KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Copyright c© 2011 Department of Production EngineeringAll rights reserved. No part of this document may be reproduced, distributed, or trans-mitted in any form or by any means, including printing, photocopying, recording, or otherelectronic or mechanical methods, without the prior consent, written or spoken, from thecreators of the document. For permission request, contact the responsible person at theaddresses above.

KIPPcolla - Knowledge Intensive Product and Production Management From Conceptto Re-cycle in Virtual Collaborative Environment

Tekes Digital Product-Process Programme

D1.3 Requirements for holons: Descriptions of Holonicentities and interoperability requirements

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Contents

1 Terms and Definitions 2

2 Introduction - Holons in Literature 3

3 Holonic Manufacturing Systems: DiMS 6

3.1 Prosa Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3.2 DiMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.3 Extension of DiMS into EDMS . . . . . . . . . . . . . . . . . . . . . . . . 9

4 Requirements for Holonic Manufacturing System 11

4.1 Requirements for System Description . . . . . . . . . . . . . . . . . . . . . 12

4.2 Requirements for Adaptiveness . . . . . . . . . . . . . . . . . . . . . . . . 13

4.3 Requirements for Changeablity . . . . . . . . . . . . . . . . . . . . . . . . 15

5 Solution Candidate for HolonicManufacturing System 17

5.1 Knowledge-based manufacturing processes . . . . . . . . . . . . . . . . . . 17

5.2 SOA and Holons in technical sense . . . . . . . . . . . . . . . . . . . . . . 19

5.3 Technical Holon description . . . . . . . . . . . . . . . . . . . . . . . . . . 21

6 Conclusions 25

7 References 26

1/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Appendix Contents

2/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

1 Terms and DefinitionsAdaptive system: A system must continuously adapt itself to the changing nature of itsenvironment or it will become extinct. Similarly a manufacturing system must change con-tinuously to meet changing requirements of the environment. This adaptation is possibleonly through learning. Learning translates into adaptiveness. An adaptive organizationwould be able to survive under an unfavorably changing business environment.

BMS: Bionic Manufacturing System

CAD: Computer Aided Design

CAM: Computer Aided Manufacturing

CAPP: Computer Aided Process Planning

CAX: Computer Aided X

Complex system: Open Complex system is defined as a system that has a large numberof elements interacting nonlinearly and that it gets inputs from surrounding environment

Data: single values, filtered values, average values

DiMS: Distributed Manufacturing System

EDMS: Extended Digital Manufacturing System

FF: Fractal Factory

HMS: Holonic Manufacturing System

Holonic Framework: Holonic Framework is the ICT implementation for message bro-kering and knowledge management.

Information: values connected to the content (such as parameters of specific machines)Knowledge: information that is understood similarly and used/re-used by other parties

Ontology: Formally represented and machine interpretable representation.

Project: Is seen as series of events that together have a start and end.

SOA: Service Oriented Architecture

System Intelligence: Where the systems intelligence means intelligent behavior in thecontext of complex systems involving interaction, feedback and learning.

3/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

2 Introduction - Holons in LiteratureTechnical developments in recent years have produced stand-alone systems where per-formance is routinely reached. This solid background has allowed the extension of thesesystems into networks of components, which gather very heterogenous elements, each incharge of only a part of the holistic action of the system. As the systems are processoriented instead of knowledge oriented systems, the interaction between tasks cannot bemodeled, thus the effect of single interactions and relationships cannot be representedin the full systems scale. The types of interactions are changing into a complex net-work of possibilities within certain limits instead of a steady and predefined process flow.This situation is relatively new and causes pressures for defining the role of intended in-teraction. According to Chavalarias et al. [3], there is no doubt that one of the maincharacteristics of complex and adaptive production platforms in the future will be theever- increasing utilization of ICT. However, while the industrial world has seen the pos-sible advantages, the implementations fall short as a result of the changes to the wholeproduction paradigm that are required, going from preplanned hierarchical systems toadaptive and self-organizing complex systems [3, 4].

Complex systems are described as the new scientific frontier which has been advancingin the past decades with the advance of modern technology and increasing interest to-wards natural systems’ behavior. The main idea of the science in complex systems is todevelop through a constant process of reconstructing models from constantly improvingdata. The characteristics of a multiple- component systems is to evolve and adapt dueto internal and external dynamic interactions. The system keeps becoming a differentsystem. Simultaneously, the connection between the system and its surroundings evolvesas well. When a multiple-component system is manipulated it reacts by feedback, withthe manipulator and complex system inevitably becoming entangled.

In complex systems, reconstruction is searching for a model that can be programmed asa computer simulation that reproduces the observed data ’well’. The ideal of predictingthe multi-level dynamics of complex systems can only be done in terms of probability dis-tributions, i.e. under non- deterministic formalisms. An important challenge is, contraryto classical systems studies, the great difficulty in predicting the future behavior from theinitial instant as by their possible interactions between system components is shieldingtheir specific individual features. In this sense, reconstruction is the inverse problem ofsimulation. This naturally indicates that the complex system cannot be understood asdeterministic system, since the predictions from Complex Systems Science do not saywhat will happen, but what can happen [3, 4, 24, 12].

In general, complex systems have many autonomous units (holons, agents, actors, individ-uals) with adaptive capabilities (evolution, learning, etc), and show important emergentphenomena that cannot be derived in any simple way from knowledge of their compo-nents alone. Yet one of the greatest challenges in building a science of such systems is

4/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

precisely to understand this link - how it is that micro level properties determine or atleast influence those on the macro level. The current lack of understanding presents a hugeobstacle in designing systems with specified behavior regarding interactions and adaptivefeatures, so as to achieve a targeted behavior from the whole [3]. Due to the complexityof the system behavior and its possibilities, and the lack of tangible and implementableresearch results how the complex systems theory can bring revenue to the company; theimplementations at the moment are scarce and acceptance varies.

Bionic Manufacturing System (BMS): The BMS investigates the biological systemsand proposes concepts for future manufacturing systems. A biological system includesautonomous and spontaneous behavior and social harmony within hierarchically orderedrelationships. Cell as an example is a basic unit, which comprises all other parts of abiological system and can have different capabilities from each other, and are capable ofmultiple operations. The manufacturing units can act similar to cells as building blocksto derive hierarchical control structures, such as shops, factories and business units. Insuch structures, each layer in the hierarchy supports and is supported by the adjacentlayers. The components, including the part, communicate and inform each other of thedecision made by themselves [5, 2].

• Evolving capabilities

• New methods and methods integration

• Intelligent and adaptive structures

Fractal Factory (FF): The concept of fractal factory proposes a manufacturing companyto be composed of small components or fractal entities. These entities can be described yspecific internal features of the fractals. The first feature is self-organization that impliesfreedom for the fractals in organizing and executing tasks. The fractal components canchoose their own methods of problem solving including self-optimization that takes careof process improvements. The second feature is dynamics where the fractals can adaptto influences from the environment without a formal organization structure. The thirdfeature is self-similarity understood as similarity of goals among the fractals to conformthe objectives in each unit [5].

• Horizontal and vertical self-similarity on all structuring levels

Holonic Manufacturing System (HMS): The core of HMS is derived from the prin-ciples behind the term ’holon’. It comes from the Greek word ’holos’, which is a wholeand the suffix ’-on’, meaning a part. Therefore the term holon means something that isat the same time a whole and a part of some greater whole [15]. The model of integratedmanufacturing systems consists of manufacturing system entities and related domains,

5/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

the structure of individual manufacturing entities, and the structuring levels of the enti-ties. A manufacturing system is, at the same time, part of a bigger system and a systemconsisting of subsystems. Each of the entities posses self-description and capability forself-organization and communication [24, 20].

• Autonomous and co-operative entities

• Network-based teams

• Modular system structure

6/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 3.1: PROSA domains extended with DiMS domains

3 Holonic Manufacturing Systems: DiMS

3.1 Prosa Architecture

The model of manufacturing systems explains the system with manufacturing entitiesand their related domains, see Figure 3.1. The basic entities are products, resources, andorders, based on the reference architecture of HMS: the Product-Resource-Order-Staffarchitecture (PROSA) [23, 1]. In the PROSA-architecture the manufacturing entities(holons) are connected by the exchange of knowledge about the manufacturing systemwhilst in this approach their relations are achieved by focusing on manufacturing do-mains. The domains, namely process, production, and business, have specific roles ina manufacturing system and are responsible for their manufacturing activities. All ofthe entities and domains have to be considered in an integrated environment in order tooperate efficiently and successfully.

Products are what the manufacturing system is offering to its customers. Orders areinstances of products the customers are purchasing. The customers can be other entitieswithin the same enterprise, or entities in the enterprise network. The ordered productswill be manufactured with the resources existing in the manufacturing system.

The business domain connects products and orders. In the markets where a manufactur-ing system exists, the demand of customers has to be met with sufficient supply. In theprocess domain, the capability to manufacture the products is defined. The system needs

7/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 3.2: Digital, Virtual and Real dimensions

to be able to manufacture all of the features of the products, i.e. the resources should beassociated with corresponding methods. The resources, having the needed capabilities,also define the capacity of the system in the production domain. It is responsible for man-ufacturing right orders at the right time. It should have enough capacity to manufacturethe volume and scalability needed to handle any variation in orders. The competence ofa manufacturing system is defined by the skills needed in each of the different types ofentities and their related domains. Each of them has to be competent in order to achievefeasible and efficient results.

The manufacturing entities, despite having different roles, have similar internal structures.The structure consists of digital, real, and virtual parts explaining the autonomous partof the entities, see figure 3.2. The entities are connected via the communication part,which makes it possible to co-operate with other entities existing in the system.

The digital part includes all the digitally presented information and knowledge. It is usedfor developing and controlling the real system, as well as building the virtual models. Thereal part represents what exists physically in the real system, such as machines and tools,humans, and products to be manufactured. The virtual part is a representation of thephysical part as a computer model. This includes, for example, computer aided design(CAD) models of products and production facilities and simulation models of robots,machine tools, and manufacturing systems. The communications part is responsible forthe co-operation on the physical and information levels. The information part of thecommunication is the language and content of the data that are transferred within thesystem. The amount of information transferred between the system entities is kept to aminimum in order to reduce the complexity of the operations.

In the context of a currently operating manufacturing system, the information for the realand virtual parts is the same as that existing in the digital part. New information andknowledge, gathered from either the real or virtual worlds, is added to the digital part

8/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 3.3: The concept of distributed manufacturing systems

and made available for both. In future design and development cases, a copy of the digitalpart is used to avoid inaccurate information being added to the current system. This isdone to eliminate false information from failed ideas for future design and developmentcases.

3.2 DiMS

The concept of distributed manufacturing systems (DiMS), developed in a scientific re-search project FMS 2010, aims to create a more adaptive and autonomous manufacturingsystem. The intention is to integrate the design and development of products, productionsystems, and business processes into one environment where the parts of the system canexist in a distributed network both on the physical and information levels. This providesmore effective use of existing knowledge and skills as duplicate design and developmentprocesses can be reduced and more cost-effective solutions achieved.

Figure 3.3 illustrates the process of the DiMS concept. The process combines three mainphases: synthesis, solution creation, and the use of the system created. The phasesare connected with processes of emergence as implementation concepts, the implementa-tion of the new system, and the growth of skills and knowledge as the system is operat-ing. In the synthesis phase, the existing skills and knowledge of a manufacturing system

9/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

are combined with new requirements and possibilities, derived from ideas and needs forchange. In the process of creating the implementation concepts, the solution principlesare used to create the solution. In the event of contradictory situations, the differentgoals are analyzed, using the principle of positive intention [21]. This is done to achieve amutually acceptable solution that can be considered for implementation. When the newlyimplemented system is in operation, it is constantly developed. The knowledge and skillsof the manufacturing system are updated. During each of the phases, accepted principleswill be added to the existing skills and knowledge and they will form the basis for howfuture design and development challenges are met.

The process is iterative both at the whole process level and also in the steps of the process.For example, a synthesis can be repeated until acceptable solution alternatives are found.In a similar fashion, a whole loop can be repeated to achieve a feasible solution. In orderto meet the new requirements set by the evolving environment, several new manufacturingparadigms, which follow characteristics of natural systems have been introduced. Theseparadigms are:

The model of integrated manufacturing systems consists of manufacturing system entitiesand related domains, the structure of individual manufacturing entities, and the structur-ing levels of the entities. A manufacturing system is, at the same time, part of a biggersystem and a system consisting of subsystems.

The research on digital manufacturing systems, factories, and enterprises has no commonlyused definitions. Typical issues mentioned are, for example: the collection of systemsand methods [?]; the integration of current engineering practice and technology [19]; themapping of all the important elements of the enterprise processes [18]; an integratedapproach to enhance the product and production engineering processes (Kuhn, 2006),and the entirety of all methods and tools.

Similar to the definitions is the aim of managing the typically isolated and separateactivities as a whole by the means of Information and Communications Technology (ICT).

3.3 Extension of DiMS into EDMS

A digitally presented manufacturing system contains the information and knowledge ofmanufacturing entities and activities that it is reasonable to represent in a digital form.This, at its best, makes possible efficient collaboration between all the manufacturingactivities and related parties. An EDMS can briefly be defined as follows:

• integrated and collaborative environment for humans, machines, and infor-mation systems to act and interact;

10/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

• to enhance the research, development and management activities of prod-ucts, production systems, and business processes,

• supporting knowledge-intensive decision-making in the entirety of their life-cycles.

The main areas of EDMS are modeling, simulation, analysis, and the management ofchanging information and knowledge. The enablers for these are the possibilities providedby ICT, together with human knowledge and skills.

11/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

4 Requirements for Holonic Manufac-turing System

Holonic system is seen as evolving system where each entity has its own lifecycle phase.In order to compare dynamically changing system the following paragraph will show themain differences between stable (all legacy systems) and dynamic (agents and holons)system.

The characteristics of a stable system such as used in design and planning operations are:

• Environment the system (ob)serves is stable

• Environment is also closed from inputs, so the system it self is a closed world system

• Once the environment is stable, naturally it can be optimized

• The stable system is hierarchical system that is highly rigid

• Only pre-defined interactions between system parts.

• No interruption of pre-defined processes

• Linear thinking (outcome is result of multiple factors)

• System is fully controlled by human

The characteristics of a dynamic system are:

• Environment the system (ob)serves is dynamic

• Environment adapts due to the input from outside

• System the serves the adaptive environment cannot be rigid

• System like environment require feedback of actions

• Interrelated elements are working towards common goal; there are multiple causesand multiple effects

• The system is also seen as dynamic and responsive to the feedback

• Joining to the processes is more flexible

• Human is the input not the control

12/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

4.1 Requirements for System Description

Requirements from customers, needs for change, and general requirements combined withideas turn into solution principles. Further on in the process, the solution principlestranslate to solution alternatives, which define the manufacturing requirements. Figure8 presents an example process from the requirements towards an efficient solution tomeet customer demands. It includes the verification and validation of manufacturingcapabilities and the capacity to manufacture the new product. The three loops in theprocess are the product requirements loop, capability loop, and capacity loop.

The product is divided into features which form the service requests, the requirements forthe system. For each requirement there must exist a corresponding capability, a methodto manufacture the product. The resource having this capacity is the service provider.The first decision is made in the product requirements loop, where it is decided if theproduct design alternative is worth going on with. It is possible to go back and modifythe design or to check if there are capabilities. In the capability loop the result betweeneach of the requirements and capabilities can be classified, for example, as one of thefollowing five categories:

• Existing capabilities: The capabilities exist for all of the product requirementswithout any need for changes to the system. The products can be manufactured asthe service requests have service providers.

• Possible existing capabilities: At least some of the product requirements needfurther investigation as to whether the capabilities exist. The requirements are closeto the existing capabilities and, using modeling and simulation, the capabilities canbe verified.

• Capabilities after reconfiguration: There is no existing capability but it maybe possible to reconfigure the system so that it has the capabilities. By modelingthe reconfigured system the possibility can be verified.

• Capabilities after implementation: The system does not have the needed ca-pability. It may be possible if new capabilities are added to the system. Again thiscan be verified using modeling and simulation.

• No capability: The result may also be that there are no capabilities and theycannot be implemented either. This leads to the need for an alternative solution,which leads to a result that fits into one of the first four categories.

When it is known that the capabilities exist for all the product requirements, the efficiencyof the capabilities still needs to be evaluated against factors such as cost, quality, and

13/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

time. It has to be decided if the solution alternative is good enough. It can be furtherinvestigated in the capacity loop or it can be rejected and sent back to the capability loop.

If all the needed capabilities exist, the capacity of the system has to be checked. Thesame five categories can be used in capacity evaluation. If it is known that there isenough capacity, nothing else has to be done. Modeling and simulation can be used toverify that there is enough capacity. It can also be used in capacity reconfiguration andimplementation issues. Modeling and simulation of capacity has the same constraints asin the case of capabilities. The capacity for existing volume and variation still has to existwhen new products are considered as an addition to existing products. In the capacityloop, the solution can be accepted or rejected, as in the capability loop. If the solutionis rejected, it can be sent back to the capability loop or further back into the designrequirements loop.

All the solutions are the results of decisions which combine existing digital informationwith the new requirements and possibilities. The digital information and knowledge isinput as it is used as support for decisions when existing knowledge is combined withnew knowledge gathered by the new product requirements. It is also output from thesolutions, as the system is updated to include the new information and knowledge.

4.2 Requirements for Adaptiveness

If we want to create a system that is no longer a rigid hierarchical one option is to approachthe field of complex systems and system intelligence. The Holonic Manufacturing Conceptfor today is first; open world system and second it has characteristics of a complex system.

In literature the definitions are as following:

• Adaptive system: A system must continuously adapt itself to the changing natureof its environment or it will become extinct. Similarly a business system must changecontinuously to meet changing requirements of the environment. This adaptationis possible only through learning. Learning translates into adaptiveness. An adap-tive organization would be able to survive under an unfavorably changing businessenvironment.

• Complex system: Complex system is defined as a system that has a large numberof elements interacting nonlinearly.

• System Intelligence: Where the systems intelligence means intelligent behaviorin the context of complex systems involving interaction and feedback.

14/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 4.1: Changes that affect to the system

Changeability is widely discussed in [11] from several different aspects and is generallydefined as: characteristics to accomplish early and foresighted adjustments of the fac-tory’s structures and processes on all levels to change impulses economically. Similarly,changeability herein is used as a general term to explain the ability of a manufacturingsystem to change or to adapt to meet the requirements of the labile markets.

Figure 4.1 presents several typical sources of change that manufacturing systems facewhen they operate in a constantly changing environment. The changes can be external orinternal, direct or indirect [11]. Typical external and indirect sources for change are in thecontext of STEEP (society, technology, economy, environment, and politics) [13]. Lawsand different rules are examples of external and direct sources for change. These sourcescan be mandatory or voluntary. Mandatory sources force the manufacturing system toadapt to the changes. For those changes that are voluntary, it depends on the strategy anddecisions within the manufacturing system how to react to the situation as the decisionswill have an impact on the competence of the manufacturing system.

Competition, customers, partners, and suppliers are examples of external and directsources for change from the viewpoint of a manufacturing system. They differ from theother external and direct sources in their nature, as they are similar entities communi-cating in the same environment as the manufacturing system itself. Similarly, new ideas,strategy, and future visions can derive from the manufacturing system itself or influencedfrom the context. The external changes will cause internal changes that will change thesystem. The changes can affect the system entities as well as their related domains. Achange within a system will almost always cause a chain of change events until the systemhas adapted to the new situation.

15/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

4.3 Requirements for Changeablity

The changeability of manufacturing systems and entities can be classified into changingby requirements, changing by learning, and changing structure during the lifecycle of theentity. The entities face requirements for change during their life-cycles, and typically theentities must change during their existence both to meet the new requirements and toimprove their actions.

An entity may have to change because its requirements change. The need for change canbe seen from the vertical structuring levels of manufacturing resource entities:

• Industrial ecosystem - Being aware of existing and future possibilities and re-quirements.

• Enterprise network - To rapidly form a new enterprise network structure whenmarkets change.

• Enterprise - Transparent co-operation with suppliers, partners, and customers toget better results.

• Factory - Rapid response to changing product families.

• Manufacturing system - Flexibility to change manufacturing processes with min-imal reconfiguration.

• Manufacturing unit - To rapidly change the system configuration for the require-ments of new part families.

• Manufacturing device - Ability to change between work pieces with minimalsetup times.

Changing by learning can be understood as the evolution of skills and knowledge fromunknown towards core skills and knowledge; see Figure 6. An unknown activity cannot beconsidered until the possibilities are known. It requires new information and knowledgeto be acquired from the enterprise network or industrial ecosystem. When the activity isconsidered to be feasible, the needed technologies can be investigated. By having a widenetwork of knowledge, it is possible to gather information on the technologies, skills, andknowledge existing in the enterprise network. When the technologies are available, thesystem may be configured and the capabilities achieved. When the actual possibility isimplemented and integrated into the system, the capability exists in the system. As thesystem operates, the capabilities are constantly improved towards core information andknowledge by learning from actions.

16/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

The structure of a manufacturing entity consisting of digital, virtual, and real parts willchange during the life cycle of the manufacturing entity, and all of the parts don’t have toexist the whole life cycle. For example a product entity that is in the early phases of thedesign process, is an idea or a vision of what it could be, and has only a vague descriptionthat can be presented digitally. When the design evolves towards more detailed solutionprinciple, the digital part is more precise and there could be a virtual part, a computermodel that can be used to test the functionality and present the product idea to otherpeople. The physical part exists for the first time if prototypes are manufactured. Whenthe product entity is accepted into production, instances of product entities, the orderentities, are realized. They can have the physical product and also digital informationand virtual models of the product as parts of the service to a customer.

17/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 5.1: The past, present, and future time dimensions of EDMS

5 Solution Candidate for HolonicManufacturing System

5.1 Knowledge-based manufacturing processes

EDMS can be utilized in the past, present, and future time dimensions, illustrated inFigure 5.1. The past represents what has happened. It is the digital memory of thesystem. The time dimension of the present, what is happening now, is used to operatethe current system by monitoring the state of the system and comparing it to the desiredstate. The future dimension makes it possible to plan future manufacturing activitiesahead and to compare different changes in strategies.

The past exists as the history data and information collected from the manufacturingactivities when they happened. It is used to analyze previous activities in order to findout what happened and the reasons why it happened. In finding the root causes for phe-nomena, the system can learn from its past and prevent unwanted situations in the future.Rules for the autonomy of the manufacturing entities, as well as for their collaboration,can be enhanced and new rules can be created.

The present here means the near future, where no major changes are planned. It is, forexample, the use of existing resources and the planning and scheduling of customer orders

18/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

that have already been placed. In the present the digital and real existences co-exist. Asthe system operates the activities are logged, creating new history data to be analyzedand to aid decision-making. The state of the real manufacturing system can be seen inthe digital manufacturing system and actions can be taken with the state of the systemas a starting point.

The dimension of the future relies on the information and knowledge gathered from thesystem previously. Future design and development decisions are syntheses of existingcapabilities and requirements combined with future goals and possibilities. The viewpointof the future can be divided into tactical decisions and visions. Tactical decisions considerthe near future into which the manufacturing system is heading. Future visions are similarto tactical decisions, the difference being the time horizon. The outcome of future visionsis more obscure but there are more possibilities to be investigated.

The information and knowledge from analyzing the past, collecting data from the present,and forecasting the future is stored in the form of receipts. A receipt holds the capabilitiesof a system, constantly updating and refining the best practices in conjunction with humanskills and know-how. The receipts are the basis of the operations in the real present, theonly time dimension in the real world.

As a manufacturing system is a part of some larger system and at the same time consistsof subsystems, it can be examined on different vertical structuring levels. The right sideof the Figure 5.2 presents the hierarchy of the vertical structuring levels of manufacturingresource entities.

A manufacturing system entity consists of manufacturing unit entities, i.e. it is a networkof resources needed to manufacture all the features of product entities. It also has resourcesfor storing and handling material and transferring it between manufacturing units. Theproduct entities are typically part families from which the volume and variation of orders iscomposed. Manufacturing units consist of resource entities of machines, devices, workers,and other entities required, such as robots, fixtures, sensors, readers etc. The units aredesignated to manufacture certain product entities, i.e. work pieces that have similaritiesin size, shape, features, material properties etc. They are also required to produce a certainamount of order entities to keep the material flowing between manufacturing units.

A factory entity consists of manufacturing and assembly systems, as well as storage areasfor blank parts and final products, including both manufacturing and assembly units. Theproducts are typically final products and the customers are the final users of the products.Enterprises consist of factory units, which can exist globally. The distance between theentities brings logistics into the picture as an important factor. The structuring levels areself-similar externally in terms of the structure of the entities as they communicate in thesame environment. Despite their self-similarity, internally their autonomy can vary andthey can be different from each other, even when they have a similar role in the system.

19/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 5.2: The vertical structuring levels of manufacturing resource entities

The structuring level of enterprise networks differs from the levels below in that theenterprise entities in the network may have different owners and possibly contradictorygoals. The level of industrial ecosystems is similar to the level of enterprise networksas it consists of enterprise entities. At the same time it differs from enterprise networksas it is not under any administration. A manufacturing enterprise can have a certainamount of control over its own enterprise network, but it cannot control other entities inthe ecosystem. The levels of enterprise networks and industrial ecosystems are partiallyoverlapping. The whole industrial world consists of several regional ecosystems whilst anenterprise network can be distributed among several ecosystems.

From another viewpoint, manufacturing entities can be similar or different, depending onwho they are examined by. A product in a manufacturing system can be a resource in thesystem of the customer. Similarly, the resources in a manufacturing system are productsfrom the viewpoint of resource suppliers.

5.2 SOA and Holons in technical sense

The communication activities between the manufacturing entities are based on service-oriented architecture (SOA), which consists of self-describing components that support thecomposition of distributed applications enabling the autonomous manufacturing entities

20/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

to negotiate and share their information and knowledge [?]. The basic conceptual modelof SOA architecture consists of service providers, service requesters, and service brokers[7]. The entities in a digital manufacturing system based on SOA have the following roles:

• Service provider entities, which are typically the resource entities having thecapabilities of fulfilling the service requests.

• Service requester entities i.e., the product entities realized as order entities thatneed to be manufactured.

• Service broker represents the rules and logic of how the requester can find providers.The rules can be based on the competitive and sustainable factors important thatare strategically important.

In the proposed model of manufacturing systems having the basic building blocks, prod-ucts, resources, and orders, services happen in the domains of process, production, andbusiness. Knowledge-based services can be seen in three dimensions: role, context, andreceipt. They are based on the distributed character of knowledge: normative expec-tations, interactive situations, dispositions, objects, cognitive state, and capability [22].Each entity has a role in the system in which it exists, i.e. it has expectations of the otherentities. It is also one of the objects existing in the system. In the context an entity isperforming its activities as interactive situations where the cognitive state of the entitycollects data and information. The dimension of dispositions is seen as a receipt, data andinformation collected from the system, to learn and improve the capability of the entityas more rich information and knowledge.

Figure 5.3 presents an example service in a process domain between product and resourceentities. The resource entity is providing a service as requested by the product entity. Theservice is a manufacturing process occurring in a certain context. The actions during theservice depend on the skills of the resource entity and the state of the system. When theservice is functioning, both resource and product entities collect data from the process.By analyzing the collected data they can learn and update the receipt. When a certainproduct entity uses a service provided by a certain resource entity, the phases includeadding the same data and information to the knowledge of both entities’. The knowledgeof a resource entity is updated with several product entities using the services it provides.In a similar fashion, the knowledge of a product entity consists of all the services it requestsusing one or more service providers.

Holonic Definition can be summarized as following:

holon (philosophical) = role = object(ICT) = dynamic module(tech.) = service = costcenter = small world

puuttuu

21/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 5.3: Services in a process domain between product and resource entities

5.3 Technical Holon description

The basic building blocks of Holonic Framework are entities of products, resources, andorders [1], as well as their related manufacturing domains of process, production, andbusiness. The entities, despite being involved in different manufacturing activities, havegenerally similar structures [20]. Similarly, the manufacturing domains have differentroles, and the combined collaborative activities of the entities and domains make upthe total manufacturing system. The manufacturing entities are autonomous and co-operative as they collaborate within the same manufacturing system. Products representwhat is offered to customers and resources are static entities of a manufacturing system,for example, machines, tools, and devices. Orders as instances of products, the dynamicentities of a manufacturing system that appear in the system, flow through the systemand disappear. The entities are connected with the process, production, and businessdomains. The process domain offers the manufacturing capabilities and the productiondomain uses the resources efficiently to manufacture the right orders at the right times.The business domain is responsible for markets. The entities have internal structuresconsisting of digital, virtual, and real parts. The digital part contains the informationand knowledge, which can be presented digitally, whilst the real part represents what existsin the real manufacturing system. The virtual parts usually present the physical partsas simulation models. While the digital, virtual, and real parts present the autonomousnature of a manufacturing entity, the communication part is the capability of the entity

22/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 5.4: Generic ICT Architecture including the Processes

to collaborate in the system together with other entities.

The Holonic Framework as a core of ICT infra, illustrated in Figure 5.4, is intended to beused in dynamically changing design, planning and control environment. The framework’score is intended to stay same. It functions as a platform for adding and removing neededapplications and services. As the devices and device combinations join to the processesand leave the processes the focus of this framework is to support knowledge share above all,thus processes are considered as knowledge sharing request-service combinations. The coreoperations are based on Service Oriented Architecture approach, the holonic frameworkwill also act as message broker for the interoperating partner networks.

Purpose of the Holonic Framework is to combine all services and with its resources to singleentity that present all capabilities and restrictions. In right side of Figure 5.5 companiesA and B are inside the framework and they are provides combined services A, B and C.Now, when customer requests a service, or combination of services, the companies insidethe framework try to satisfy the service needs together. When an external service requestarrives to the Holonic Framework, the best company in the framework, to handle theservice, is determined. The selection of the provider is depending on the characteristicsof the request. Inside the selected company, it is decided what factory floor is best suitedto handle the service and the most suitable system inside the floor is selected. If a servicecannot be completed with only one resource, the service request is broke down into severalsub-requests. The result of the negotiation is sent back to upstream following the samepath the service was broke down. Requests are always delegated from an abstract level

23/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

Figure 5.5: Holonic Framework

to a more concrete level. Also the services can and should have same kind of hierarchy.We might start from service "make a car" that breaks down to "make two doors", "makechassis" and "make engine" and so on until the level of "make hole in position X, Y, Z,depth".

When actually realizing the system, the companies might have different implementationstrategies. In left side of Figure 5.5, company A has decided to separate the digital andreal models, so in this system the actual intelligence is located, for example, in a serverinside the server rack and in factory floor there is little or no intelligence. Company Bhas opted to integrate digital and real entities together. The systems know themselvesand their own capabilities, and when a service request is verified, it is done near themachine. Both have their benefits and drawbacks. A separate system is easy to migrateto an existing factory and this also makes practicable to handle larger datasets, sinceall information is located in one place. Integrating the real and virtual entities, enablesthe system to be more ready for expansion, but this option has more requirements forthe machines. If the service is not made in a holonic way, it is still possible to make itcommunicate with the rest of the system by using translation layer that implements the

24/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

holonic interfaces and services and then uses original service to acquire the results. Thisway from the outside view the system appears to be totally holonic although there arenon-holonic entities.

25/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

6 ConclusionsThe document listed the characteristics of holonic entities as those are seen in this re-search. It is understood that the holon as a term is very philosophical entity that canhave different concrete appearances in different levels of organization. In general, complexsystems have many autonomous units (holons, agents, actors, individuals) with adaptivecapabilities (evolution, learning, etc), and show important emergent phenomena that can-not be derived in any simple way from knowledge of their components alone. Yet one ofthe greatest challenges in building a science of such systems is precisely to understandthis link - how it is that micro level properties determine or at least influence those onthe macro level. The current lack of understanding presents a huge obstacle in design-ing systems with specified behavior regarding interactions and adaptive features, so asto achieve a targeted behavior from the whole. Due to the complexity of the system be-haviour and its possibilities, and the lack of tangible and implementable research resultshow the complex systems theory can bring revenue to the company; the implementationsat the moment are scarce and acceptance varies.

26/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

7 References[1] Van Brussel, H., Wyns, J., Valckenaers, P., Bongaerts, L. and Peeters P., 1998,

Reference architecture for holonic manufacturing systems: PROSA, Computers inIndustry, 37: 255- 274

[2] Ueda, K., Vaario, J. and Ohkura, K., 1997, Modelling of Biological ManufacturingSystems for Dynamic Reconfiguration, Annals of the CIRP, 46/1: 343-346

[3] Chavalarias, D., Cardelli, L., Kasti, J., et al., 2006, Complex Systems: Challenges andopportunities, an orientation paper for complex systems research in fp7, EuropeanCommission

[4] Cotsaftis, M. , 2009, A passage to complex systems, in Complex Systems and Self-organization Modeling, C. Bertelle, G. H. E. Duchamp, and H. Kadri-Dahmani, eds.,Springe

[5] Tharumarajah, A. , Wells, A.J. and Nemes, L., 1996, A Comparison of the Bionic,Fractal and Holonic Manufacturing Concepts, International Journal of ComputerIntegrated Manufacturing, vol.9, no.3/96

[6] Zaeh, M.F., Lau, C., Wiesbeck, M., Ostgathe, M. and Vogl, M., 2007, Towards theCognitive Factory, (Keynote Paper) Proceedings of the 2nd International Conferenceon Changeable, Agile, Reconfigurable and Virtual Production, Toronto, Canada, July22-24: 2- 16

[7] W3C, 2004, Web Services Architecture, http://www.w3.org/TR/ws-arch/, accessed18th July 2008

[8] Law, A.M. and Kelton, W.D., 2000, Simulation modeling and analysis, 3rd ed.McGraw-Hill, New York

[9] Law, A.M. and McComas, M.G., 1999, Simulation of Manufacturing Systems, In:Farrington PA, Nembhard HB, Sturrock DT and Evans GW (eds), Proceedings ofthe 1999 Winter Simulation Conference, Squaw Peak, Phoenix, AZ, 56-59

[10] Papazoglou M.P. and Georgakopoulos, D., 2003, Service Oriented Computing, Com-munications of the ACM, Vol.46, No.10, pp.25-28

[11] Wiendahl, H.-P. , ElMaraghy, H.A., Nyhuis, P., Zaeh, M.F., Wiendahl, H.-H., Duffie,N., and Brieke, M., 2007, Changeable Manufacturing - Classification, Design andOperation, Annals of the CIRP, 56/2 (2007) 783-809.

[12] Lanz, M. , 2010, Logical and Semantic Foundations of Knowledge Representation forAssembly and Manufacturing Processes, PhD thesis, publication 903

27/27

KIPPcolla DOC Date

1-2-2011Approved

FinalRevision

0.6Document NO

PB/100111Preparer

H. Nylund,M. Lanz

Reviewers

Kai Salminen, ReijoTuokko

e-mail address

hasse.nylund@tut.fi,minna.lanz@tut.fi

N.B.

PUBLIC

[13] Tuokko, R., 2009, Competitive Sustainable Manufacturing, Mass Customization &Personalization Conference, Helsinki, Finland, 2009, 44 p.

[14] Hellmann, A., Jessen, U., and Wenzel, S., 2003, e-Services - a part of the DigitalFactory, The 36th CIRP-International Seminar on Manufacturing Systems, pp 199-203

[15] A. Koestler, 1989, Ghost in the Machine, published by Arkana books

[16] KuehnW, 2006, Digital Factory - Integration of simulation enhancing the product andproduction process towards operative control and optimization, International Journalof Simulation, Vol. 7, No. 7, 2006, pp 27-39

[17] Maropoulos, P.G., 2003, Digital enterprise technology-defining perspectives and re-search priorities, International Journal of Computer Integrated Manufacturing, Vol.16, Nos. 7-8, 2003, pp 467-478

[18] Monostori, L., Haidegger, G., Vancza, J., and Viharos, Z.J, 2002, Digital Enterprises:A national project in Hungary, Proceedings of 1st CIRP(UK) International Seminaron Digital Enterprise Technology, September 2002, Durham, UK, pp. 269-272

[19] Reiter, W.F., 2003, Collaborative engineering in the digital enterprise, InternationalJournal of Computer Integrated Manufacturing, Vol. 16, Nos. 7-8, 2003, pp 586-589

[20] Salminen, K. ., Nylund, H., and Andersson, P., 2009, Role based Self-adaptationof a robot DiMS based on system intelligence approach, Flexible Automation andIntelligent Manufacturing (FAIM2009)

[21] Dilts, R., 1998, The Principle of Positive Intention, Robert Dilts NLP Home Page,Retrieved: 15th July 2008, http://www.nlpu.com/Patterns/pattern2.htm

[22] Dilts, R., 2005, Modeling, Robert Dilts NLP Home Page, Retrieved: 15th July 2008,http://www.nlpu.com/Articles/artic19.htm

[23] Wyns, J., 1999, Architecture for Holonic Manufacturing Systems - The Key to Sup-port Evolution and Reconfiguration, PhD thesis, K.U. Leuven, Belgium

[24] Valckenaers, P., Van Brussel, H., Bongaerts, L. and Wyns J., 1994, Results of theHolonic Control System Benchmark at the K.U. Leuven, Proceedings of the CIMATConference, Troy, NY, USA, 128-133

[25] Warnecke, H.J., 1993, The Fractal Company: A Revolution in Corporate Culture,Springer- Verlag, Berlin, Germany

28/27

Appendix