and notation Time consuming
GRAI process focused of activities between period and and mechanisms, describe a Possible when modelling
nets on decisions event driven activities overview and process to build a software Complexity
details

145
 ARTICLE IN PRESS

146 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

GRAI-GIM
general and simple as possible: change model
Workflow permissiveness. This characteristic pays attention
UML
Model change permissiveness

Shlaer-Mellor OOT to the level to allow and facilitate model changes.

Active

Rumbaught OMT Analysis of the techniques identified has indicated

Coad andYourdon’s OOA/OOD
that some of the models developed may be
Booch’s OOD
IDEF 3 Coloured Petri nets classified as passive. That is, they do not have the
Gantt Chart SSM-Rich Pictures capability to allow the user to interact with, or
IDEF 0 IDEF 3 change them without totally remodelling the
Passive

Role Interaction Diagram

process. In contrast, other models allow users to
DFD - Yourdon
Role Activity Diagram
make changes, or are dynamic themselves. Exam-
Flow Chart Gantt Chart ples include simulation and enactable models.
Descriptive DS for process DS for process IT Enactment Such models may be classified as active. Hence,
for learning develop/design execution support the distinction between active and passive models
Purpose of the model is what we call change model permissiveness, and
Fig. 10. Classification framework to select among business constitutes the vertical axis of our framework (see
process modelling techniques. Fig. 10).

the design, or on the development (changes, 6. Conclusion and further research

improvements or re-design) of processes. The aim
in this case is to develop adequate business Business process modelling is a much-researched
processes, so the purpose of these models is for field but is neither well structured nor classified.
analysis. Sometimes however when executing a There exists considerable confusion on terminol-
process some decisions might be required to ensure ogy. For example, OO has different interpretations
its correct performance. Hence the need for some and definitions. Some consider OO as a general
process models to control and monitor processes as methodology for process modelling. Others how-
well as to give right information in order to ever may consider it a philosophy that shows how
support those decisions. Interactive models are the ‘‘real world’’ behaves or may consider it to be
often of great use here. Finally, for the software just a simple technique. The same applies to
development process, which supports business SSADM, Workflow and GRAI. Hence there is a
processes, enactable models are essential for need to clarify, classify, organise and structure this
programming. Therefore, uses or purposes of field of research. The review and resultant frame-
business process models might be divided into four work presented in this paper is an attempt to fill
main categories as follows: this gap.
A classification framework to aid selection of
(1) descriptive models for learning;
process modelling techniques based on the purpose
(2) descriptive and analytical models for decision
and type of model has been proposed. However,
support to process development and design;
further research is required in order to classify the
(3) enactable or analytical models for decision
techniques according to other criteria such as
support during process execution, and con-
experience, difficulties in use and suitability.
trol; and
Comparisons among the techniques would be of
(4) enactment support models to Information
great help too.
Technology.
There is still lack of a general framework on
They will constitute the horizontal axis of our what a business process modelling technique
framework (see Fig. 10). must include to be successful. Some efforts
Another specific model characteristic is consid- carried out in this sense have given partial results
ered important for the present proposal frame- so far. However, when integrating an enter-
work always looking to make the framework as prise, business process modelling techniques and
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 147

tools cannot in themselves provide ‘the solution’. provide a comparative assessment and aid in their
They are an aid to business analysts to design selection.
and manage the processes, whose understanding
is an essential function of communication and
consensus in an enterprise. Hence, the capa- Appendix
bility of a tool to support communication
and enhance understanding is of the highest Business process modelling techniques and their
importance. associated tools is given in Table 2. The purpose of
Further research is required to analyse in this table is limited to inform readers about
detail the available process modelling tools in available tools. It is not intended to be a guide
order to give users a complete description of the for users to select the appropriate tool. In this
purpose, scope and use of each tool. This will also sense further research is needed.

Table 2

Technique Tools/Trademarks

Flowchart ABC Flow Charter 4.0, ABC Graphics Suite, ABT Project Workbench, AWD and Workflow Analyzer, Bench
Marker Plus, BPM, Business Object Modelling Workbench, Cap Web-Flow, CLEAR, COI-Business Flow,
CORE, COSA, CSEWorkflow 5.0, Docu Flow, EPM SuiteFlow Maker, Flow Path, Flow PATH IMAGEWorks,
Flowcharter, Flowmark, Form Flow, Free Flow, GOOFEE Diagrammer, IBMBusiness Process Modeler, Ithink
(HPS), Jet Form Server, MAXIM, Net Prophet, OCTOFlow, Optix Workflow, PAVONE Group Flow,
PFTamptrade, Power Flow, Power Flow Team Flow Process Wise, Pro Model, Process Charter, Process Maker,
RKB Work Frame, SA/BPR Professional, Smart Flow 98, Vectus, Visual Thought, Work Flow Analyzer, Work
FLOW SQL, Work Flow.2000, Work Flow.2020, Work Xpert, Workflow FONT , CESymbolmiddot
FONTBPR, Workflow Modeler, Workflow.BPR, Trampolin.
DFD— ARIS-Tools, CASE Tool, 4Keeps, BONAPART, GRADE, INCOME, IEW, Paradigm Plus, Popkins Systems
Yourdon Architect, Softwarethrough Pictures SE , ProcessWise, With Class 98, Graphics Toll
Role activity RADitor (Co-ordination Systems Ltd.)
diagrams RAD
Role interaction RADitor (Co-ordination Systems Ltd.)
diagram RID
Gantt chart ABT Proyect Workbench, PFTamptrade, Project Scheduler7, Team Flow, Workflow BPR
IDEF 4Keeps, AI0WIN, BPWin, Business Object Modelling, orkbench, CORE, Design IDEF, Design Leverage, IDEF
Tools, Popkins Systems Architect, Pro CAP Pro SIM, Process Maker, SA/BPR Professional and Workflow
Modeler.
Petri-net–CPN Desigh CPN, UNCOME, PACE, Process Maker and Process Weaver
Booch OOD 4Keeps, lass Designer, Paradigm Plus,
Softwarethrough Pictures Booch, With Class 98

Coad/Yourdon 4Keeps, Paradigm Plus, Together C , With Class 98

OOA/OOD
Rumbaught 4Keeps, Paradigm Plus, Select Enterprise
OMT
Shlaer-Mellor 4Keeps, Bridge Point Automation Tools, Paradigm Plus, SES/objectbench, With Class 98
OOM
UML OOM 4Keeps, Class Designer, COOLJex, Innovator, j-vision, Javision, LOREx2 for Java, Magic Draw UML, Object
Plant, Objecteering, Paradigm Plus, Pragmatica, Real-time Studio, Rhapsody, SDT, Soft Modeler Business,
Softwarethrough Pictures UML, Together C, Together J, Visual UML, With Class 98
Workflow View Workflow, ABSI-Docss, Action Request System, Action Workflow Analyst, Action Workflow Application
Builder, Action Workflow Enterprise Series, Action Workflow Workflow Manager, ARIS, TIWorkflow
Distributor AWD, AWD and Workflow Analyzer, Beyond Mail, BONAPART, Business Object Modelling
Workbench, Cap Web-Flow, CMSWorkflow, COI-Business Flow, Computron Workflow, COOL, COSA,
 ARTICLE IN PRESS

148 R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149

Table 2 (continued)

Technique Tools/Trademarks

CSEWorkflow 5.0, Designer2000, Docu Flow Document Manager, Documentrix Workmanager, EDI36, EDI38,
EDI400, EDIe Qmail, Engineering Workflow System, Ensemble, Enterprise Analyst, Entire Workflow, EPM
Suite, Extend BPR, Fabasoft Components, File Net Work Flow, Flo Ware, Flow Maker, Flow Man, Flow PATH
IMAGEWorks, FLOWBuilder, Flowmark, Form Flow, FORO, FYI, FYI Workflow, Group Wise, IBIsys,
IBMBusiness Process Modeler, IBS Workflow Manager, Image Fast, Image Master, In Concert, Inter Office, Jet
Form Server, Key Workgroup, Keyflow, Lifeflow, Life FLOW, Link Works Team Links, Linkworks, Livelink
Intranet, MAVIM 3, Memo, Message Driven processor MDp, Metaphase 2.0, Metaview FOLDERS, METEOR,
Metis, Navigator 2000Document Management Systems Navigato, Navigator 2000Workflowm, Nova Manage,
OCTOFlow, ODMS, Office.IQ, Open Image, OPENworkflow, Optix Workflow, PANOVE Group Flow, Plexus
Flo Ware, Power Flow, Power Flow Team Flow Process Wise, Power Work, Process IT, Protos, Radica, Regata,
Route Builder Omni Desk, SAP Business Workflow, Smart Flow 98, Smart Stream, SPARKS G2, Struct Ware,
The Vantive System, Ultimus, Viewstar Workbench, Win Work, Wizdom Works, Work Fast, Work Flow
Analyzer, Work MAN, Work Party, Work Xpert, Workflow FONT FACESymbolmiddot, FONTBPR,
Workflow.BPR, WORKlogik TM, World Wide Web Flow W4, Xworkflow
SSADM 4Keeps, SSADM
Soft System Group Decision Support System (GDSS), Group system (Ventura Corp)
Methodology
GRAI GIM IMAGIM, CAGIM (Computer Aided GIM), DGRAI
Simulation AWD and Wordflow Analyzer, BONAPART, BPSimulator Template, Business Object Modelling Workbench,
Business Process Analyzer Bwise Toolkit, CABRE-Witness Cinderella SDL, CLEAR, Clear Process, Design
CPN, Design Leverage Dress Rehearsal, EPM Suite, First STEP, Flowcharter GRADE HITSoft BIZ, HOCUS, i-
think Ithink, Live Analyst, METIS, Micro SAINT Object GEODE, Optima, Optima Express, Oracle Process
Manager, PACE, PAVONE Group Flow, Powersim, ProModel, ProModel2.0, Process Charter Prophesy,
PROSIM Process Modeling Software, Quick CRC, RDD-100, SES/Workbench, SIMAN amp ARENA,
SIMPROCESS, Soft Modeler Business, SPARKS G2, Statemate Magnum, Struct Ware, Surveywin Taylor II, TI
BDF, Vectus, Vensim, Witness, Workflow Analyzer, Workflow FONT FACESymbolmiddotFONTBPR,
Workflow BPR

References Booch, G., Rumbaugh, J., Jacobson, I., 1999. The Unified
Modelling Language User Guide. Addison-Wesley, USA.
Adams, J.D., 1984. Transforming Work. Miles River Press, Boma, 1996. Process definition, available at http://www.sesh.
Alexandria, VA, USA. com/procdef.html.
Aguilar-Sav!en, R., 2001. Business process modelling techniques Bruce, P.D., 1998. Real-Time UML, Developing Efficient
and tools. Department of Production Economics. WP291, Objects for Embedded Systems. Addison-Wesley, USA.
.
Linkoping Sweden. Chan, M., 2002. A framework to develop an enter-
Aguilar-Sav!en, R., Olhager, J., 2002. Integration of product, prise information portal for contract manufacturing.
process and functional orientations: Principles and a case International Journal of Production Economics 75 (1–2),
study. Preprints of the International Conference on 113–126.
Advanced Production Management System, APMS 2002, Coad, P., Yourdon, E., 1991. Object-Oriented Analysis, 2nd ed.
IFIP, September, The Netherlands. Prentice-Hall, Englewood Cliffs, NJ, USA.
Al-Mubarak, F., 2003. A simulation study of focused cellular Coad, P., North, D., Mayfield, M., 1995. Object Models:
manufacturing as an alternative batch-processing layout. Strategies, Patterns and Applications. Prentice-Hall, Engle-
International Journal of Production Economics 83 (2), wood Cliffs, NJ, USA.
123–138. Davenport, T.H., 1993. Process Innovation: Reengineering
Artiba, A., 2001. Productive systems: Strategy, control, and Work through Information Technology. Harvard Business
management. International Journal of Production Econom- School Press, Boston, MA, USA.
ics 74 (1–3), 1–4. Doumeingts, G., et al., 1996. State of the art on models,
Barros, A.P., Hofstede, A., 1998. Towards the construction of architectures and methodologies. In: Bernus, P., Nemes, L.,
workflow suitable conceptual modelling techniques. Infor- Williams, T.J. (Eds.), Architectures of Enterprise Integra-
mation Systems Journal 8 (4), 313–337. tion. IFIP/Chapman & Hall, London, pp. 223–255.
 ARTICLE IN PRESS

R.S. Aguilar-Sav!en / Int. J. Production Economics 90 (2004) 129–149 149

Downs, E., Clare, P., Coe, I., 1992. Structured Systems Journal of Management Information Systems 14 (1),
Analysis and Design Method, 2nd ed. Prentice-Hall, 119–154.
Englewood Cliffs, NJ, USA. Kettinger, W.J., Teng, J., Guha, S., 1997b. Appendices business
ENV 12 204, 1995. Advanced Manufacturing Technology— process change: A study of methodologies, techniques and
Systems Architecture—constructs for Enterprise Modelling. tools. Management Information Systems Quartely Archivist
CEN/CENELEC. 14 (1), Appendices 1–8.
Fischer, L.J., 1995. New Tools for new Times: The Workflow Lakin, R., et al., 1996. BPR enabling software for the financial
Paradigm. Future Strategies. ISBN 0-9640233-2-6. services industry. Management services, ISSN: 0307-6768.
Falkenberg, E.D., et al., 1996. A framework of Information Levitt, J., 1960. Marketing myopia. Harvard Business Review
System Concepts. IFIP WG 8.1 Task group FRISCO. (July/August), pp. 45–56.
Leiden University, Leiden. Macintosh, A.L., 1993. The need for enriched knowledge
Giaglis, G., Doukidis, G., 1997. Simulation for intra- and inter- representation for enterprise. Artificial Intelligence in
organizational business process modelling. Informatica Enterprise Modelling, IEE Colloquium on, pp. 3/1–3/3.
(Ljublana) 21 (4), 613–620. Nicholas, D., 2003. Introducing SSADM available in http://
.
Goldkuhl, G., Rostlinger, A., 1988. For. . andringsanalysis. www.bdris.com/SSADM1.html.
.
Linkopings University, Studentlitteratur, Sweden. Mart!ınez, M.T., et al., 2001. Virtual enterprise—organisation,
Guinet, A., 2001. A design and application methodology for evolution and control. International Journal of Production
hierarchical production planning decision support systems Economics 74 (1–3), 225–238.
in an enterprise integration context. International Journal of Oakland, J.S., 1993. Total Quality Management: The Route to
Production Economics 74 (1–3), 5–20. Improving Performance, 2nd ed. Nichols Publishing, New
Hammer, M., 1990. Reengineering work: Don’t automate. Jersey, USA.
Obliterate. Harvard Business Review 68 (4), 104–112. Pandya, K., 1995. Review of modelling techniques and tools for
Hammer, M., Champy, J., 1993. Reengineering the Corpora- decision making in manufacturing management. IEE
tion: A Manifesto for Business Revolution. New York, Proceedings: Science, Measurement and Technology 142
USA. (5), 371–377.
Harmsen, A.F., 1997. Situational Method Engineering, Ph.D. Phalp, K.T., 1998. CAP framework for business process
thesis. Moret, Ernst & Young Management Consultants, modelling. Information and Software Technology 40 (13),
Utrecht, Holland. 731–744.
Harrington, J., 1991. Business Process Improvement: The Phalp, K., et al., 1999. Quantitative analysis of static models
Breakthrough Strategy for Total Quality, Productivity and of processes. Journal of Systems and Software 52 (2),
Competitiveness. McGraw Hill, New York, USA. 105–112.
Hollingsworth, D., 1995. The Workflow Reference Model. PROCESS, 2003. Modelling and mapping the business process,
Workflow Management Coalition, UK. Available at http:// EPSRC project, home page at: http://www.ecs.soton.ac.uk/
www.wfmc.org. Bkp/process.html.
Holt, A., et al., 1983. Coordination systems technology as a Rajala, M., et al., 1996. A framework for customer oriented
programming environment. Electrical Communication 57 business process modelling. Computer Integrated Manufac-
(4), 307–314. turing Systems 9 (3), 127–135.
Hommes, B., et al., 2000. Assessing the quality of business Roboam, M., 1993. La Methode GRAI. Principes, Outils,
process modelling techniques. Proceedings Hawaii Interna- D!emarche et Pratique. Teknea, Toulouse, France.
tional Conference on Systems SCI, IEEE. Los Alamitos, Rumbaugh, J., Blaha, M., Premerlani, W., Eddy, F., Lorensen,
CA, USA, p. 5. W., 1991. Object-Oriented Modelling and Design. Prentice-
http://panoramix.univ-paris1.fr/CRINFO/dmrg/MME/misop025/ Hall, Englewood Cliffs, NJ, USA.
info.html. Soft Systems Methodology web page: http://members.tripod.
http://threesl.com/data flow diagrams.htm. com/SSM Delphi/ssm4.html.
http://www.daimi.au.uk/PetriNets/tools/quick.html. Swanson, L., 2003. An information-processing model of
http://www.slac.standford.edu/Bmarino/html/booch/method. maintenance management. International Journal of Produc-
html. tion Economics 83 (1), 45–64.
http://cheetah.sdd.sri.com/eliot/ads/shlaer-mellor.html. Toussaint, P.J., Bakker, A., Groenewegen, L., 1997. Construct-
IDEF (2003). Family of Methods web page: http://www. ing an enterprise viewpoint: Evaluation of four business
idef.com. modelling techniques. Computing Methods and Programs
Jarzabek, S., et al., 1995. Model-based support for business re- in Biomedicine 55 (1), 11–30.
engineering. Information and Software Technology 38 (5), UML, 2003. Resource Center web page: http://www.rational.
355–374. com/uml/index.jsp.
Kelton, W., Sadowski, R., Sadowski, D., 1996. Simulation with Workman, J.C., et al., 2000. On the relation between business,
Arena. McGraw Hill, New York, USA. business model, software, and ACT-platform architectures.
Kettinger, W.J., Teng, J., Guha, S., 1997a. Business process Information and Technology Division. Eindhoven Univer-
change: A study of methodologies, techniques and tools. sity of Technology, The Netherlands.