HOLONIC ARCHITECTURE FOR A MULTIAGENT-BASED
SIMULATION TOOL
Nancy Ruiz, Adriana Giret and Vicente Botti
Department of Informatics Systems and Computing, Polytechnic University of Valencia, Camino de Vera SN, Valencia, Spain
Keywords:
Multi-agent systems, Simulation Tool, Holonic Architecture.
Abstract:
In the last few years Holonic Manufacturing Systems (HMS) have been used to improve some areas such
as control systems of manufacturing systems, monitoring and diagnosis systems, and Computer Integrated
Manufacturing architectures. This paper presents an application of the holonic approach together with the
Multi-agent Systems (MAS) to develop a Simulation Tool for manufacturing systems. The architecture for
this tool is related to functionalities that support the Manufacturing Model created by the User. This proposal
allows the User to take advantage of these two approaches to simulate a manufacturing environment according
to the manufacturing requirements.
1 INTRODUCTION
Current simulation tools usually offer useful tech-
niques to simulate manufacturing environments in
static models. Their main feature is the simulation
process based on static models. Some tools offer an
optimization function to improve the model before ex-
ecuting the simulation process, and other tools allow
the User to modify the model “on-the-fly”. However,
since manufacturing modeling changes according to
customer needs, simulation tools must also adapt to
them (Nikoukaran and Paul, 1999). To the best of
our knowledge, there are currently no simulation tools
that provide this kind of flexibility (Klingstam and
Gullander, 1999). However, we assume that if there
is a Simulation Tool that provides additional features
such as proactivity, flexibility, agility and reconfig-
urability it is possible to improve the Manufacturing
Model Creation and the Simulation Process for a User
(Law and McComas, 1999). Therefore, we propose
the application of Multi-Agent Systems (MAS) the-
ory and the Holonic Approach (HMS, 2004) to add
flexibility to manufacturing system simulation (Miller
and Pegden, 2000). The proposal is based on the unity
of two notions: holon and agent. On one hand, a
holon refers to a system (or phenomenon) that is both
This work has been partially supported from the Span-
ish government and FEDER funds under CICYT TIN2006-
14630-C03-01 and TIN2005-03395 projects, and by the
Mexican CONACYT under research grant N178874.
a whole in itself as well as part of a larger system. It
can be viewed as systems that are nested within each
other (Koestler, 1990). On the other hand, an agent
is an autonomous and flexible computational system
that is able to act in an environment (Wooldridge and
Jennings, 1995). Based on a comparative study of
these two notions made by Giret (Giret and Botti,
2004), in this work it can be concluded that an Agent
can be treated as a Holon. Since MAS is an exten-
sive approach that can be used for the distributed and
intelligent control of general systems, and since the
Holonic Manufacturing System (HMS) is a specific
manufacturing approach for distributed control sys-
tems, thus, it is possible to conclude that the devel-
opment of a Simulation Tool based on a Multi-agent
Methodology for Holonic Manufacturing Systems is
viable. In this Work the ANEMONA methodology
(Giret et al., 2005) is used for the tool specification.
Section 2 of this paper presents the basis of the ap-
proach of the HMS architectures and the MAS that
will support the Simulation Tool. Both the Model
of the shop floor and the Simulation Tool are treated
as a MAS that interact to provide additional features
such as flexibility and proactivity during the Simu-
lation Process. The proposed Simulation Process is
composed of two phases: Model Creation and Model
Simulation. Section 3 presents the HMS Architecture
for the Simulation Tool. Finally, some conclusions
and further research are outlined.
395
Ruiz N., Giret A. and Botti V. (2007).
HOLONIC ARCHITECTURE FOR A MULTIAGENT-BASED SIMULATION TOOL.
In Proceedings of the Ninth International Conference on Enterprise Information Systems - AIDSS, pages 395-398
DOI: 10.5220/0002360603950398
Copyright
c
SciTePress
2 A MAS BASED SIMULATION
TOOL
This paper presents an extension of a previous work
(Ruiz et al., 2006) in which a brief study was made
to identify the manufacturing simulation process. The
study distinguished two main phases in the simulation
process of a manufacturing system: Model Creation
and Model Simulation. The goal of the first phase is
to create a model that represents a real manufactur-
ing environment, the goal of the second phase is to
simulate this model where the user can observe the
system behavior, detect bottle necks, and adjust the
system to improve its functionality based on the sim-
ulation results. The purpose of the Simulation Tool
is to provide a Service Platform that executes specific
system processes that can be automatically improved
by adding proactivity and flexibility features. These
features will help the User during the Simulation Pro-
cess. The descriptions of the functional blocksassoci-
ated to the Simulation Process is presented below: In
the First Phase, the User builds a model while a Man-
ager coordinates the model creation by monitoring
when it is necessary to execute specific tasks accord-
ing to the current state of the system. The first step is
the importation and analysis of information from the
real world. This information is included in the ele-
ments to be used by a Modeller to create the Model.
The Model can be edited by the User whenever nec-
essary, and the User can specify the behaviour (at-
tributes) of each one of the model elements as desired.
In the Second Phase, the simulation of the model ob-
tained in Phase 1 is configured by the User (event
type, speed, and type of animation). Then, the sim-
ulation is executed according to the established con-
figuration, taking into account the attributes of each
model element. The User can add events in order
to include additional information during the simula-
tion process. When the simulation process is finished,
the results are analyzed. The analyzed results are in-
corporated into business reports and graphics to be
viewed by the User and can be reused by the system.
It is also possible to export the analyzed results for
external analysis.
3 USE CASE INTERACTIONS IN
THE SIMULATION TOOL
The ANEMONA methodology (Giret et al., 2005)
was used for the specificaction of the Simulation Tool
which includes the specification of System Goals, Use
Cases, and Holons related to the architecture of the
Simulation Tool. The identification of the System
Goals is made by using the HMS-CU Guidelines from
1 to 6 (Giret and Botti, 2006). 1. Importation of
the information from the real manufacturing system
to be modelled in order to analyze and associate it to
the specific icons of the Library Icons; 2. Creation
of a Model using an icon library by providing func-
tional schemas that will help the User to customize
each icon; 3. Simulation of a Model based on a spe-
cific configuration made by the User; 4. Supervision
of the simulation process by the automatic creation
or elimination of instances according to the needs of
the current state of the system; 5. Automatic analysis
of the results to be viewed by the User (Reports and
Graphics); 6. Reuse of the analyzed results by associ-
ating them to the corresponding icon behavior in the
Icon Library; 7. Exportation of the analyzed results
to external analysis tools; 8. Editing of interfaces to
extend the current framework.
3.1 Use Cases
The cooperation domains (-Use Cases- required
to satisfy the System Goals) are specified using
the HMS-CU Guidelines from ID 7 to 13 of the
ANEMONA methodology (Giret and Botti, 2006).
Specification of Use Cases and their Roles: As
result of using the HMS-CU guidelines, the iden-
tified roles that appear during the Simulation Pro-
cess are: the Import/Export Manager (IEM); the In-
put/Output Data Analyzer (IODA); the Synchroniza-
tion Manager (SYNM); the Modeller (MOD); the
Icon Manager(ICOM); the Model Checker (CHECK);
the Speed Manager (SPEM); the Animation Man-
ager (ANIM); the Event Generator (EVEG); the Sim-
ulation Planner (SIMP); the Report/Graph Manager
(REGM); and the Interface Manager (INTM).
In order to fulfil the communication needs among
entities in a Use Case, the following interaction rela-
tions (Fig. 1) have been detected.
- Manage the Data Importation. According to a
User request, the Synchronization Manager (SYNM)
asks the Import/Export Manager (IEM) to import data
from a real manufacturing system. Then the SYNM
asks the Input/Output Data Analyzer (IODA) to an-
alyze the imported data, and asks the Icon Man-
ager(ICOM) to associate this data to the correspond-
ing icons;
- Manage the Model Creation. When the data of
the real manufacturing system has been included in
the system, the SYNM generates instances of the
Modeller (MOD), the Model Checker (CHECK), the
ICOM and the Interface Manager (INTM) which all
cooperate with the User during Model Creation;
ICEIS 2007 - International Conference on Enterprise Information Systems
396
- Model Creation. The User creates a model using
the Modeller (MOD), which communicates with the
Icon Manager(ICOM). The ICOM provides the icons
that represent specific elements of the system (hu-
man, machinery, tool, etc.). Then, the Model Checker
(CHECK) verifies that the model is free of mistakes,
and the Interface Manager INTM provides the appro-
priate interfaces;
- Model Simulation Configuration. When the Model
has been created, the User requests the configuration
of the model simulation and the SYNM activates the
Modeller (MOD). The User requests the speed, type
of animation, and type of events required from the
MOD. The MOD communicates the selected speed to
the Speed Manager (SPEM), the type of selected an-
imation to the Animation Manager (ANIM), and the
selected type of events to be associated to the Model
during the simulation process to the Event Generator
(EVEG);
- Model Simulation. When the configuration of Model
Simulation has finished, the SYNM activates the
Simulation Planner (SIMP). The SIMP receives the
Model and starts to plan the simulation. It commu-
nicates with other components during model simula-
tion: the Speed Manager (SPEM) indicates the speed
at which events and their animation must be gener-
ated; the Event Generator (EVEG) generates events
according to the type of events selected and speed in-
dicated by the SPEM; and the Animation Manager
(ANIM) animates the icons according to the events
generated by the EVEG;
- Manage Analysis of Results Simulation. When
the simulation has finished, the SYNM generates in-
stances of the Input/Output Data Analyzer (IODA),
and the Report/Graph Manager (REGM), which both
cooperate during the Results Analysis and the Re-
port/Graph Generation;
- Analysis of Simulation Results. The Input/Output
Data Analyzer (IODA) analyzes the simulation results
to extract valuable data to be reused by the system
and delivers the analyzed results to the Icon Manager
(ICOM) which then associates this data to their corre-
sponding icons in the Icon Library;
- Report/Graph Generation.When the analysis has
finished, the Input/Output Data Analyzer (IODA)
delivers the analyzed results to the Report/Graph
Manager (REGM) which then generates reports and
graphs for the User;
- Results Exportation. When the User requests the ex-
portation of the results, the Synchronization Manager
(SYNM) generates an instance of the Import/Export
Manager (IEM), which then exports the results into
files for external analysis.
- Interface Edition. When the User requests to extend
an interface to improve system functionality, the Syn-
chronization Manager (SYNM) generates an instance
of the Interface Manager (INTM), which allows the
User to edit the selected interface. When the Synchro-
nization Manager (SYNM) detects that an instance is
no longer required it eliminates it.
3.2 Holon Identification
The refinement of the Organization Model is done
by following the ANEMONAs PROSA Guidelines
from 1 to 13 and 22 to 25 to define holons (abstract
agents). a) Specification of Holons: Both the defini-
tion of holons and the refinement of the Organization
Model are done by following the PROSA Guidelines
from 1 to 13 and 22 to 25. Figure 1 shows the main
interactions between the specified holons. Resource
Holons and their Roles: The Import/Export Holon
plays the Import/Export Manager (IEM) role; the
Analysis Holon plays the Input/Output Data Analyzer
(IODA) role; the Icon Holon plays the Icon Manager
(ICOM) role; the Modeller Holon plays the Modeller
(MOD) and Model Checker (CHECK) roles; the Sim-
ulation Holon plays the Simulation Planner (SIMP)
role; the Results Holon plays the Report Generator
(REGM) role; the Event Holon plays the Event Gen-
erator (EVEG) role; the Animation Holon plays the
Animation Manager (ANIM) role; the Speed Holon
plays the Speed Manager (SPEM) role; and the In-
terface Holon plays the Interface Manager (INTM)
role; Staff Holon and their Roles: Synchronization
Holon plays the Synchronization Manager (SYNM)
role. Product Holons: Model Holon, Report Holon,
Graph Holon, Event Holon, Icon Holon, and Inter-
face Holon; Work Order Holons: Importation Or-
der Holon, Model Order Holon, Report/Graph Order
Holon, Event Order Holon, Exportation Order Holon;
b) Refinement of Interactions: The specification of
new interactions is identified by using the PROSA
guidelines 26 to 28. PROSA guideline 25 is also used
to complete the definition of the new interactions. The
refinement of interactions includes the identification
of goals that the interactions pursue, the identification
of exchanged messages, the identification of source
and receiver of that messages, and the identification
of temporal constraints.
4 CONCLUSION
This work presents a new holonic architecture for
a multiagent-based simulation tool. The specifica-
tion of the Holonic Architecture of the Simulation
Tool has been made possible by the Analysis Phase
HOLONIC ARCHITECTURE FOR A MULTIAGENT-BASED SIMULATION TOOL
397
I
S
t
a
r
t
s
coordination
To Manage the Model
Creation
I
S
t
a
r
t
s
I
S
t
a
rt
s
cooperation
Model Simulation
Configuration
coordination
To Manage Analysis of
Results Simulation
I
S
t
a
r
t
s
cooperation
Analysis of Simulation
Results
I
S
t
a
r
t
s
collaboration
Model Simulation
collaboration
Model Creation
cooperation
Results Exportation
I
S
t
a
r
t
s
collaboration
Interface Edition
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
IC
ol
l
a
bo
r
at
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
ll
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
la
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
ol
l
a
b
or
a
t
es
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
S
t
a
r
t
s
I
S
t
a
r
t
s
I
C
o
l
l
a
b
o
ra
t
e
s
A
Modeller
Holon
A
Import/Export
Manager Holon
A
Icons
Holon
A
Interfaces
Holon
A
Speed
Holon
A
Animation
Holon
A
Events
Holon
A
Results
Holon
A
Synchronization
Holon
A
Simulation
Holon
A
Importation
Order Holon
A
Exportation
Order Holon
A
Model Order
Holon
A
Event Order
Holon
A
Report/Graph
Order Holon
A
Model
Holon
A
Interface
Holon
A
Icon
Holon
A
Event
Holon
A
Report
Holon
A
Graph
Holon
I
S
t
a
r
t
s
I
C
o
l
l
a
b
o
r
a
t
e
s
A
Import/Export
Manager Holon
A
Interfaces
Holon
A
Interface
Holon
A
Synchronization
Holon
A
Synchronization
Holon
A
Synchronization
Holon
A
Icons
Holon
A
Icon
Holon
A
Modeller
Holon
A
Icons
Holon
A
Interfaces
Holon
coordination
To Manage Data
Importation
A
Analysis
Holon
A
Icons
Holon
A
Synchronization
Holon
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
I
S
t
a
r
t
s
cooperation
Report/Graph
Generation
I
C
o
l
l
a
b
o
r
a
t
e
s
I
C
o
l
l
a
b
o
r
a
t
e
s
Figure 1: Holon Interactions.
of the ANEMONA Methodology (Giret et al., 2005).
The architecture takes advantages of the benefits of
two representative approaches, HMS and the MAS
to propose a more powerful tool than current simu-
lation tools. The Simulation Tool includes features
such as flexibility, scalability, reconfigurability, and
adaptability that can be observed during the Model
Simulation and also offers additional features such
as: a) Simple Architecture: Since a Holon can play
more than one role, it allows the system to reduce
the risk of having a great number of entities that are
not required all the time. Control can be distributed
thereby avoiding bottle necks which slow the system
performance down. b) Synchronization: Due to the
supervision of the Synchronization Holon, it is pos-
sible to detect needs that are related to the genera-
tion or elimination instances of a specific holon or
holons according to the current state of the system. c)
Modelling: The Modeller Holon provides flexibility
when the User is creating the Model avoiding possible
omissions(Checker role). d) Negotiation Processes:
To optimize the use of current resources holons per-
form negotiation processes during the whole simula-
tion cycle based on their goals, beliefs, and tasks. We
are currently working on the documentation of the de-
sign phase of the Simulation Tool, which is also based
on the ANEMONA methodology.
REFERENCES
Giret, A. and Botti, V. (2004). Holons and agents. Journal
of Intelligent Manufacturing, 15(5):645–659.
Giret, A. and Botti, V. (2006). From system requirements
to holonic manufacturing system analysis. Int. J. of
Production Research, 44(18-19):3917–3928.
Giret, A., Botti, V., and Valero, S. (2005). Mas methodology
for hms. In Verlang, S., editor, Proc. de 2nd Int. Conf.
on Applications of Holonic and MAS (HoloMAS05).
LNAI 3593, pages 39–49.
HMS (2004). Holonic manufacturing systems consortium
web site. http://hms.ifw.uni-hannover.de.
Klingstam, P. and Gullander, P. (1999). Overview simu-
lation tools for computer-aided product engineering.
Computers in Industry, 38:173–186.
Koestler, A. (1990). The Ghost in the Machine. Penguin
Group, reprint edition edition.
Law, A. M. and McComas, M. G. (1999). Simulations of
manufacturing systems. In Proc. of the 1999 Winter
Simulation Conference, pages 56–59.
Miller, S. and Pegden, D. (2000). Introduction to manufac-
turing simulation. In Proc. of the 2000 Winter Simu-
lation Conference, pages 63–66.
Nikoukaran, J. and Paul, R. J. (1999). Software selection
for simulation in manufacturing: a review. Simulation
Practice and Theory, 7(1):1–14.
Ruiz, N., Giret, A., and Botti, V. (2006). Towards an agent-
based simulation tool for manufacturing systems. In
Proc. de 11th IEEE Int. Conf. on Emerging Technolo-
gies and Factory Automation (ETFA06), pages 797–
804.
Wooldridge, M. and Jennings, N. (1995). Intelligent agents:
Theory and practice. The Knowledge Engineering Re-
view, 10(2):115–152.
ICEIS 2007 - International Conference on Enterprise Information Systems
398