Journal of Software Engineering and Applications, 2011, 4, 137-145
doi:10.4236/jsea.2011.43015 Published Online March 2011 (http://www.SciRP.org/journal/jsea)
137
Agent-Based Synthesis of Distributed Controllers
for Discrete Manufacturing Systems
Ernesto López-Mellado
CINVESTAV Unidad Guadalajara, Zapopan, Mexico.
Received February 4th, 2011; revised February 23rd, 2011; accepted February 25th, 2011.
ABSTRACT
A method for designing real-time distributed controllers of discrete manufacturing systems is presented. The approach
held is agen t based ; the controller strategy is distributed into several interacting agents that operate each one on a part
of the manufacturing process; these agents may be distributed into several interconnected processors. The proposed
method consists of a modelling methodology and software development framework that provides a generic agent archi-
tecture and communica tion facilities supporting the interaction among agents.
Keywords: Age nt -B ase d Sof t w are Desi g n, Distributed Knowledge Based Controllers, Dis c rete Manuf acturing Systems
1. Introduction
Nowadays discrete manufacturing systems are large and
complex sy stem s that integra te several kinds of devi ces of
miscellaneous nature and behaviour, namely robots, con-
veyors, machines, sensors, etc. Additionally, the produc-
tion require ments are often c hanged; t hese fa cts im pose to
the system components to be versatile, and to the coor-
dination system or controller to be highly flexible.
The core of a control system is complex software,
which determines at last, the flexibility and the perform-
ance of the automated system. This control software pro-
vides several functions: tasks execution, monitoring, de-
cision making, and planning; it is generally distributed
into a four layer hierarchy [1]; in this scheme the control
function is decomposed into four levels in which the re-
sponse time is shortest in the lower levels.
The lowest level includes the local controllers of the
physical devices in the cell (robots, conveyors, machines,
sensors, etc. The task coordination level or cell level
manages and supervises the activities of the local control-
lers involved in a cell by the gen eration of pertinent com-
mands according to events issued from the local control
level. The task planner generates the strategy of the con-
troller for the cells according to the s pecifications from the
productio n pl anning level.
Due to the complexity of the tasks, especially which
performed at the cell level, the synthesis of the con troller
is a difficult job often addressed through planning tech-
niques. One of the problems found in the synthesis real-
time controllers is that the representation of the qualitative
controller often includes a large amount of knowledge
whose processing is time consuming.
This work deals with the task coordinator level. The
functions of this control level can be decomposed mainly
in a) the sequencing of operations to accomplish the as-
sembly task in a normal functioning regime, and b) the
handling of exceptions representing operation failures [2].
In this paper the case of normal execution is addressed.
The design and implementation of tasks controllers of
complex manufacturing systems has been addressed by
using several approaches. The object oriented approach
has been held for modelling [3], simulation [4-6], and con-
trol [7] of manufacturing systems. The agent based ap-
proach [8] has been adopted to address some problems in
manufacturing systems [9]; D e Loach [10] p roposes a mo-
delling lan guage for describing the dive rse kinds of agents,
and defines a m ethod ology ( MaSE) for t he form al sy nthe-
sis of agent systems; in [11] Bussmann focuses on deci-
sion making issues during the planning stage, and in [12]
he addresses t he tas k program m ing issue pro posin g a syn-
thesis method that leads to concurrent centralised control
software; in [ 13] Ouelha dj propose d a dynam ic control ar-
chitecture for m anufacturing s yst em s organi s ed into cell s,
but the programming of the agents is not reported.
In this work we also profit of the agent-based approach
for conceiving a control software as a composition of in-
teracting modules, defined as reactive agents, which may

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
138
be distrib uted i nto seve ral p rocessors; it i s propose d a m e-
thod that supports the complete development life cycle of
distributed controllers of discrete manufacturing systems;
this lifecycle is shown in Figure 1, in which the stages of
the method are pictorially overviewed. The proposed me-
thod for sequencing the activities of the cell components
allows buildi ng rapidly p rototypes of so ftware controll ers.
The remainder of this paper is organised as follows:
Section 2 describes the proposed methodology for mod-
elling both the manufacturing system and the tasks to be
executed; Section 3 presents the proposed method for the
design of distributed control software: first the decompo-
sition of the task model is described, then agent based
solution to the synthesis of distributed software is out-
lined.
2. Controller Modelling
This section presents the methodology that helps to ob-
tain systematically the contents of the knowledge bases
from a model of the manufacturing/assembly system; this
model, close to that presented in [7], includes the system
description and the tasks specification. The aim of this
stage is to obtain in a structu red way the system functio n-
ing and production requi rements.
The description consists in a component classification
of the m anufacturi ng proce ss, and a str ucturing of the sys-
tem workspace.
The component classification leads to a taxonomy of
the system components organised as a hiera rchy includi ng
capabilities and features of each component, and the total
quantity of devices. The hierarchy is useful to program the
necessary classes in the control software. Figure 2 shows
an example of components hierarchy.
2.1. System Description
The structuring of the system workspace consist of a defi-
nition of releva nt physical emplacem ents where operations
are perform ed on the work pieces or parts; rather t han spa-
ce partition into regions, the structuring is a discrete assi-
gnment of positions. The key element for structuring the
workspace is the notion of site, which is defined as the
place where parts can be tem pora ry held or st ored in a sta-
ble position (a table, a magazine, a robot gr ipper, ...) [2].
The sites may be single or composed (macro-site); sin-
gle sites held one part or subassembly; composed sites
have two or more emplacements which manage the infor-
mation attached to a set of sites closely located and func-
tionally equivalents. The sites that are associated to effec-
tors are named active sites; otherwise they are called pas-
sive sites. A site co ntains inf ormation a bout the w ork zone
were it is emplaced that can be used as mutual exclusion
resource (for robot collision avoidance, for example).
2.2. Task Specification
The flow of material is described by a flow of parts graph
(FPG), and t hen a set of dis patching rules, which represent
the controller strategy, is obtained.
2.2.1. The Flow of Parts Graph
A FPG is a directed graph whose nodes are all the sites of
the system; the arcs joining the nodes represent eith er the
operations needed to transfer the parts from one site to
another one or to modify the properties of the part held
into a site. For example, Figure 3 describes the operations
pick and place performed by a robot (R1); three sites are
involved: two passive sites (CONV2 and TAB1), and an
active site (GRIP1) associated to the robot gripper. The
definition of FPG is given below:
Definition. A flow of parts graph is the tuple F = (G,
SITES, OPER , , , ), where
 G is a c onnected dir ected graph G = (V, A), wher e
—V is a finite set of vertex,
—A  V  V is a set of edges or arcs.
 SITES = {site1, ···, siten} is a finite set of site names,
which are not input or output sites.
  = {sitein1, ···, siteinp, siteout1, ···, siteoutq} is a fini-
te set of site names labelling sites where the parts
entry or leave the FMS.
 : V  SITES  {} is a labelling function that
assigns name sites to the vertex of G.
 : A  OPER is a labelling function that assigns
operations names to the arcs of G.
2.2.2. Sequencing the Ope rations
The dispatching rules are antece dent-consequent rules that
state the conditions in which an operation must be exe-
cuted; they are obtained directly from the FPG, and the
number of rul es i s t he sam e t han the number of a rcs in t he
FPG. The antecedent part is compos ed by conditions that
involve mainly sensory conditions and tests (contents, part
posture, …) on the sites related by the operation; other
conditions may involve tests on sites located upstream the
FPG. The consequent part includes the request of execu-
tion of the associated operation and the updating of the
involved sites.
2.3. A Modelling Example
For illustrating purposes we include an example regarding
a simple assembly system; it will be addressed through the
rest of the paper.
2.3.1. Sys tem and Task Description
1) The system: Consider the assembly cell sketched in
Figure 4; it consist of three conveyor belts B1, B2, and
B3, two robots R1, R2, two assembly tables A1, A2, an
storing table ST. Each assembly table has two positions;

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
139
Figure 1. Software development life cycle.

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
140
Figure 2. Taxonomy of the assembly cell.
CONV2 GRIP1 TAB1
Pick.R1 Place.R1
Figure 3. Flow of parts graph.

B1 B2
B3
R1
R2
ST A1
A2
C1
Figure 4. Assembly cell layout.
the storing table has four positions. The parts to be as-
sembled arrive through the conveyor B1; two kinds of as-
sembled products the leave the cell through B2 and B3.
In the front of R1, an optical sensor C1 detects the arrival
of parts; over this zone a camera of a location-recogni-
tion system is emplaced.
2) The task: Eight types of parts (A, B, C, D, E, F, G,
and H) constitute the input flow in B1; they arrive at ran-
dom order. R1 gets the parts and bui lds asse mblies (stacks )
on A1 (with the parts A, B, C, and D) or A2 (with the parts
E, F, G, and H) according to a predefined order for each
product. The detection of parts in C1 stops B1; the identity
of the parts is determined b y the v ision system when th ey
arrive at the position C1. R1 gets part only if it can be as-
sembled or temporary stored in STi. Otherwise the part is
left in B1. R2 gets completed assemblies from A1 or A2
and places them on B2 and B3 respectively.
2.3.2. System and Task Modelling
1) System taxonomy. The components of the assembly
system are classed from a functional point of view: sen-
sors, effectors, etc. This classification is useful to struc-
ture the factual knowledge of the assembly system: task
state, component capabilities and relationships, etc. Fig-
ure 2 shows the hierarchy concerning the assembly sys-
tem of the example; the items in the lowest level of the
hierarchy can be object instances of the upper concept
(class).
2) Workspace modelling. In the example the following
sites are defined: CONV1, CONV2 and CONV3 are the
sites associated to the place where the parts stops in front
of the robot ; G RIP1 an d GR IP2 a re associat ed to the grip-
pers of R1 and R2 res pect i vely. The sto ri ng table has four
sites: STi ( i = 1, · ·· , 4); the y can be manage d b y the macro
site ST. The sites associated to assembly tables are ASSB1
and ASSB2.
3) Flow of parts. The FPG shown in Figure 5 describe s
the flow of parts required in the assembly task; the sites
defined in the model are related by the operations whose
outcome is, mainly, the tran sferring the parts between the
sites. Operations may only modify the properties of the
part into a site; as an example notice that the operation
Ident-Loc does not transfer the part to another site but it
changes the attributes of the unknown part.
Operations may also put parts into the flow model or
drawn out parts (or products) from the model. The FPG
for this example is defined as follows:
F = (G, SITES, OPER, , ) where G = (V, A) with
V = {v0, v1, v2, v3, v4, v5, v6, v7, v8, v9, v10, v11}
A = {(v0, v1), (v1, v2), (v1, v1), (v1, v3), (v3, v4), (v4, v3),
(v3, v5), (v3, v6), (v5, v7), (v6, v7), (v7, v8), (v7, v9), (v8, v10),
(v8, v11)}
SITES = {conv1, conv2, conv3, grip1, grip2, assb1,
assb2, st1}
 = {input1, out put 1 , output2, output3}
OPER = {B1. feed(), B 1.a dvance(), R 1.pick(conv1), R1.
store(st), R1.recover(st), R1.assembly(assb1), R1. assem-
bly(assb2), R2.pick(assb1), R2.pick(assb2), R2.place-

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems 141
Figure 5. Flow of parts graphs for the assembly cell.
(conv2), R2.place2(conv3), B2.advance(), B3.advance(),
C1.ident&Loc()}
 = {(v0, input1), (v 1, conv1), ( v2, ou tput1), ( v3, grip1),
(v4, st1), (v5, assb1), (v6, assb2), (v7, grip2), (v8, conv2),
(v9, conv3), (v10 , output2), (v11, output3)}
 = { ((v0, v1), B1.feed()), ((v1, v2), B1.advance()), ((v1,
v3), R1.pick(conv1)), ((v3, v4), R1.store(st)), ((v4, v3),
R1.retreive(st)), ((v3, v5), R1.assembly(assb1)), ((v3, v6),
R1.assembly(assb2)), ((v5, v7), R2.pick(assb1),) , ((v6, v 7),
R2.pick(assb2)), ((v7, v8), R2.place(conv2)), ((v7, v9),
R2.place(conv3)), ((v8, v10), B2.advance()), ((v8, v11),
B3.advance()), ((v1,v1), C1.ident&Loc())}
3. Distributed Software Design
This section deals with the design of the distributed soft-
ware that implements the task controller of a manufactur-
ing system. First the modularisation of the task model is
presented, and then the resultin g partition is taken for im-
plementing the agents [14].
3.1. Task Model Partition
The task m odel must be decom posed into su btasks in s uch
manner that every subtask may be assigned to an agent;
this decomposition is ach ieved by a partition of the FPG.
Several strategies for obtaining sub-graphs from the FPG
may be adopted according to the number of processors, the
geographical distribution on the components, or the simi-
larity of the sub-graphs.
In this work the strategy held for decomposing the gra-
ph is creating the maximal number of sub-graphs; each
sub-graph must in volve one activ e site. This approach al-
lows defining agents capable to control one effector. A
three-step algorithm is described below.
Algorithm. Partitioning the task model.
1) Identify active and passive sites. Let Act  SITES,
the set of active sites and Pass  SITES the set of passive
sites.
2) Sub-model creation. For every sAct, create a sub-
graph gk including s and its predecessors and successors.
In gk it is included an active site and several passive sites.
SG = {g1, g2, ···, gr} is the set of sub-models.
3) Simplification of sub-models. The operations asso-
ciated to the arcs of a graph gk must be executed by the
effectors or sensors associated to the active site of gk. Thus
the arcs labelled with other operations must be withdrawn
from gk; consequently isolated vertex must be eliminated
too.
The sites belonging to two or more gk are called inter-
face sites; they are in the boundary of the graph and they
must be car efully m anaged be cause they are considere d as
shared resources.
Example. Consider the FPG of Figure 5; defining Act =
{conv1, conv2, conv3, grip1, grip2}, and Pas = {asb1,
assb2, st}, the decomposition procedure yields the sub-
graphs depicte d in Figure 6; so SG = {g1, g2, g3, g4, g5}.
3.2. Task Programming Framework
Once the task model is decomposed, each sub-model is
used for defining an agent. Since every sub-model invol-
ves an actuator, it must be controlled by the corresponding
agent avoiding situations in which two or more agents
handle the same effector. The knowledge base of every
agent corresponds to the set of rules associated to the arcs
of the pertaining subtask.
The developed platform supports the programming of
the agents that implement th e subtask s of the syste m; this
platform is based on JADE V1.2 (Java Agent Develop-
ment fram ework) [15], w hich meets the standards of FIPA
[16].
3.2.1. Requirements
Agent requirements. Each agent
 has a unique name
 manages the correspondi ng su btask; i t initi alises and
updates the contents of its sites and macro-sites.
 exchanges messages with other agents; it in terprets
the message and perform the requests such as pro-
vide information about the contents of a site or about
the execution of an operatio n.
 coordinates with other agents for allocating shared
resources.

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
142
Figure 6. Graph decomposition.
 manages the associated devices (effectors and sen-
sors involved in the subtask).
Sites requirements. Each site or macro-site
 has a unique name.
 provides facilities for managing its state, i.e. the ini-
tialising and updating of the associated attributes
such as contents, position, sensor values, trace of the
operations performed on it, etc.
 provides facilities for managing special features,
such as flags for mutual exclusion, the agent names
that share it, or the assembly patterns if it is a site for
assembly.
Devices requirements. For every device (effector or
sensor) in the system, one must create an interface mod-
ule that allows consulting the device state and handling
the messages representing actions requests or responses.
Every module has a unique na me.
3.2.2. Definition of Cl asses
Component architecture. The implemented components
are integrated into a package organised in four sub-pack-
ages; it is shown in Figure 7. The sub- pack age MApack -
age.sites contains the classes Site and Part, MApack-
age.macrosites contains the class Macrosite and the class-
es related with the access of sites contained in the macro-
site. Three kinds of sites are considered: interface sites
(shared by two or more sub-tasks), internal sites, and re-
mote sites (non shared but belonging to other sub-tasks).
Figure 8 shows a detail of the above packages.
The sub-package MApackage.devices contains the class
Device and other sub-packages. MApackage.devices.
effectors and MApackage.devices.sensors contain sub-
classes illustrated in Figure 9. This organisation is strongly
suggested by the taxonomy of the manufacturing system
obtained during the modelling stage.
Agent class. The class AgentBas e is an abstract class
that specialises the class Agent from JADE. This class em -
Figure 7. Component organisation.

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
143
Figure 8. Relationships among sub-packages.
sensorseffectors
devices
Device
Effector Sensor
CameraRobot Conveyor Detector
Figure 9. Devices sub-package.
beds the behav iour of a generi c agent that m anages all the
activities related to a subtask. Every agent must be pro-
grammed by extending AgentBase, declaring the knowl-
edge of the corresponding subtask, and instancing the ex-
tended subclass. Th e generic agent provides the facilities
for the ma nagement of si t e s and the ha n dl i n g o f messages
related with the manufacturing task and messages for in-
teracting with other agents; interaction among agents is
performed through four kinds of messages regarding in-
formation requests and sending about sites, and request/
confirmation on the use of sites shared by two or more
agents.
Agent prog ramm ing . The programming of each agent
is done based in the information obtained from the cor-
responding subtask graph.
3.2.3. Comp onent Identific ation
The first step for programming an ag ent is to identify the
rules, the si tes, the acti ons, a n d de vices rega rdi ng t he su b-
task. For every site in the graph one must declare which
agent shares this site, and define if it is a remote site. If a
site is used for assembly declare the assembly pattern to
follow during the execution of the task. When an in terface
site is associated to sensor, only one agent must manage
Operation Ident-Loc
Description If during the feeding a part is detected, the con-
veyor 1 stops and the part is identified. The sen-
sor value is reset.
Related sites Conv1
Related devicesConveyor1, Vision system, Detector.
Pre-conditions Conv1 is empty, the state of conveyor is FEED- ING, the presence sensor is ON, and the subtask
state is INITIAL.
Actions Stop Conveyor1. Identify and locate the part with
the vision system. Reset sensor.
Post-conditions Conv1 holds a part, the state of Conveyor1 is
STOP, the vision system is IDLE, and the sensor
is OFF. The state of the task is HOLDINGP ART.
Figure 10. Frame for a rule description.
such sensor.
3.2.4. Buil di ng the Rul e B as e
The rules define the strat egy of the controller. Every agent
has a small set of rules corresponding to the arcs of the
sub-graph. Befo re the writing of the ru les it is convenient
to enumerate all the info rmation regarding each ru le. This
may be systematised by the filling of frames, such as
shown in Figure 10 for the operation Ident-Loc of the
sub-task 1.
The Java coded rules for the a gent coordinating the sub-
task 3 is given below.
// implementing rules for the agent named
// ManufacturingAgent3
void rules()
{ // declaring reference to site for place //the
part
Site destinationPart = null;
// GRIP2.TAKE1: Grip2 takes the assembled //part
from ASSB3
if(assb3.getSite(0).isAssembled() &&
grip2.isEmpty() &&
conv2.isEmpty() && destinationPart==null)
{
grip2.setContent(assb3.getContent(0));
assb3.removeContent(0);
sendUpdateMessages(grip2);
sendUpdateMessages(assb3);
//
destinationPart = conv2;
}
// GRIP2.TAKE2: Grip2 takes the assembled part
// from ASSB4
if(assb4.getSite(0).isAssembled()&&
grip2.isEmpty() &&
conv3.isEmpty() && destinationPart==null)

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
144
{
grip2.setContent(assb4.getContent(0));
assb4.removeContent(0);
sendUpdateMessages(grip2);
sendUpdateMessages(assb4);
destinationPart = conv3;
}
// GRIP2.TAKE: Grip2 places the took part in
// CONVin takes the assembled part from ASSB4
if(!grip2.isEmpty() && destination-
Part.isEmpty() && destinationPart!=null)
{
destination-
Part.setContent(grip2.getContent();
grip2.removeContent();
sendUpdateMessages(destinationPart);
sendUpdateMessages(grip2);
destinationPart = null;
}
}
3.3. Implementation Issues
This method has been demonstrated through the software
implementation of several case studies. The software has
been written in Java using the JADE framework for sup-
porting the agent definition and task interaction.
The distribut ed software has bee n tested on several per -
sonal computers interconnected through a local area net-
work; every agent was assigned to single PC.
For executing the controller, first the JADE framework
is initiated in a computer where an agent may be executed,
and then the rest of the agents are started in their compu-
ters.
For every agent the interaction with a device has been
simulated through a visual interface on screen; during a
test, users simulate the devices response to commands sent
by the controller through the keyboard.
4. Conclusions
In this work a method to develop distributed software for
manufacturing control systems has been presented. One
important issue of the proposed method is the modelling
stage; the specifications are transformed into graphical
models that contribute to reduce the problems due to am-
biguities and incompleteness. Th e partitioned task model
leads to obtain systematically the knowledge of every
agent in the control software, allowing modifying easily
the strategies of the subtasks; this feature is useful when
task reprogramming is needed.
The programming methodology takes advantage of the
JADE framework facilities, which permit to the defined
agents to be executed into different kinds of platforms.
The case studies have been tested by simulating the
manufacturing system devices through the console com-
puters; however a real application could be developed by
the implementation of the interface modules that com-
municate the controller with the actual devices.
REFERENCES
[1] S. B. Gershwin, “Hierarchical Flow Control: A Frame-
work for Scheduling and Planning Discrete Events in
Manufacturing Systems,” IEEE Proceedings of S. I. on
Discrete Event Systems , Vol. 77, No. 1, January 1989, pp.
195-209.
[2] E. López and R. Alami, “A Failure Recovery Scheme for
Assembly Workcells,” IEEE International Conference on
Robotics and Automation, Cincinnati, May 1990, pp
702-707.
[3] E. Arjona-Suárez and E. López-Mellado, “A Computer
Language for the Modelling of Flexible Manufacturing
Systems,” Proceedings of the 13th IASTED International
Symposium on Robotics and Manufacturing Santa Bar-
bara, California, November 1990, pp. 183-187.
[4] C. R. Glassey and S. Adiga, “Conceptual Design of a
Software Object Library for Simulation of Semiconductor
Manufacturing Systems,” Journal of Object Oriented
Programing, Vol. 2, No. 4, 1989, pp. 39-43.
[5] M. Bakalem, G. Habchi and A. Courtois, “PPS: An Inte-
grated Object Oriented Approach for Modelling and
Simulation of Manufacturing Systems,” IEEE Interna-
tional Conference on Systems, Man and Cybernetics.
Texas, 2-5 October 1994, pp. 2184-2189.
[6] G. Ramzi, M. Baka le m an d G. Habchi, “An Object Model
for Simulation of Manufacturing Systems,” IEEE Inter-
national Conference on Systems, Man and Cybernetics,
Vancouver, 22-25 October 1995, pp. 137-142.
[7] E. López and E. Medrano, “Object-Based Design of FMS
Controllers,” Proceedings of IASTED International Con-
ference on Robotics and Manufacturing, Cancun, May
1997. pp. 342-345.
[8] N. R. Jennings, K. Sycara and M. Wooldridge, “A Road-
map of Agent Research and Development,” International
Journal of Autonomous Agents and Multi-Agent Systems,
Vol. 1, No. 1, 1998, pp. 7-38.
doi:10.1023/A:1010090405266
[9] W. Shen and D. H. Norrie, “Agent-Based Systems for
Intelligent Manufacturing: A State-of-the-Art Survey,”
Knowledge and Information Systems, Vol. 1, No. 2, 1999,
pp. 129-156.
[10] S. A. DeLoach, “Multiagent Systems Engineering: A Me-
thodology and Language for Designing Agent Systems,”
Agent-Oriented Information Systems’99, Seattle, 1 May
1998, pp. 45-57.
[11] S. Bussmann, H. Baumgärtel and M. Klosterberg, “Multi-
Agent Coordination of Material Flow in a Car Plant,”
Second International Conference on Practical Applica-
tions of Intelligent Agents and Multi-Agent Technology,
London, 1997, pp. 227-236.

Agent-Based Synthesis of Distributed Controllers for Discrete Manufacturing Systems
145
[12] S. Bussmann, “Agent-Oriented Programming of Manu-
facturing Control Tasks,” Proceedings of the 3rd Interna-
tional Conference on Multi-Agent Systems, Paris, 1998,
pp. 57-63.
[13] D. Ouelhadj, C. Hanachi and B. Bouzouia, “Multi-Agent
System for Dynamic Scheduling and Control in Manu-
facturing Cells,” Proceedings of the IEEE International
Conference on Robotics and Automation, Leuven, 16-20
May 1998, pp. 2128-2133.
[14] J. G. Morales-Montelongo, “Agent-Based Distributed Co-
ordination of Manufacturing Systems,” MSC Thesis,
Cinvestav Unidad Guadalajara, Mexico, November 2002.
[15] F. Bellifemine, A. Poggi and G. Rimassa, “JADE—A
FIPA—Compliant Agent Framework,” Proceedings of
International Conference on Practical Applications of
Intelligent Agents and Multi-Agent Technology, London,
April 1999, pp. 97-108.
[16] “Foundation for Intelligent Physical Agents,” Specifica-
tions, 2005. http://www.fipa.org