See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/258570130

A Comparison of BPMN 2.0 with Other

Notations for Manufacturing Processes
Article April 2012
DOI: 10.1063/1.4707613

CITATIONS

READS

36

3 authors:
Antonio Garca-Domnguez

Mariano Marcos-Brcena

The University of York

Universidad de Cdiz

51 PUBLICATIONS 111 CITATIONS

249 PUBLICATIONS 1,913 CITATIONS

SEE PROFILE

SEE PROFILE

Ivan Vladimir Medina

Hospital Infantil de Mxico Federico Gmez
4 PUBLICATIONS 7 CITATIONS
SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

Available from: Antonio Garca-Domnguez

Retrieved on: 26 April 2016

A Comparison of BPMN 2.0 with Other Notations for

Manufacturing Processes
Antonio Garca-Domnguez(1), Mariano Marcos-Brcena(2),
Inmaculada Medina-Bulo(3)
(1)

Department of Computer Languages and Systems, University of Cdiz, C/Chile 1, CP 11002,

Cdiz, Spain, [email protected]
(2)
Department of Mechanical Engineering and Industrial Design, University of Cdiz, C/Chile 1,
Cdiz, Spain, CP 11002, [email protected]
(3)
Department of Computer Languages and Systems, University of Cdiz, C/Chile 1, CP 11002,
Cdiz, Spain, [email protected]

ABSTRACT
In order to study their current practices and improve on them, manufacturing firms need to view their
processes from several viewpoints at various abstraction levels. Several notations have been developed
for this purpose, such as Value Stream Mappings, IDEF models or the Process Specification Language.
More recently, the BPMN 2.0 standard from the Object Management Group has been proposed for
modeling business processes. A process organizes several activities (manual or automatic) into a single
higher-level entity, which can be reused elsewhere in the organization. Its potential for standardizing
business interactions is well-known, but there is little work on using BPMN 2.0 to model manufacturing
processes. In this work some of the previous notations are outlined and BPMN 2.0 is positioned among
them after discussing it in more depth. Some guidelines on using BPMN 2.0 for manufacturing are
offered, and its advantages and disadvantages in comparison with the other notations are presented.
Keywords: manufacturing process, modeling, PSL, IDEF, BPMN.

RESUMEN
Para estudiar sus prcticas actuales y mejorarlas, las empresas de fabricacin necesitan representar sus
procesos bajo distintos puntos de vista y niveles de abstraccin. Dentro de la Ingeniera de Fabricacin
se han aplicado diversas tcnicas para ello, como Value Stream Mapping, IDEF o el Process
Specification Language. Recientemente ha surgido BPMN 2.0, un estndar del Object Management
Group para modelar procesos de negocio. Estos renen una serie de tareas o "servicios" manuales y/o
automticos en un proceso de mayor nivel, que pasa a formar parte del catlogo de servicios de la
empresa y puede ser reutilizado en un modelo de nivel superior. Esta capacidad de BPMN 2.0 y su
potencia de representacin lo hacen atractivo en el modelado de procesos de negocio en general, pero
hay pocos trabajos que investiguen sobre la aplicacin de BPMN 2.0 para representar procesos de
fabricacin. En este trabajo se da una visin general de los diversos lenguajes de modelado disponibles
y se sita a BPMN 2.0 entre ellos, tras discutirlo ms en profundidad. Se sugiere cmo usar BPMN 2.0
en fabricacin, y se presentan ventajas e inconvenientes respecto a otras notaciones.
Palabras clave: procesos de negocio, procesos de fabricacin, modelado, PSL, IDEF.

1. Introduction
In order to study their current practices and improve on them, manufacturing firms need to view their
processes from several viewpoints at various abstraction levels. Several notations have been used for this
purpose, such as IDEF3, the Process Specification Language or Value Stream Mappings. More recently,
the BPMN 2.0 standard from the Object Management Group has been proposed for modeling business
processes, using three kinds of views: collaborations, processes and choreographies. BPMN is intended as
a bridge between business process design and process implementation [1]. It has gained considerable

momentum in the recent years, with over 73 implementations by various vendors. However, there is little
work on comparing BPMN with previous notations.
In this work some of the existing notations are outlined and BPMN 2.0 is positioned among them after
discussing it in more depth. Some guidelines on using BPMN 2.0 for manufacturing are offered, and its
advantages and disadvantages in comparison with the other notations are presented.
The rest of this work is structured as follows: Section 2 briefly introduces the basic elements of IDEF3,
PSL and VSM. Section 3 describes BPMN 2.0 more in depth. Section 4 includes a case study based on a
textual description of a manufacturing process, which is mapped to each of the notations under study.
Section 5 sums up the results from Section 4 and offers some additional remarks. Section 6 lists some of
the related work in the literature. Finally, Section 7 offers some conclusions on the present study.
2. Selected Notations
In this section, three of the notations previous to the inception of BPMN 2.0 will be described: IDEF3,
PSL and VSM. The authors believe that these three notations are a representative sample of the existing
notations, as they cover textual and graphical notations for various purposes: process specification,
process reengineering, reasoning about processes and process interchange formats.
Many other notations exist. The survey by Aguilar-Savn in [2] covers an extensive range of flow-based
notations: however, it predates BPMN 2.0. Zor et al. have described a limited manual mapping from
VSM to BPMN models in [3]: our approach will focus on the relative strengths of the two notations,
rather than try to define a mapping between them. Initial work on PSL produced an extensive comparison
of the capabilities of the notations available at the time [4].
2.1 Integrated DEFinition for Process Description Capture Method (IDEF3)
According to the original report, IDEF3 was created specifically to capture descriptions of sequences of
activities [5]. IDEF3 uses two kinds of models: process schematics and object schematics. Process
schematics describe the valid sequences of the Units of Behavior (UOBs) in the process. Object
schematics describe the kinds of objects present in the system, their relationships and their state
transitions. Node and link shapes for IDEF3 process and object schematics are shown in Figure 1.
Process schematics represent the UOBs as boxes with textual labels and unique identifiers. Precedence
links specify valid sequences of UOB activations. There are two types of precedence links: simple and
constrained. A simple precedence link from A to B only indicates that whenever A and B both happen, A
must happen before B. A may happen and not B, B may happen and not A, or any number of UOBs not
included in the process schematic may happen between A and B. Constrained precedence links can
further limit the valid possibilities. Finally, junctions can split or join paths. AND junctions activate or
join all related paths, OR junctions only some, and XOR junctions exactly one.
Object schematics represent the possible states for each object in the system. Links relate different
objects, represent their state transitions or classify them. A state transition from A to B means that object
b can only be in state B after object a has been in state A. Object a may be the same as object b or not.
Users set conditions on transitions or states by linking them to UOBs from the process schematics.
IDEF3 allows for a hierarchical decomposition of both process and object schematics: starting with a
high-level view, modelers can drill down into the detailed descriptions. It also allows modelers to
indicate where information is hidden about parts, object categories and other constructions.
2.2 Process Specification Language (PSL)
The Process Specification Language, also known as ISO 18629:2004, is a textual notation for describing
manufacturing processes [6]. Its goal is to allow different applications to exchange process data, as STEP
is used to exchange information about parts. To achieve that interoperability, PSL is organized as an
ontology of concepts, related to each other using axioms and definitions. PSL is organized into several
layers. This work only provides a brief introduction: other works describe PSL in more depth, such as [7].
The main concepts in PSL constitute the PSL-Core. There are four kinds of entities in PSL processes:
activities, objects, activity occurrences and timepoints. Activities can have zero or more occurrences: the
activity Stamp may happen once, several times, or not at all. Timepoints are linearly ordered from a
timepoint before all others in the past (inf-), to a timepoint after all others in the future (inf+). Every
activity occurrence and object happens or exists between two timepoints.

Figure 1. Selected subset of the IDEF3 notation

The next layer in PSL is the Outer Core, with additional concepts and definitions that are commonly
used, such as subactivities: activities nested inside others. The PSL-Base layer extends the PSL Outer
Core with more specialized terms, such as durations (intervals between timepoints) and resource
handling, among others. The outermost layer only includes optional extensions.
2.3 Value Stream Mapping (VSM)
Lean manufacturing strives to reduce costs and increase flexibility by removing waste (muda) from the
manufacturing process. VSM is a tool for identifying issues and creating improvement plans to reduce
waste. A value stream contains all the actions required to bring a product to the customer. This section
will refer to the VSM workbook from the Lean Enterprise Institute, focused on the production flow [8].
Factory icons represent external plants. Incoming and outgoing shipments are represented using a truck
icon and a broad arrow. The manufacturing process is divided into process boxes: each box is a sequence
of steps in which materials flow continuously. Elements may be connected by information flows (regular,
electronic or go see) or material flows (push, pull, FIFO or sequenced pull). Material flows usually
indicate the kind of inventory handling involved: accumulated inventory, supermarkets or buffer stock.
Specific icons are available for load-leveling (heijunka) boxes, kanban-based systems, operators and
possible improvements (kaizen bursts). The graphical icons to be used in Section 4 are listed in Figure 2.

Figure 2. Selected subset of the Value Stream Mapping notation

3. Business Process Modeling Notation 2.0 (BPMN 2.0)
In the recent years, interest in modeling business processes for reengineering, simulation and execution
has steadily increased. The proliferation of low-level notations for process execution motivated the
creation of BPMN as a high-level notation which could be used by both business analysts and software
developers. BPMN 2.0 has added formal execution semantics based on Petri networks and several file
formats to the specification, making it usable both for process design and process enactment [1]. Part of
the notation is shown in Figure 3.
Activities represent units of work in the process. Activities may have sub-processes describing them in
more detail. Some of these sub-processes may be started in response to an event. Activities performed by
different stakeholders will be usually placed in different pools, which may be further divided into lanes.

The type of a task is noted by decorating it with an icon in the upper left corner. Decorations can indicate
aspects such as parallelism or iteration: for the sake of brevity, they have been omitted.
Events are situations to which the BPMN process reacts. Events are drawn as circles: the line style of the
circle indicates if it is a start event, an intermediate event (handled during the execution of a process or
activity) or a finish event. Inside the circle, an icon indicates what kind of event is handled.
Finally, activities and events are connected together through flows and gateways. Message flows model
the information exchanges between the participants, and sequence flows control the execution of the
activities. Sequence flows may converge or diverge through gateways, similar to IDEF3 junctions.

Figure 3. Selected subset of the BPMN 2.0 notation

4. Case Study
In the previous sections, IDEF3, PSL, VSM and BPMN 2.0 were presented. This section shows how to
use each notation to model a hypothetical manufacturing process described in natural language. In the
next section, the models will be used to compare the notations.
4.1 Textual description
The company under study (Company A) receives tobacco and cellulose acetate and produces cigarettes.
Tobacco preprocessing consists of several steps and slightly varies from product family to product family.
180-200kg boxes of raw tobacco are regularly received from external suppliers. First, the moisture in the
raw tobacco is increased and casings are added. Next, tobacco is blended, cut, compressed and packaged.
Optionally, the tobacco may be expanded before packaging to produce the light variants.
Cigarette filters are produced from cellulose acetate tows, separating the fibers before adding a plasticizer
and cutting the filter rods into individual filters. After letting the filters harden on trays, they are sent to
the cigarette making machine.
Filters and processed tobacco are received by another department, which wraps the tobacco and adds the
filters, joining the cigarette with the filter using tipping paper. These cigarettes are then packed into
boxes, which are bought by distributors and finally sold by retailers.
It is important to note that Company A recently joined a larger group and needs to synchronize its inhouse information system with the SAP R/3 installation in use within the group. This includes inventory
levels, manufacturing reports and production forecasts. Shipments from suppliers and to distributors are
handled by an external company, part of the same group.
4.2 IDEF3 model
Figure 4 is the IDEF3 object schematic for the manufacturing process. Objects represent intermediate
products, from raw materials up to packaged goods. The UOB boxes have a slightly different notation, as
they refer to UOBs in the omitted process schematic. They describe the process steps required for each
state transition. There are two types of processed tobacco (regular and expanded), and therefore two types
of cigarettes (regular and light). Most UOBs have a single digit: their contents have not been expanded.
However, as an example, the Make filters UOB (#2) has been expanded into four nested UOBs, with

identifiers from 2.1.1 to 2.1.4. Additionally, the Filters object node has a different line style and is
decorated with a C, indicating there are several types of filters not shown in the diagram.

Figure 4. IDEF3 model for the case study

4.3 PSL description
Due to space constraints, this section only includes the fragment dedicated to preprocessing the tobacco.
To simplify the discussion, it is shown by parts. occurrence_of has been shortened to oof,
subactivity_occurrence to sao) and min_precedes to mpr, to save space. The next part
describes the activities and subactivities dedicated to preprocessing tobacco and how they are ordered:
(forall (?opt)
(implies (oof ?opt PreprocessTobacco)
(exists (?oim ?oac ?orc ?ob ?orb ?ocut ?ocomp)
(and (oof ?oim IncreaseMoisture) (oof ?oac AddCasings)
(oof ?orc RefillCasings) (oof ?ob Blend)
(oof ?orb RefillBlender) (oof ?ocut Cut) (oof ?ocomp Compress)
(sao ?oim ?opt) (sao ?oac ?opt) (sao ?orc ?opt) (sao ?ob ?opt)
(sao ?orb ?opt) (sao ?ocut ?opt) (sao ?ocomp ?opt)
(mpr ?oim ?oac) (mpr ?oac ?orc) (mpr ?oac ?ob)
(mpr ?ob ?orb) (mpr ?ob ?ocut) (mpr ?ocut ?ocomp)))))
These constraints ensure the occurrences of the subactivities in Preprocess Tobacco go in the order
Increase Moisture, Add Casings, Blend, Cut and Compress. In addition, Refill Casings
should happen some time after Add Casings, and Refill Blender some time after Blend.
The next part describes the machines and materials used at each step, using PSL resource theory.
Resources are described in terms of how an activity using a resource affects other activities which require
that resource. Reusable resources can be always used after the activity which uses them completes.
Possibly reusable resources require that a setup activity completes before they can be reused. Wearable
resources may not be usable at some point in the future. Consumable resources can never be reused.
Finally, an activity may also create or modify a resource.
(reusable Moisturizer IncreaseMoisture)
(possibly_reusable CasingSpreader AddCasings)
(possibly_reusable Blender Blend)
(wearable Cutter Cut)
(reusable Compressor Compress)
(consumable DryTobacco IncreaseMoisture)
(creates IncreaseMoisture MoistTobacco)
(consumable MoistTobacco AddCasings)
(creates AddCasings MoistTobaccoWithCasings)
(consumable MoistTobaccoWithCasings Blend)
(creates Blend BlendedTobacco)
(modifies Cut BlendedTobacco)
(modifies Compress BlendedTobacco)

4.4 VSM model

Figure 5 presents a VSM schematic describing the different material and information flows in the plant.
Suppliers provide the required tobacco and cellulose acetate tows once a week and these are pushed
through the process, which performs 2 weekly shipments of cigarette boxes. There are two information
systems communicating with the plant: an in-house system sends daily orders to the tobacco
preprocessing area and weekly orders to the filter manufacturing area, and receives regular notifications
about the shipments. The SAP/R3 system from the parent company sends weekly manufacturing
schedules and receives periodic production and inventory status reports.
4.5 BPMN 2.0 model
Figure 6 is a BPMN 2.0 model of the manufacturing process. The model is divided into several lanes: one
for each participant in the process. Lanes do not need to represent every action taken by a participant: for
instance, this diagram only shows the activities from the parent company and in-house IT directly related
to this manufacturing process. The lane for the logistics company is completely empty: all the model
shows is that the plant sends shipment requests to it after a batch is done.
The model indicates that the plant receives every day the batches to be produced, and repeats the basic
manufacturing process for each of them. Repetition in BPMN 2.0 is modelled by marking the repeated
activity (Produce batch) with a small circle-shaped arrow. The contents of Produce batch are very
similar to the IDEF3 process schematic from which Figure 4 was produced. The BPMN model adds the
capability to model the messages sent to the other participants. An event-based subprocess (marked with a
dashed rectangle with rounded corners) indicates that when there is a fault, a message is sent to the inhouse IT system notifying that the manufacturing of a certain batch was aborted.

Figure 5. VSM model for the case study

5. Results and Discussion
In the previous section, the same manufacturing process has been described from several viewpoints,
using IDEF3, PSL, VSM and BPMN 2.0. This section will compare the expressive capabilities of these
notations for several important aspects in manufacturing processes. Table 1 summarizes these results.
All the notations allow for defining valid sequences for the tasks in the manufacturing process. VSM uses
very high-level tasks, dividing the process only where continuous flow is interrupted. IDEF3 and BPMN
2.0 model sequences of activities, which can diverge into different paths or converge into one path using
junctions (IDEF3) or gateways (BPMN). BPMN can also describe what to do if something goes wrong
(faults), how to undo changes (compensation) and how to respond to signals. Though IDEF3 and BPMN
2.0 allow unspecified activities to be inserted between those in the models, PSL is the most flexible
notation for activity sequencing, due to its use of precedence relations.
IDEF3 and VSM do not explicitly model timing constraints: they can only be emulated through textual
descriptions. BPMN 2.0 allow for setting alarms at certain times, frequencies or delays, as shown in the
models. PSL includes a rich set of theories for the duration of activities.

Figure 6. BPMN 2.0 model for the case study

Table 1. Comparison summary
IDEF3

PSL

VSM

BPMN 2.0

Activity sequences

Fine-grained
(control flows)

Fine-grained
(precedence
constraints)

Coarse
(material
flows)

Fine-grained
(control flows,
events)

Timing constraints

Implicit (text)

Explicit
(durations)

Implicit (text)

Explicit (alarms)

Machine/operator
assignments

Implicit (objects)

Explicit
(resources)

Implicit (data
boxes)

Implicit (pools)

Material flows

Implicit (object
transitions)

Explicit
(resources)

Explicit

Implicit (messages)

Information flows

Needs
IDEF0/IDEF1X

Needs extensions

Explicit, no
internal
structure

Explicit, relies on
extensions for
internal structure

Machine/operator assignments can be emulated in IDEF3, relating the object node with the machine to
the object node. VSM does not model assignments and only includes the parameters of the process which
affect material flow, such as changeover or cycle time. BPMN does not explicitly model
machine/operator assignments, but they can be emulated using pools and lanes if desired. PSL models
machines and operators and resources, and includes several theories for describing constraints on them.
Material flows can be emulated in IDEF3 using state transitions between object nodes, as in Figure 4.
BPMN cannot accurately model continuous material flows, but can emulate material flow in discrete
manufacturing through messages with the part information. Material flows can be explicitly modeled in
VSM. PSL models normally describe materials as consumable or renewable resources.
Information flows cannot be described with a single IDEF3 model: supporting IDEF0
models will be usually required. PSL does not model the information exchanged
manufacturing steps directly: however, a new extension has been proposed for that [9].
information flows directly, but does not provide any formal mechanisms to describe

and IDEF1X
between the
VSM models
their internal

structure. BPMN explicitly models the messages exchanged between each of the participants, but relies
on vendor-specific extensions to describe the structure of the messages.

6. Conclusions and future work

The aim of this work was to position BPMN 2.0 among the existing notations. IDEF3 was selected as a
graph-based notation for modeling the expected behavior of a process, PSL as a constraint-based textual
notation, and VSM as a graph-based notation for the material and information flows in a manufacturing
process. After introducing the notations, a case study modeling a tobacco manufacturing process was
presented. Using the obtained models, the expressive power of each of these notations was compared.
BPMN 2.0 can be seen as a superset of IDEF3 process schematics, adding explicit support for modeling
the participants in the process, event handlers and message exchanges. However, BPMN cannot model
the existing objects and their transitions, like IDEF3 object schematics can.
As a constraint-based notation, PSL has more expressive power than BPMN. However, it is also much
harder to use than BPMN, and there are fewer tools for it: the authors have only found Tau and Vampire,
two theorem provers. Messaging is still not part of the PSL standard: once it is included, it would be
interesting to translate BPMN to PSL and then enhance the PSL description.
VSM is a much simpler notation than BPMN and only provides a very high-level picture of the process,
focusing on the material and information flows rather than the exact sequence of operations. For this
reason, VSM is observed to complement BPMN: the former is a quick pen-and-paper tool for iterative
process improvement, and the latter is for detailed process design and enactment.
As it is, the authors would recommend BPMN 2.0 in two areas: describing the information-intensive
activities which support the manufacturing process, and describing repetitive manufacturing processes
with few variations. For describing families of interrelated manufacturing processes with high degrees of
flexibility, PSL would be the best choice: however, the tools for PSL could be improved.
From the present study, several lines of future work can be drawn. Once the messaging extensions are
integrated into the standard PSL ontology, it would be interesting to define an automated translation of an
appropriate subset of BPMN into PSL. Users could define the information environment of the
manufacturing process graphically, and then describe the process itself in more detail through PSL.
Another interesting line of work would be developing more accessible tools for PSL, which performed
specific kinds of analysis on the manufacturing process.

7. Acknowledgements
This paper was funded by the research scholarship PU-EPIF-FPI-C 2010-065 of the University of Cdiz.
8. References
[1] Object Management Group, Business Process Modeling Notation 2.0, 03-Jan-2011. [Online].
Available: http://www.omg.org/spec/BPMN/2.0/Beta2/. [Accessed: 17-Apr-2011].
[2] R. S. Aguilar-Savn, Business process modelling: Review and framework, International Journal of
Production Economics, vol. 90, no. 2, pp. 129-149, Jul. 2004.
[3] S. Zor, K. Grlach, and F. Leymann, Using BPMN for Modeling Manufacturing Processes, in
Proceedings of the 43rd CIRP Conference on Manufacturing Systems, Vienna, Austria, 2010.
[4] A. Knutilla et al., Process Specification Language: An Analysis of Existing Representations.
Gaithersburg, MD, USA: National Institute of Standards and Technology, 1998.
[5] R. J. Mayer, C. P. Menzel, M. K. Painter, P. S. de Witte, T. Blinn, and B. Perakath, IDEF3 Process
Description Capture Method Report. Texas, USA: Knowledge Based Systems Inc., 1995, p. 236.
[6] International Standards Organization, ISO 18629-1:2004 - Process specification language - Part 1:
Overview and basic principles. 2004.
[7] C. Bock and M. Gruninger, PSL: A semantic domain for flow models, Software & Systems
Modeling, vol. 4, no. 2, pp. 209-231, 2005.
[8] M. Rother and J. Shook, Learning to See: Value Stream Mapping to Add Value and Eliminate
MUDA. Lean Enterprise Institute, 1999.
[9] C. Bock, Interprocess Communication in the Process Specification Language. Gaithersburg, MD,
USA: National Institute of Standards and Technology, 2006.