MODELING OF COLLABORATIVE PRODUCTION SYSTEMS

USING COLOURED PETRI NETS

Percy Igei, Carlos E. Cugnasca, Fabrício Junqueira, Paulo E. Miyagi

Escola Politécnica, University of São Paulo, São Paulo, Brazil

José I. García

Mechanical Engineering School, University of Valle, Cali, Colombia

Keywords: Production system, Publish/Subscribe, Collaborative production system model, Coloured Petri Net.

Abstract: Production Systems are undergoing continuous changes in order to be more competitive in a globalized

market. In this scenario, there is a tendency to geographically distribute the production process to reduce

development and production costs. The interaction between these components is characterized by the

presence of a high number of concurrent and asynchronous processes. As a result, the development of

production systems is a complex and difficult task. To cope with this complexity, this paper proposes a

systematic approach for the modeling of collaborative production systems. This approach explores the

potential of formal modeling language such as Colored Petri net to represent and analyze the dynamic

behavior of the system. Additionally, the publish/subscribe paradigm is introduced to establish the

communication in the coordination process of the system.

1 INTRODUCTION

Production systems (PS) perform a process using

materials, equipment, human resources and other

physical entities so as to produce goods or services

(Villani et al., 2007). Increased competition in

productive organizations triggered the development

of different approaches in order to conceive new

kinds of PS. In this sense, initially, a change from a

PS centralized structure to a distributed structure

was stimulated by economic factors, in which a

displacement of manufacturing plants was made to

countries with reduced operating costs (Shah, 2005).

However, nowadays, the paradigm changes are

more stimulated by strategic parameters, seeking a

quick production response, focused on the demand

for customized products (Grefen et al., 2009,

Shimizu et al., 2007). In this scenario, according to

Ko and Nof (2010), recent and emerging advances in

hardware (e.g., pervasive computing devices,

wireless sensor networks, nano-electronics) and

software (e.g., multi-agent systems, artificial

intelligence, workflow and information integration)

enables to a new class of PS, called Collaborative

Production System (CPS).

The CPS has a distributed structure with a high

degree of flexibility, in which customers, suppliers,

operators and production facilities are

geographically dispersed, interacting with the intent

to develop a customized product. The CPS integrates

new information technologies and controller devices

in order to reach a flexible structure (Godara, 2010).

The structural flexibility permits the expansion of

the CPS and allows rapid reconfiguration to produce

a wide range of products (Leitao and Colombo,

2006).

The distributed structure of the CPS makes

necessary the use of flexible communication

paradigms that address the dynamic and decoupled

nature of these applications. The publish/subscribe

paradigm is receiving increasing attention and is

claimed to provide the loosely coupled form of

interaction required in large distributed systems

(Schneider and Farabaugh, 2004, Eugster et al.,

2003).

The design and implementation of the CPS are

complex tasks. In this sense, several approaches

were proposed for modeling CPS. In Matsusaki and

257

Igei P., E. Cugnasca C., Junqueira F., E. Miyagi P. and I. García J..

MODELING OF COLLABORATIVE PRODUCTION SYSTEMS USING COLOURED PETRI NETS.

DOI: 10.5220/0003353302570260

In Proceedings of the 1st International Conference on Pervasive and Embedded Computing and Communication Systems (PECCS-2011), pages

257-260

ISBN: 978-989-8425-48-5

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

Santos Filho (2006), a multifaceted decomposition

of a system was introduced in order to meet the CPS

requirements of CPS. Thus, the logic control of the

system is classified according to the semantics and

functionality. In Junqueira and Miyagi (2009), a

framework for distributed simulation was defined

which the Petri net is used as a tool to model large

and geographically dispersed PS. In Garcia et al.

(2009), a characterization of a framework was

presented for integration and coordination of

services provided by components of a

geographically dispersed PS. However, although

these studies address the issue of distributed and

dispersed PS from different points of view, they do

not approach adequately the complexity of the

communication process in terms of their dynamic

and quantity of components that are involved in the

coordination process.

Motivated by these facts, a systematic approach is

introduced here for the modeling of CPS. To

represent the behavior of the CPS, the use of two

abstraction models is proposed. Initially, Message

Sequence Chart (MSC) is used to represent the

interaction between the components in different

operation scenarios. Then, the functionality of each

component is modeled by CPN. Finally, the

components are integrated into a single model to

analyze the properties and behavior of the overall

system by means of CPNTools (Jensen et al., 2007).

The text is organized as follows: section 2 presents

the fundamental concepts that support the work.

Section 3 presents the architecture of the distributed

SP. The 4

section presents a systematic approach

for modeling and analysis based on CPN. Finally, in

section 5 are the main conclusions obtained in this

work.

2 BASIC CONCEPTS

2.1 Publish/Subscribe Paradigm

In the last years, this paradigm of communication

has gained importance in the design and

development of different class systems like stock

exchange systems, air traffic control systems,

defense systems and CPS and pervasive computing

applications (Abawajy, 2009). This is due to the

capacity of the publish/subscribe paradigm to

completely decouple communication participants,

allowing the development of applications with a

high degree of flexibility (Ryll and Ratchev, 2008).

This is one of the motivations for the use of this

paradigm in this work.

In the publish/subscribe paradigm, communication

can be anonymous, asynchronous and multicasting.

The combination of these characteristics makes the

publish/subscribe paradigm well suited to a variety

of application areas, allowing the development of

systems with a high degree of flexibility. However,

this advantage is counterbalanced by increased

complexity in the sense of understanding the

functioning of the system as a whole. Furthermore,

although the components of a collaborative system

can work properly when examined individually, they

may have an inappropriate behavior when working

in a cooperative manner (Baresi et al., 2005).

2.2 Coloured Petri Net (CPN)

The Coloured Petri net (CPN) is a graphical

language that combines the capabilities of PN with

the capabilities of a high-level programming

language, in which the PN provides a graphical

environment and supports the formal description of

the system and the programming language facilitates

the definition of data types (Jensen and Kristensen,

2009). Thus, a compact and parametric model can be

created. This type of PN allows the construction of

hierarchical models that facilitate the description of

the system. The CPN models can be hierarchically

structured as a collection of small connected

components in which each component corresponds

to a model in CPN. This model structuring can be

performed by replacing substitution transitions

(specific terms of Petri net are in Arial Narrow). A

substitution transition is a special type of transition used

in the CPN to represent an instance of another

model. This feature of the CPN allows the reuse of

components already built, reducing the time and cost

of development (Jensen et al., 2007). A more

detailed description about the elements of CPN can

be found in (Jensen and Kristensen, 2009).

3 PROPOSED ARCHITECTURE

The proposed architecture of the CPS considers that

the dynamic evolution of the states is performed

concurrently. Thus, CPS performs tasks in a parallel

and independent way. Fig. 1 shows the architecture

of the CPS.

The communication between the components that

integrates the CPS is performed by exchanging

asynchronous messages according to the

publish/subscribe paradigm. The messages are

grouped into topics. The applications are not

directly coupled to each other and communication is

PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems

258

Figure 1: Proposed architecture of CPS.

carried out through the publication and subscription

of data topics. The term component represents an

independent unit that is functionally autonomous,

which encapsulates its internal state and behavior

(Szyperski, 2002). The proposed architecture is

characterized by the composition of reusable

components that are developed independently and

are integrated to achieve a final goal.

The CPS applications are composed of

components that can be reused to build other

applications. Thus, a reduction of cost and time is

achieved (Sommerville, 2007). Fig. 1 shows the

utilization of “Component 1” in the construction of

three others components. The models of the

components are developed using the CPN. The input

and output interfaces of the component are

represented by places and the data are modeled by

colored tokens. Once the models of the components

are constructed, they can be simulated and analyzed

in CPNTools to verify their static and dynamic

properties.

One of the main features of the CPN is the

possibility to assign attributes to the tokens, in order

to differentiate one token from another. This feature

is used to perform the communication component

model. Thus, a message can be modeled as a colored

token that identifies the data type (topic) and the data

value (content).

4 MODELING PROPOSED

PROCEDURE

A modular approach based on CPN is adopted here,

and the complete system is divided into functional

independent modules, which interact by exchanging

messages via a communication network. This

modeling procedure allows identifying each

functional independent module, with different

degree of detailing (i.e. assembly system, robot,

actuator, sensor, etc.), which can be divided into

reusable components. The behavior of these

components and their interaction with the others is

represented by CPN models. Hence, it is possible to

represent the whole system in a computational

model that can be used to perform different kinds of

analysis (Jensen et al., 2007). The proposed

modeling procedure comprises six steps, briefly

described as follows:

Step 1: The scope and the main functionalities of the

CPS are identified and delimited. At this stage, the

stakeholders involved in the project discuss the CPS

objectives and requirements.

Step 2: The topics of data are defined in order to

establish the communication between the functional

independent modules. This stage uses the

publish/subscribe paradigm to perform the

communication process.

Step 3: The message sequence chart (MSC) is used

to model the interaction of the CPS. Thus, the

behavior of the system can be represented by the

interaction of the components in different scenarios.

Step 4: The basic components of the CPS are

identified. The components are constructed

considering the reuse of these models to develop

others CPS.

Step 5: This step consists of the integration of the

components in order to obtain the complete model of

each CPS.

Step 6: The CPS is integrated into a single model

that can be simulated. The simulation of the

computational model allows the designer to refine or

to create new strategies or specifications of the

system, detecting errors and mistakes before the

CPS implementation (Junqueira and Miyagi, 2009).

5 CONCLUSIONS

Although the use of distributed and intelligent

components facilitates the production process, it

introduces difficulties in the design of the control

system, making the understanding of the overall

CPS process a complex task. In this sense, this paper

presents a systematic procedure for specifying the

CPS. The approach proposes the use of the

publish/subscribe paradigm to establish the

communication between the components of the

distributed system. This approach aims to facilitate

the understanding of the communication process in

systems that are complex in terms of their dynamic

and number of components. In order to validate the

proposed procedure, it is been applied in the

modeling of a manufacturing system that emulates a

dispersed CPS. This system is installed at the

MODELING OF COLLABORATIVE PRODUCTION SYSTEMS USING COLOURED PETRI NETS

259

Department of Mechanical and Mechatronic

Engineering at the University of São Paulo

(Junqueira and Miyagi, 2009).

ACKNOWLEDGEMENTS

The authors would like to thank the Brazilian

governmental agencies CAPES, CNPq and FAPESP

for their partial support to this work.

REFERENCES

Abawajy, J. H. 2009. Advances in Pervasive Computing:

Guest Editorial. International Journal of Pervasive

Computing and Communications, 5, 4-8.

Baresi, L., Ghezzi, C. & Zanolin, L. 2005. Modeling and

Validation of Publish/Subscribe architectures. Testing

Commercial-off-the-shelf Components and Systems, 3,

273-291.

Eugster, P. T., Felber, P. A., Guerraoui, R. & Kermarrec,

A. M. 2003. The many faces of publish/subscribe.

ACM Computing Surveys (CSUR), 35, 114-131.

Garcia, J. I., Fattori, C., Junqueira, F. & Miyagi, P. E.

Year. Framework for Collaborative Manufacturing

Systems Based in Services. In: 20th International

Congress of Mechanical Engineering (COBEM 2009),

2009 Gramado, Brazil.

Godara, V. 2010. Pervasive Computing for Business:

Trends and Applications, Sidney, Igi Global.

Grefen, P., Mehandjiev, N., Kouvas, G., Weichhart, G. &

Eshuis, R. 2009. Dynamic business network process

management in instant virtual enterprises. Computers

in Industry, 60, 86-103.

Jensen, K., Kristensen, Lars, M. & Wells, L. 2007.

Coloured Petri Nets and CPN Tools for modelling and

validation of concurrent systems. International

Journal on Software Tools for Technology Transfer, 9,

213-254.

Jensen, K. & Kristensen, L. 2009. Coloured Petri Nets:

Modeling and Validation of Concurrent Systems,

Springer-Verlag New York Inc.

Junqueira, F. & Miyagi, P. E. 2009. Modelagem e

simulação distribuída de sistema produtivo baseados

em rede de Petri. Sba: Controle & Automação

Sociedade Brasileira de Automatica, 20, 1-19.

Ko, H. & Nof, Y. 2010. Editorial-Special Issue on

Collaboratin Support Systems (CSS). International

Journal of Computers, Communications and Control,

V, 6-7.

Leitao, P. & Colombo, A. Year. Petri net based

methodology for the development of collaborative

production systems. In: IEEE Conference on

Emerging Technologies and Factoring Automation,

ETFA´06 2006 Prague. 819-826.

Matsusaki, C. & Santos Filho, D. F. Year. Modeling of

Distributed Collaborative Control Systems of

Production Systems. In: ABCM Symposium Series in

Mechatronics, 2006. 345-352.

Ryll, M. & Ratchev, S. 2008. Application of the Data

Distribution Service for Flexible Manufacturing

Automation. International Journal of Mechanical,

Industrial and Aerospace Engineering, 2, 187-194.

Schneider, S. & Farabaugh, B. 2004. Is DDS for You?

White paper by Real-Time Innovations.

Shah, N. 2005. Process industry supply chains: Advances

and challenges. Computer Aided Chemical

Engineering, 29, 1125-1235.

Shimizu, Y., Zhang, Z. & Batres, R. 2007. Frontiers in

Computing Technologies for Manufacturing

Applications, Springer Verlag.

Sommerville, I. 2007. Software Engineering, Addison

Wesley.

Szyperski, C. 2002. Component Software: Beyond Object-

Oriented Programming Addison-Wesley.

Villani, E., Miyagi, P. E. & Valette, R. 2007. Modelling

and analysis of hybrid supervisory systems: a Petri net

approach, Springer Verlag.

PECCS 2011 - International Conference on Pervasive and Embedded Computing and Communication Systems

260