Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 62 (2017) 165 – 169

10th CIRP Conference on Intelligent Computation in Manufacturing Engineering - CIRP ICME '16

Analysis of control architectures in the context of Industry 4.0

Hermann Meissnera,*, Rebecca Ilsena, Jan C. Auricha
a
Institute for Manufacturing Technology and Production Systems (FBK), University of Kaiserslautern, Erwin-Schrödinger-Straße 57, Kaiserslautern, 67663,
Germany
* Corresponding author. Tel.: +49-631-205-4068; fax: +49-631-205-3304. E-mail address: publications.fbk@mv.uni-kl.de

Abstract

Industry 4.0 is a current research topic in the field of production engineering. One common characteristic of Industry 4.0 is decentralization which
can be implemented by a decentralized production control. Several researchers have already addressed decentralized production controls. This
paper focuses on the characteristics of Industry 4.0 as well as decentralized control approaches and hierarchies. Different properties of approaches
and architectures are compared to the objectives of Industry 4.0. Based on this comparison conclusions are drawn about how different
architectures suit Industry 4.0, and need for action for the development of production controls of Industry 4.0 is derived.
©©2017
2016TheTheAuthors. Published
Authors. by Elsevier
Published B.V. This
by Elsevier B.V.is an open access article under the CC BY-NC-ND license
Selection and peer-review under responsibility of the International Scientific Committee of “10th CIRP ICME Conference".
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 10th CIRP Conference on Intelligent Computation in Manufacturing Engineering
Keywords: Production planning; Scheduling; Planning; Industry 4.0

1. Introduction and to determine their main characteristics. Subsequently, in

As a response to the current challenges in fast changing
environments, the concept of Industry 4.0 has been discussed
in research for several years. With this concept, several
objectives are pursued. One major objective is to improve the
cost situation of a company by decreasing the costs per piece
[1]. Moreover, the production systems should become more
flexible regarding customer demands [2]. This results in small
batch sizes which increases complexity. To manage this
complexity another objective of applying Industry 4.0 is to
decentralize the structures of production [3]. Consequently,
Industry 4.0 focuses to improve the competitiveness by
reducing costs and increasing flexibility in decentralized
production systems to offer customized products, which is an
advantage in satisfying customer markets. Therefore, a high
level of productivity has to be reached to stay competitive.
The motivation to decentralize production systems by applying
decentralized production control is not new, but was discussed
already many years ago. Due to this history, an analysis of the
control architectures for production control is conducted with
respect to their suitability for Industry 4.0. To this end, the
terms Industry 4.0 and related cyber-physical production
systems (CPPS) are analyzed to investigate their backgrounds
chapter 3 the theory of centralized and decentralized
production planning and control is presented and the different
control architectures are explained. Finally, conclusions about
the control architectures with respect to Industry 4.0 are derived
in chapter 4 and an outlook for future research is given.
2. Industry 4.0
Despite the high relevance of Industry 4.0 in production
research and the high number of publications on this topic,
Industry 4.0 has not been clearly defined [4]. Moreover, the
understanding of what exactly pertains to this topic differs and
the affected research fields differ. In contrast to the past three
industrial revolutions, for the upcoming fourth revolution no
single technology has been identified that triggers Industry 4.0.
Instead, it can be described more precisely by a conjunction of
many technologies – both existing and new – which now work
together [5]. This combination of technologies leads to new
opportunities and applications in production systems.
Despite the lack of a standard definition, as previously
mentioned, there are technologies and characteristics that
clearly belong to Industry 4.0. One of the key technologies in
the fourth industrial revolution is the internet [6]. In a more
general way, Industry 4.0 can be described as a conjunction of

2212-8271 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the 10th CIRP Conference on Intelligent Computation in Manufacturing Engineering
doi:10.1016/j.procir.2016.06.113
166 Hermann Meissner et al. / Procedia CIRP 62 (2017) 165 – 169
information and communication technologies (ICT) with
production systems [7].
One of the key enablers of Industry 4.0 are cyber-physical
systems (CPS) [8]. CPS describe the amalgamation of
embedded systems and the internet to connect physical objects
[9]. If these systems are applied in production systems, the
products, the machines and other facilities become smart to
some degree, then cyber-physical production systems arise [8].
Thereby, the machines and corresponding facilities of the
production systems as well as the products are interconnected
with each other through the internet and information can be
exchanged. The connection through the internet allows on the
one hand more flexible interconnections of machines or other
facilities and offers a higher degree of flexibility as an ad-hoc
connection is enabled [6]. On the other hand, it enables a data
exchange in real-time so that information is available to every
element at all times.
To realize CPPS, several developments must be made to
today’s production systems. An important prerequisite is that
the elements of a production system need to become smart [7].
To this end, the elements will be equipped with memory
processors and the ability to communicate, and are additionally
equipped with sensors and actuators [10]. Sensors enable the
machines to collect information about their environment or
particular conditions that have to be monitored, whereas
actuators offer the opportunity to interact with the production
system.
In addition to the ability to process information, elements also
need to be clearly identifiable. The identification of every
element within the production system enables targeted
communication and data exchange of the machines. Moreover,
the elements need to be aware of their own conditions. Each
element must contain information about its original, current
and final condition, and include steps that allow autonomous
control of production [8].
Another important characteristic of Industry 4.0 is the already
mentioned ability to collect and use a large amount of data [11].
The act of data collection alone does not offer advantages.
Instead, to make use of the data and generate additional value,
it is important to analyze and process the data in real-time and
separate unimportant and important information [6]. Industry
4.0 offers potential to analyze the data of the production system
for patterns which can either be used to make better forecasts
for the future, or to improve decision making by discovering
weaknesses and taking into account the system’s current status
[12]. Moreover, monitoring enhances the understanding of the
machines in the production system or the system itself [13].
Hence, data collection and processing is a big advantage of
Industry 4.0, as it provides a better database for decision
making. This is especially true of the opportunity to constantly
gather refreshed data, which offers new applications and
possibilities for decision making in production systems.
The properties of Industry 4.0 respective to CPPS described so
far lead to another one of its central characteristics,
decentralization [14]. Smart elements, which are able to make
decisions on their own in combination with the information of
the production process, enable an autonomous, self-controlled
production done by machines and products. Decentralization is
pursued to manage complexity and requires a complete change
of the current centralized production control.
3. Architectures for production planning and control
approaches
Despite the theoretical benefits for decentralized production
control, today’s production control is predominantly
centralized [15]. The characteristics of the varied production
control styles will be subsequently described to better
understand the way they work.
3.1. Centralized production control
Centralized production planning and control is characterized by
the fact that all decisions are made by one centralized control
unit, which plans and schedules different production orders and
hands them over to the machines [16]. To handle the
complexity arising from comprehensive decisions, the
decisions are made in several steps. The hierarchical control
architecture does a partition of the overall problem into smaller
subproblems over the hierarchical levels [17]. Thereby,
information pathways over the hierarchical levels are well
defined, namely that commands always flow downwards in the
hierarchy and sensory data flows upwards [18].
The commands sent by the control unit mostly concern
determining scheduling, quantity and capacity planning.
Therefore the centralized approach requires stable and
predictable conditions. The schedules fixed by the central
control unit are handed over to the machines at the shop floor
level, which execute the defined schedules without any further
adjustment. Furthermore, the machines give feedback to the
central control unit. If changes occur for any reason, the central
control unit conducts a complete replanning with updated data.
In this process, it is necessary that the unit is continuously
provided with updated data [18].
The implementation of the centralized approach as done by a
hierarchical architecture is shown in fig. 1 on the left side. This
part of the figure illustrates the elements receiving commands
from the control unit to just execute the commands. Thereby,
the elements are dedicated to several defined tasks like machine
control or gathering information from the sensors.
Hierarchical Heterarchical
control structures control structures
Control system Product Resource
FBK/014-007
Fig. 1. Types of control architectures based on [19]
The strengths of a centralized approach and hierarchical
architecture are their access to global information, which is a
basic requirement for global optimization, and that all
information can be retrieved from a single source [20]. But the
single source and the centralized approach also have
disadvantages. Although centralized planning leads to precise
 Hermann Meissner et al. / Procedia CIRP 62 (2017) 165 – 169
production plans in stable conditions, in dynamic situations,
such as a breakdown of a machine, high deviations from the
predetermined plan may occur [21]. Moreover, if the deviation
from the original plan reaches a certain extent, a replanning of
the entire schedule for all machines is necessary [22]. This is
because the scheduling is always done by the centralized
decision-making unit alone, and the single machines do not
have the ability to immediately make corrections at shop floor
level themselves. This is exacerbated by the problem that the
planning data for rescheduling are very often already outdated
when they are passed through the hierarchies and reach the
centralized control unit [23]. Therefore, the centralized
production systems are inflexible with respect to changes and
thus constrained [24]. Finally, the modifiability and
extensibility is determined and limited by the central control
unit [20].
3.2. Decentralized production control
Due to the need for more reactive approaches and the long time
span of feedback loops in centralized approaches, decentralized
approaches come into consideration [25]. In contrast to the
centralized approach, in decentralized approaches the decision
power is shifted towards the elements of the production system,
i.e. that machines, products or other elements in the production
system can make decisions on their own without any superior
control unit [26]. As discussed in the description of Industry
4.0, to this end the single elements need to have the ability to
process information, and make and execute decisions, thereby
becoming autonomous. To make decisions, it is essential that
all elements have access to relevant information which means
that the elements in a decentralized production control have to
communicate with each other [27]. Due to this requirement,
communication also needs cooperation in the systems to
prevent disorder [28]. That is why decision making can also
mean that an element makes a suboptimal decision for itself to
achieve an overall objective [29].
With the power of communication and of decision making
directly controlled by the elements, and not a superior control
unit, the decentralized approach offers short decision-making
procedures [30]. This is supported by the fact that decentralized
decision-making needs less calculation due to the limited
extent of elements’ decisions, which need less time
computational effort [26]. From a structural point of view, the
decentralized approach is usually applied by heterarchical
architectures (fig. 1).
The decentralized control approach is fit for dynamic
environments in which many changes occur and corrections are
required as it can quickly adapt to changes [31]. The quick
reactions also facilitate enhanced stability of plans and delivery
dates as changes can promptly be made [23]. Furthermore,
replanning is conducted very frequently [24]. The production
control also improves in quality as the data are gathered
immediately where they occur, i.e. at shop floor level, which
leads to a more accurate database as well as increased
transparency [31]. Despite these advantages, decentralized
control systems also have their disadvantages. The main
problem here is the increased effort needed to coordinate the
single elements as each of them tries to pursue its own goals,
167
which does not necessarily lead to the global optimization of
the system [23]. Hence, in decentralized systems an alignment
of the goals of the individuals is needed so that they can
collectively achieve the goal of the production system.
3.3. Hybrid control architectures
Besides the centralized and decentralized approach, there is a
wide range of hybrid forms (fig. 1, in the middle). Generally,
the hybrid approaches try to combine the benefits of the
centralized and the decentralized approach [32]. This means
that hybrid approaches have some degree of hierarchy and top-
down control by a superior unit as well as autonomy of the
elements at shop floor level. On the one hand, the fixed
direction of communication of the hierarchical approach is
eliminated [33]. On the other hand, the elements can make
decisions on their own to a certain degree. For this purpose,
they also need to be smart to a certain extent, in contrast to the
hierarchical approach.
An exemplary hybrid form could be designed like this [34]: The
superior unit creates the schedules for the elements of the
production system. But in contrast to the centralized approach,
the schedules only advise the executing elements. As in the
decentralized approach, elements can depart from this advice
and instead decide in a different way. So if changes occur after
the schedules were fixed by the centralized control unit, the
elements on the shop floor can reject the proposed schedule and
make the best decision in the respective situation and based on
the available information.
3.4. Evolution and characteristics of past control approaches
Besides the current intention to use decentralized control
structures in Industry 4.0, there was a shift from centralized to
decentralized control structures 25 years ago to gain higher
local reactivity [18]. The motivation came from the weaknesses
of centralized controls in that their structure is determined at an
early phase in the design, and changes (i.e. extensions) are
difficult to conduct when the design is fixed and require high
costs [35]. Aside from extensions, the complexity resulting
from big hierarchical systems and the difficulty to obtain fault
tolerance in the system led to the development of heterarchical
control approaches.
To reach the objective of an increased fault tolerance in
heterarchical systems, the goal is to eliminate global
information which is needed by elements at shop floor level
[18]. The attempt to eliminate global information affected other
system properties and led to an improved modularity of the
system, modifiability and extendability as well as reduced
complexity [18]. Moreover, the decentralized structure also
promotes further beneficial properties and offers the possibility
of self-configurability, responsiveness and adaption in real
time to unexpected events like rush orders or machine
breakdowns [36].
Performance comparisons were conducted to compare the
different control architectures. In these comparisons [34],
classical hierarchical and heterarchical as well as hybrid
architectures were compared. A testbed was created simulating
a simple manufacturing system containing four machines and a
168 Hermann Meissner et al. / Procedia CIRP 62 (2017) 165 – 169
robot. In the simulations, the flow time and the tardiness of the
control architectures were compared. Deterministic and
stochastic scenarios were created and the planning horizons
were varied. The comparison showed that the hybrid
architecture performed better regarding flow time and tardiness
for most scenarios. Only in the stochastic scenario the
heterarchical architecture had better tardiness performance, and
when the planning horizon was very short, the fully
heterarchical architectures showed a shorter flow time. But for
the majority of scenarios, the hybrid architecture performed
better. The conclusion drawn from the performance
comparison was that heterarchical approaches do not deliver
satisfactory performance and planning to a certain degree is
necessary, even if the environment is unpredictable [34].
4. Conclusions
The results from past research, the characteristics of the
architectures, and the performance comparisons, deliver useful
insights for today’s research. The current situation in which the
shift to Industry 4.0 will take place is characterized by a
dynamic environment and highly customized products, which
implies small batch sizes and a high number of variants.
Measured against these conditions and the aforementioned
characteristics of different control approaches, it can be said
that only approaches with a certain degree of decentralization
should be considered. Fully hierarchical architectures are not
compatible with these conditions and decentralization is the
likely strategy to cope with modern conditions. This is common
sense regarding research for Industry 4.0.
However, the description of the control architectures, their
properties, and their performance show that a great deal of
research has been done to investigate hierarchical,
heterarchical and hybrid architectures. Although the
technological basis regarding ICT of the research conducted 25
years ago cannot be compared to current opportunities such as
the internet, the general objectives of the research at that time
show high similarities to those which are currently pursued by
Industry 4.0. Objectives such as reducing system complexity,
enabling changes or extensions of the system, or ability to react
to unexpected changes in real time are pursued by Industry 4.0,
too.
Besides the similarity of the objectives of past and current
research for Industry 4.0, the most interesting insight is the
results of the performance comparison of the control
architectures. It was shown that even in unpredictable
environments some planning improved the system
performance a lot. It shows that in simulations hybrid systems
outperform fully decentralized approaches. Although the
simulations are specific test cases, and can neither be
generalized nor adapted exactly to today’s circumstances, they
point out the abilities of hybrid architectures. And it is
suspected that these architectures play an important role in
future production planning [25]. Despite these results,
decentralization is currently the main method discussed in the
context of Industry 4.0 without consideration of the use of
central planning units, although in past research they did not
lead to satisfactory results. But it is not sufficient to only
investigate fully decentralized approaches due to their
weaknesses regarding achieving the system’s overall objective.
Therefore, due to the insights gained over the past decades,
hybrid architectures should be focused on and should be
included more in Industry 4.0 research, as they can adopt many
forms of production control between fully centralized and
decentralized control and show beneficial properties.
5. Summary
Industry 4.0 is a term that is still not clearly defined, and there
are different understandings of what applies to it. However,
some main characteristics could be identified. The most
important driving factor behind Industry 4.0 is the application
of the internet on the shop floor and the related connectivity of
the elements. The connectivity of machines enables
communication and cooperation. Thus, it allows an exchange
of data between the machines. The exchange and use of data by
the elements at shop floor enables a certain degree of autonomy
and moreover, a shift to decentralized production planning and
control, which is another key characteristic of Industry 4.0.
The shift in Industry 4.0 to decentralized production control
means turning away from the current hierarchical approach. To
this end, the concepts of fully heterarchical and hybrid
architectures are investigated and their properties are
compared. With respect to performance, tests from past
research are investigated in this paper. They point out that in
most cases, hybrid architectures show better performance in
uncertain situations, and lead to shorter flow times and
tardiness. Due to these results, the conclusion is derived that
the research in the field of Industry 4.0 should not focus only
on fully decentralized architectures as is currently the case, but
also pay attention to hybrid architectures as they have good
properties with respect to Industry 4.0 and show manifold
possibilities for application.
Acknowledgements
This work was partly funded by the German Federal Ministry
of Education and Research (BMBF) within the Framework
Concept “Research for Tomorrow’s Production” (funding
number 02PJ2573) and managed by the Project Management
Agency Karlsruhe (PTKA).
References
[1] Schuh G, Potente T, Wesch-Potente C, Weber AR, Prote J. Collaboration
Mechanisms to Increase Productivity in the Context of Industrie 4.0.
Procedia CIRP 2014;19:51-6.
[2] Dombrowski U, Wagner T. Mental Strain as Field of Action in the 4th
Industrial Revolution. Procedia CIRP 2014;17:100-5.
[3] Bauernhansl T. Die Vierte Industrielle Revolution. In: Bauernhansl T, ten
Hompel M, Vogel-Heuser B, editors. Industrie 4.0 in Produktion,
Automatisierung und Logistik. Wiesbaden: Springer Fachmedien
Wiesbaden; 2014. p. 5-35.
[4] Tschöpe S, Aronska K, Nyhuis P. Was ist eigentlich Industrie 4.0. ZWF
Zeitschrift für wirtschaftlichen Fabrikbetrieb 2015;110(3):145–9.
[5] Bildstein A, Seidelmann J. Industrie 4.0-Readiness: Migration zur
Industrie 4.0-Fertigung. In: Bauernhansl T, ten Hompel M, Vogel-
Heuser B, editors. Industrie 4.0 in Produktion, Automatisierung und
Logistik. Wiesbaden: Springer Fachmedien Wiesbaden; 2014. p. 581-
597.
 Hermann Meissner et al. / Procedia CIRP 62 (2017) 165 – 169
[6] Dais S. Industrie 4.0 – Anstoß, Vision, Vorgehen. In: Bauernhansl T, ten
Hompel M, Vogel-Heuser B, editors. Industrie 4.0 in Produktion,
Automatisierung und Logistik. Wiesbaden: Springer Fachmedien
Wiesbaden; 2014. p. 625-634.
[7] Von Lukas U, Stork A, Behr J. Industrie 4.0 - Evolution statt Revolution.
wt Werkstattstechnik online 2014;104(4):255-7.
[8] Anderl R, Strang D, Picard A, Christ A. Integriertes Bauteildatenmodell
für Industrie 4.0. ZWF Zeitschrift für wirtschaftlichen Fabrikbetrieb
2014;109(1-2):64-9.
[9] Broy M. Engineering Cyber-Physical Systems: Challenges and
Foundations. In: Aiguier M, Caseau Y, Krob D, Rauzy A, editors.
Complex Systems Design & Management: Proceedings of the Third
International Conference on Complex Systems Design & Management
CSDetM 2012. Berlin: Springer; 2013. p. 1-13.
[10] Zuehlke D. SmartFactory—Towards a factory-of-things. Annual
Reviews in Control 2010;34(1):129–38.
[11] Schuh G, Reuter C, Hauptvogel A, Dölle C. Hypotheses for a Theory of
Production in the Context of Industry 4.0. In: Brecher C, editor.
Advances in Production Technology. Cham: Springer Int. Publishing;
2015. p. 11-23.
[12] Friedli T, Lanza G, Schuh G, Reuter C, Arndt T, Fränken B, Lützner R,
Wenking M. Industrie 4.0 - ein Beitrag zur Entwicklung von "Smart
Networks". ZWF Zeitschrift für wirtschaftlichen Fabrikbetrieb
2015;110(6):378–82.
[13] Shrouf F, Ordieres J, Miragliotta G. Smart factories in Industry 4.0. In:
IEEE, editor. 2014 Int. Conference on Industrial Engineering and
Engineering Management (IEEM). 2014. p. 697-701.
[14] Imtiaz J, Jasperneite J. Scalability of OPC-UA Down to the Chip Level
Enables "Internet of Things". In: IEEE, editor. 2013 11th IEEE Int.
Conference on Industrial Informatics (INDIN). 2013. p. 500-505.
[15] Schlechtendahl J, Keinert M, Kretschmer F, Lechler A, Verl A. Making
existing production systems Industry 4.0-ready. Prod. Eng. Res. Devel.
2015;9(1):143–8.
[16] Lindemann M, Gronau N. Gestaltung marktorientierter
Produktionssysteme. In: Specht D, editor. Weiterentwicklung der
Produktion. Wiesbaden: Gabler; 2009. p. 43-59.
[17] Hax AC, Meal HC. Hierarchical Integration of Production Planning and
Scheduling. In: Geisler MA, editor. Logistics. Amsterdam: North-
Holland Publishing Company; 1975. p. 53-69.
[18] Duffie NA. Synthesis of Heterarchical Manufacturing Systems.
Computers in Industry 1990;14(1-3):167–74.
[19] Scholz-Reiter B, Freitag M. Autonomous Processes in Assembly
Systems. CIRP Annals - Manufacturing Technology 2007;56(2):712–29.
[20] Dilts DM, Boyd N, Whorms HH. The Evolution of Control Architectures
for Automated Manufacturing Systems. Journal of Manufacturing
Systems 1991;10(1):79–93.
[21] Grundstein S, Schukraft S, Görges M, Scholz-Reiter B. Autonome
Produktionssteuerung bei zentraler Produktionsplanung. ZWF Zeitschrift
für wirtschaftlichen Fabrikbetrieb 2013;108(12):992–4.
169
[22] Uhlmann E, Hohwieler E, Kraft M. Selbstorganisierende Produktion mit
verteilter Intelligenz. wt Werkstattstechnik online 2013;103(2):114–7.
[23] Franke J, Merhof J, Hopfensitz S. Einsatz von dezentralen
Multiagentensystemen. ZWF Zeitschrift für wirtschaftlichen
Fabrikbetrieb 2010;105(12):1075–8.
[24] Rohloff M. Decentralized production planning and design of a
production management system based on an object-oriented architecture.
Int. Journal of Production Economics 1993;30-31:365–83.
[25] Trentesaux D. Distributed control of production systems. Engineering
Applications of Artificial Intelligence 2009;22:971–8.
[26] Windt K, Hülsmann M. Changing Paradigms in Logistics -
Understanding the Shift from Conventional Control to Autonomous
Cooperation and Control. In: Windt K, Hülsmann M, editors.
Understanding autonomous cooperation and control in logistics. Berlin:
Springer; 2007. p. 1-16.
[27] Windt K, Böse F, Philipp T. Autonomy in production logistics:
Identification, characterisation and application. Robotics and Computer-
Integrated Manufacturing 2008;24(4):572–8.
[28] Bousbia S, Trentesaux D. Self-Organization in distributied
manufacturing control: state-of-the-art and future trends. In: El Kamel A,
Mellouli K, Borne P, editors. 2002 IEEE Int. Conference on Systems,
Man and Cybernetics. 2002.
[29] Brettel M, Friederichsen N, Keller M, Rosenberg M. How Virtualization,
Decentralization and Network Building Change the Manufacturing
Landscape: An Industry 4.0 Perspective. Int. Journal of Mechanical,
Aerospace, Industrial and Mechatronics Engineering 2014;8(1):37–44.
[30] Schuh G, Brosze T, Meier C. Gestaltungsaufgaben in der PPS. In: Schuh
G, Stich V, editors. Grundlagen der Produktionsplanung und -steuerung
1. 4th ed. Berlin: Springer Vieweg; 2012. p. 297-305.
[31] Märtens A, Elsweier M, Nickel R. Adaptive, dezentrale
Produktionssteuerung. wt Werkstattstechnik online 2007;97(6):471–8.
[32] Grundstein S, Schukraft S, Görges M, Scholz-Reiter B. Interlinking
central production planning with autonomous production control. In:
Marascu-Klein V, editor. Advances in production, automation and
transportation systems. Brasov: WSEAS Press; 2013. p. 326-332.
[33] Rogers P, Brennan RW. A simulation testbed for comparing the
performance of alternative control architectures. In: Andradóttir S, Healy
KJ, Withers, DH, Nelson, BL, editors. Proceedings of the 29th
conference on Winter simulation. Piscataway; 1997. p. 880–887.
[34] Brennan RW. Performance comparison and analysis of reactive and
planning-based control architectures for manufacturing. Robotics and
Computer-Integrated Manufacturing 2000;16(2-3):191–200.
[35] Duffie NA, Chitturi R, Mou J. Fault-tolerant Heterarchical Control of
Heterogeneous Manufacturing System Entities. Journal of
Manufacturing Systems 1988;7(4):315–28.
[36] Duffie NA, Prabhu VV. Heterarchil control of highly distributed
manufacturing systems. International Journal of Computer Integrated
Manufacturing 1996;9(4):270–81.