Service Orientation in Holonic and Multi-Agent Manufacturing

336 Pages·2016·9.53 MB·English

by  Theodor Borangiu

Studies in Computational Intelligence 640
Theodor Borangiu
Damien Trentesaux
André Thomas
Duncan McFarlane Editors
Service Orientation
in Holonic and
Multi-Agent
Manufacturing
Studies in Computational Intelligence
Volume 640
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [email protected]
About this Series
The series “Studies in Computational Intelligence” (SCI) publishes new develop-
mentsandadvancesinthevariousareasofcomputationalintelligence—quicklyand
with a high quality. The intent is to cover the theory, applications, and design
methods of computational intelligence, as embedded in the fields of engineering,
computer science, physics and life sciences, as well as the methodologies behind
them. The series contains monographs, lecture notes and edited volumes in
computational intelligence spanning the areas of neural networks, connectionist
systems, genetic algorithms, evolutionary computation, artificial intelligence,
cellular automata, self-organizing systems, soft computing, fuzzy systems, and
hybrid intelligent systems. Of particular value to both the contributors and the
readership are the short publication timeframe and the worldwide distribution,
which enable both wide and rapid dissemination of research output.
More information about this series at http://www.springer.com/series/7092
Theodor Borangiu Damien Trentesaux
(cid:129)
é
Andr Thomas Duncan McFarlane
(cid:129)
Editors
Service Orientation
in Holonic and Multi-Agent
Manufacturing
123
Editors
Theodor Borangiu AndréThomas
FacultyofAutomaticControlandComputer University of Lorraine
Science Épinal
University Politehnica of Bucharest France
Bucharest
Romania Duncan McFarlane
Institute for Manufacturing Engineering
Damien Trentesaux Department
University of Valenciennes and CambridgeUniversity
Hainaout-Cambresis Cambridge
Valenciennes UK
France
ISSN 1860-949X ISSN 1860-9503 (electronic)
Studies in Computational Intelligence
ISBN978-3-319-30335-2 ISBN978-3-319-30337-6 (eBook)
DOI 10.1007/978-3-319-30337-6
LibraryofCongressControlNumber:2016933325
©SpringerInternationalPublishingSwitzerland2016
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
foranyerrorsoromissionsthatmayhavebeenmade.
Printedonacid-freepaper
ThisSpringerimprintispublishedbySpringerNature
TheregisteredcompanyisSpringerInternationalPublishingAGSwitzerland
Foreword
We are living today in an extraordinary period of growing complexity of manu-
facturing systems. This is, as a matter offact, the result of technological develop-
ment in several areas:
(cid:129) Computing systems are becoming more powerful every day. This is a conse-
quence of Moore’s law which is still valid: the key parameters of HW equip-
ment is doubled (processing speed and memory capacity) or reduced to half
(energyconsumption)every18months.TheSWarchitecturesandtoolsarealso
evolving at a rapid pace and succession.
(cid:129) Communication systems are becoming more broadband, faster, more efficient,
and are built to support the idea of cyber-physical systems.
(cid:129) Automation methodologies are becoming more intelligent, their development
being oriented towards decentralized intelligent systems organized as a com-
munityofautonomousunitswithoutanycentralelement.AIprinciplesarebeing
massively deployed and exploited.
Especially the field of distributed intelligent systems has influenced manufac-
turing and production management quite strongly. Let us summarize briefly the
shorthistoryofthisdevelopmenttobetterunderstandthestateoftheartandtrends
inabroadercontext;thiswillalsounderlinetheimportanceofscientificeventssuch
as SOHOMA.
Thefirstideasofdistributedintelligentsystemsappearedinconnectionwiththe
holonic visions 25 years ago. The very first pilot implementations led to the
development of the first holons—in principle reactive agents—and to introduction
of the first standard in the field (IEC 16499). Once engineers intended to add
intelligencetothedistributedelementswestartedtotalkaboutagents,agent-based
systems and multi-agent systems (MAS). Specific platforms to run multi-agent
systems with specialized functionalities like yellow pages, white pages, brokers,
sniffers, etc. evolved being supported by the FIPA Association; then, FIPA com-
munication and architecture standards appeared. FIPA efforts are rather frozen at
v
vi Foreword
the moment, but the FIPA standards accepted 10–12 years ago are still in use and
accepted in the field of MAS.
The MAS philosophy applied to industrial control allowed thinking about and
conceiving new approaches and solutions. Progressively, products and
semi-products started to be represented by SW agents that were able to commu-
nicate, negotiate and coordinate their activities, not only in manufacturing and
transport processes. The products became active elements during their execution
life cycle. The PROSA-like way of thinking strongly influenced the field, causing
that not only semi-products but also humans were considered as resources repre-
sented by agents. This might be considered as a significant technology break-
through in the field of decentralized control and production management.
But to make distributed solutions increasingly more intelligent, higher level
agents required deployment of more and more knowledge. This is why semantics
has been introduced and ontology knowledge structures shared by agents became
anobviousvehicletoreducethecommunicationtrafficandtomaketheagentsmore
intelligent.Insomecases,theontologyconvergedtotheWWWtechnology(orwas
combined with it). The direct communication and interaction with and among the
devices (not only with their SW modules as virtual representation) became nec-
essarytogetfasteraccesstophysicaldevices,tothephysicalworld.TheInternetof
Things appeared.
Thisdevelopment ledtothe new visionof theFactoryof theFuture formulated
in the German governmental initiative Industry 4.0, 2013. This vision is nothing
else than the extension of the trends in the field of distributed intelligent control
combined with new business models supported by accelerated development in the
domain of computing and communication. Industry 4.0 is based on the following
principles:
(cid:129) Integration of both the physical and virtual worlds using the Internet of Things
and Internet of Services.
(cid:129) VerticalIntegrationalongtheenterpriseaxis,whichmeansintegrationofallthe
information and knowledge-based systems in a company, starting with the
real-time control level of shop floor up to the ERP and managerial systems on
the top.
(cid:129) HorizontalIntegrationalongthevaluechainaxis,whichmeansintegrationofall
business activities starting from the supply chain on and up to the product
delivery phase (from suppliers to customers).
(cid:129) Engineering Activities Integration along the life cycle axis from rough idea via
design,development,verification,productionandtestinguptoproduct-lifecycle
management (from design to support).
The visions of the three integration axes are based on the following MAS
principles: cooperation of distributed autonomous units, ontology knowledge
sharing and big data analytics. Industry 4.0 solutions are more and more linked or
even coupled with the higher level information systems of the company. Their
implementations are influenced by the latest trends in SW engineering exploring
service-orientedarchitectures(SOA).TheMAStechnologyremainstorepresentan
Foreword vii
excellent and promising theoretical background for developing an Industry 4.0
solution. The MAS theory can be used to support research activities, to bring new
features to these solution explorations, e.g. AI principles, machine learning, data
mining and data analytics in general. But the implementations do explore—as a
rule—the SOA approaches in broader and broader scale. These are critically sim-
plifying real-life solutions.
This volume of the SOHOMA’15 contributions displays the current trends in
intelligent manufacturing, namely the duality of MAS and SOA approaches. It
confirms that additional techniques like ontology knowledge structures, machine
learning, etc. represent very important and promising topics for further research;
they are expected to enrich the current solutions and bring Industry 4.0 visions to
industrial practice. In addition, this volume documents many successful research
results in this direction.
December 2015 Vladimír Mařík
Preface
This volume gathers the peer reviewed papers which were presented at the fifth
edition of the International Workshop “Service Orientation in Holonic and
Multi-agent Manufacturing—SOHOMA’15” organized on 5–6 November 2015 by
the Institute for Manufacturing (IfM) of the University of Cambridge, UK in col-
laboration with the CIMR Research Centre in Computer Integrated Manufacturing
and Robotics of the University Politehnica of Bucharest, Romania, the LAMIH
Laboratoryof IndustrialandHumanAutomation Control,Mechanical Engineering
and Computer Science of the University of Valenciennes and Hainaut-Cambrésis,
France and the CRAN Research Centre for Automatic Control, Nancy of the
University of Lorraine, France.
SOHOMA scientific events have been organized since 2011 in the framework
of the European project ERRIC, managed by faculty of Automatic Control and
Computer Science within the University Politehnica of Bucharest.
The book is structured in seven parts, each one grouping a number of chapters
describingresearchinactualdomainsofthedigitaltransformationinmanufacturing
and trends in future manufacturing control: Part I: Applications of Intelligent
Products, Part II: Recent Advances in Control of Physical Internet and
InterconnectedLogistics,PartIII:SustainabilityIssuesinIntelligentManufacturing
Systems, Part IV: Holonic and Multi-Agent System Design for Industry and
Services, Part V: Service Oriented Enterprise Management and Control, Part VI:
CloudandComputing-OrientedManufacturing,PartVII:SmartGridsandWireless
Sensor Networks.
These seven evolution lines have in common concepts, methodologies and
implementing solutions for the Digital Transformation of Manufacturing (DTM).
The Digital Transformation of Manufacturing is the actual vision and initiative
about developing the overall architecture and core technologies to establish a
comprehensive, Internet-scale platform for networked production that will encap-
sulatetherightabstractionstolinkeffectivelyandscalablythevariousstakeholders
(product firms, manufacturing plants, material and component providers,
ix
x Preface
technology and key services providers) to enable the emergence of a feasible and
sustainable Internet economy for industrial production.
For the manufacturing domain, the digital transformation is based on the
following:
1. Instrumenting manufacturing resources (machines, robots, AGVs, ASRSs,
products carriers, buffers, a.o.) and environment (workplaces, material flow,
tooling,a.o.)whichallows:producttraceability,productiontracking,evaluation
of resources’ status and quality of services, preventive maintenance…
2. Interconnecting orders, products/components/materials, resources in a
service-orientedapproachusingmultiplecommunicationtechnologies:wireless,
broadband Internet, mobile applications.
3. Intelligent, distributed control of production by:
(cid:129) New controls based on ICT convergence in automation, robotics, vision,
multi-agentcontrol,holonicorganization;thenewcontrolsenablethesmart
factory.
(cid:129) New operations based on product- and process modelling and simulation.
Ontologies are used as a “common vocabulary” to provide semantic
descriptions/abstract models of the manufacturing domain: core ontology—
modellingofassemblyprocesses(resources,jobs,dependencies,a.o.);scene
ontology—modelling flow of products; events ontology—modelling various
expected/unexpected events and disruptions; these models and knowledge
representation enable the digital factory.
(cid:129) Novel management of complex manufacturing value chains (production,
supply, sales, delivery, etc.) for networked, virtual factories: (a) across
manufacturing sites: logistics, material flows; (b) across the product life
cycle.
Research in the domain of DTM is determined by the last decades’ trend in the
goods market towards highly customized products and shorter product life cycles.
Such trend is expected to rise in the near future, forcing thus companies to an
exhaustive search for achieving responsiveness, flexibility, reduction of costs and
increased productivity in their production systems, in order to stay competitive in
such new and constantly changing environment. In addition, there is a shift from
pure goods dominant logic to service dominant logic which led to service orien-
tation in manufacturing and orienting the design, execution and utilization of the
physicalproductasvehiclefordeliveringgenericorspecificservicesrelatedtothat
product (in “Product-Service Systems”).
How this new vision on digital transformation of manufacturing is achieved?
Reaching the above objectives require solutions providing:
(cid:129) Dynamic reconfigurability of production (re-assigning resource teams,
re-planning batches, rescheduling processes) to allow “agile business” in
manufacturing;
(cid:129) Robustness at technical disturbances;