Copyright © IFAC Low Cost Automation,
Buenos Aires, Argentina, 1995
THE APPLICATION OF ENHANCED MARK FLOW GRAPH
IN REAL TIME CONTROL SYSTEMS
P. E. Miyagi*; L. M. Camarinha-Matos**; D. J. Santos Filho*;
J. Barata** and J. Arakald*
*Universil/ade de Siio Paulo - Escola Politecnica
Av. Pro! Luciano Gualberto - Trav 3, N.,380, CEP 05508-900 - Sao Paulo, SP,- Brasil
**UniversUiade Nova de Lisboa and Uninova Institute
Quinta da Torre - 2825 Monte Caparica - Portugal
Abstract: The aim of this work is the evaluation of the modeling power of Enhanced Mark Flow Graph (E-
MFG) when dealing with flexible manufacturing systems and the presentation of an approach to link the models
to the real system. Main concepts regarding E-MFG and its application in the description of the control part are
described as well as the operative part of production systems. Through the information associated with the
structural components of E-MFG it is possible to represent complex control strategies involving different tasks
of the manufacturing systems in a highly flexible environment. An object oriented approach for integrating the
various heterogeneous controllers available in a manufacturing system is presented, resulting in an abstract
machine that realizes the operative part and, thus, ensures the link between the E-MFG models and the physical
system.
Keywords: discrete event dynamic system, Petri nets, Mark Flow Graph, manufacturing system, system
integration
1. INTRODUCTION
The difficulty in controlling a flexible manufacturing
system is strongly related to its level of automation
and flexibility. On the other hand, the concept of
automatic process control can be generalized to the
case in which the control object is a dynamic
production system driven by events, i.e., a discrete
event dynamic system - DEDS (Santos F", 91) (Ho,
89). Therefore, this work approaches the problem of
controlling a flexible manufacturing system with
high functional flexibility which can be treated as
DEDS.
Recently, the technological evolution of the
equipments used in automated manufacturing systems
has increased substantially the degree of automation
of these systems. In this way, increases in the
functional flexibility of manufacturing systems meet
the changes of the market dynamics with high degree
of autonomy. Nevertheless, in a shop floor we can
find a large variety of controllers, representing
various generations of technology.
147
Mark Flow Graph - MFG (Hasegawa, 84) (Masuda,
81) is a technique derived from Petri nets (Reisig, 92)
that is effective to represent and control the dynamic
behavior of DEDS in special cases where the system
must be described and controlled in a distributed way.
The major property of the MFG is that it explicitly
describes the control strategy of independent,
sequential and concurrent events. However, there are
some problems related to the modeling of systems
with complex dynamic behavior involving non
trivial control rules. To represent the control rules in
a consistent manner it is necessary to consider several
sub-graphs that need to be synchronized. The
resultant graph is difficult to be interpreted and
analyzed. This aspect motivated the proposition of
extensions to MFG to increase i;s modeling power
and a methodology to represent Integrated
Manufacturing Systems (IMS) by subsystems that
operate cooperatively. In this context, the Enhanced
Mark Flow Graph - E-MFG (Santos P, 93) is based
on individualized marks* and additional rules to
control the transitions firing.
* These concepts will be presented in detail in the
following sections.