Description
Due to the dynamic and structural complexity of today's logistics systems and networks, central planning and control of logistic processes becomes increasingly difficult. Thus, decentralised and autonomous control of logistic processes is required.
Autonomous Logistic Processes
– New Demands and First Approaches –
B. Scholz-Reiter (2), K. Windt, M. Freitag
University of Bremen, Bremen, Germany
Abstract
Due to the dynamic and structural complexity of today’s logistics systems and networks, central planning and
control of logistic processes becomes increasingly difficult. Thus, decentralised and autonomous control of
logistic processes is required. Based on recent IC technologies such as RFID and wireless communication
networks, intelligent items, which can communicate and coordinate each other, are possible. These techno-
logical developments require novel concepts and strategies to implement autonomy in logistic processes.
This paper sketches the vision of autonomy in logistics, describes the new demands on autonomous logistic
processes and introduces a new German Collaborative Research Centre, which investigates autonomy as a
new control paradigm for logistic processes.
Keywords:
Production, Control, Autonomy
1 INTRODUCTION
There is an ongoing paradigm shift from centralised con-
trol of ‘non-intelligent’ items in hierarchical structures to-
wards decentralised control of ‘intelligent’ items in heterar-
chical structures in logistic processes. Those intelligent
items could either be raw materials, components or prod-
ucts as well as transit equipment (e.g. pallets, packages)
or transportation systems (e.g. conveyors, trucks). Reichl
describes such items as things that think [1]. The main
characteristic of an intelligent item is its capability to con-
trol itself, which means that these items act autonomous in
their planning and production processes.
Autonomy in general means the capability of a system,
process or an item to design its input-, throughput- and
output-profiles as an anticipative or reactive answer to
changing constraints of environmental parameters. One
specific criterion of autonomous processes or items is to
render a decision by itself on the basis of parameters,
which can lead to different but in principal predetermined
process or order fulfilment steps. The dynamic develop-
ment of information and communication technologies, e.g.
the RFID (Radio Frequency Identification) technology,
makes intelligent processes (and therefore intelligent
items or autonomy) possible.
Since January 2004, a German Collaborative Research
Centre (CRC 637) funded by the German Research Foun-
dation (DFG) has been established at the University of
Bremen. It is named “Autonomous Cooperating Logistic
Processes – A Paradigm Shift and its Limitations”. The
interdisciplinary project with the participating faculties of
engineering technology, economics, informatics, mathe-
matics and electrical engineering concentrates on model-
ling of autonomous cooperating logistic processes, de-
signing methods and adequate tools as well as on evalua-
tion for practical use.
This paper will explain the meaning of autonomy in logistic
processes on different examples and will give an overview
of possible logistics applications in the near future. One
key problem of autonomy is the detection of borderlines
from conventional management to autonomy.
The next chapter will give a description of the actual situa-
tion of logistic processes. The dynamics in logistics struc-
tures are characterised and the new technologies, impor-
tant for autonomous logistic processes, are addressed.
Next, a vision of autonomous processes is formulated in
response to existing and emerging problems.
The new demands on logistic processes are derived and
structured in technological, organizational and process
related demands followed by an overview of the organisa-
tional and topical structure of the collaborative research
centre 637.
The ideas and research objectives of the different sub
projects of the CRC will be presented in chapter 5. The
paper ends with first approaches demonstrated on an
example of a production process and with a closing sum-
mary.
2 CURRENT SITUATION IN LOGISTICS
2.1 Increasing Dynamics and Complexity
The more and more rapidly changing conditions in present
markets have an extensive impact on logistic processes.
The worldwide presence and market development of a
growing number of companies implicate the development
of global and complex intra- and inter-corporate logistics
networks [1].
The shift from seller to buyer markets pushed by the emer-
gence of the internet economy and the increasing impor-
tance of customer orientation and individualisation involve
a simultaneous atomization of payloads and increase of
shipment frequency as well as overall transport volume
[2].
There are naturally and economically limited infrastructure
expansion possibilities. In such an environment, today’s
concepts of planning and controlling of logistic processes
are starting to fail. The complexity and dynamics of widely
ramified and distributed value chains complicate the sup-
ply of all relevant information to a central entity.
This development requires new methods of planning and
control in production, transportation and procurement,
which means that instead of having central control of com-
plexity and dynamics, decentralized autonomous logistic
processes are implemented. One assumed characteristic
of autonomous processes or items is the adaptation of its
behaviour in a more effective manner on changing con-
straints. Due to this development, the achievement of
Published in: Monostori, L. (ed.): Proc. 37th CIRP International Seminar on Manufacturing Systems. Hungarian Academy of Science,
Budapest, Hungaria, 2004, pp. 357-362
This example demonstrates the combination of logistics
structures and new technologies to carry out autonomous
logistic processes and to comply with demands of today’s
markets.
4 NEW DEMANDS ON LOGISTIC PROCESSES
Establishing autonomous logistic processes requires a set
of general conditions on the logistics system, which must
be fulfilled. Figure 3 gives an overview of these demands
according to the holistic perspective on logistics.
Task
Layers
Logistic
System
Organisational demands
• Definition of autonomous logistics processes
• Definition of limits of conventional and autonomous control
• Availability of adequate information at the correct place in time
• Ability to measure and evaluate autonomous logistics processes
• Design and structuring of dynamic distributed targets
• Management strategies to consider external and internal risks
• Methods to establish efficient distributed Quality Management Systems
Technological demands
• System items’ ability to communicate and cooperate
• Distributed data management and date handling
• Mobile data communication technologies
• Data security guidelines regarding establishing mobile data communication
• Localisation’s ability
• Mobile hardware components (transponder etc.)
• Software requirements (new PPC-/ERP-functions)
Process-related demands
• Development of autonomous decision algorithms
• Development of strategies to use the process immanent intelligence
• Ability to model autonomous logistics processes
• Adaptation/development of PCS- and logistics-functions
• Robustness (resource, object, part)
• Divisibility of orders / mergence of intelligent items (e.g. assembly stage)
• Logical and physical reactivity
New demands on Logistics Processes
Organisation
and Management
Decision
System
Informatics
Methods and
I&C Technologies
Information
System
Material Flow
and Logistics
Execution
System
Task
Layers
Logistic
System
Task
Layers
Logistic
System
Organisational demands
• Definition of autonomous logistics processes
• Definition of limits of conventional and autonomous control
• Availability of adequate information at the correct place in time
• Ability to measure and evaluate autonomous logistics processes
• Design and structuring of dynamic distributed targets
• Management strategies to consider external and internal risks
• Methods to establish efficient distributed Quality Management Systems
Technological demands
• System items’ ability to communicate and cooperate
• Distributed data management and date handling
• Mobile data communication technologies
• Data security guidelines regarding establishing mobile data communication
• Localisation’s ability
• Mobile hardware components (transponder etc.)
• Software requirements (new PPC-/ERP-functions)
Process-related demands
• Development of autonomous decision algorithms
• Development of strategies to use the process immanent intelligence
• Ability to model autonomous logistics processes
• Adaptation/development of PCS- and logistics-functions
• Robustness (resource, object, part)
• Divisibility of orders / mergence of intelligent items (e.g. assembly stage)
• Logical and physical reactivity
New demands on Logistics Processes
Organisation
and Management
Decision
System
Organisation
and Management
Decision
System
Informatics
Methods and
I&C Technologies
Information
System
Informatics
Methods and
I&C Technologies
Information
System
Material Flow
and Logistics
Execution
System
Material Flow
and Logistics
Execution
System
Figure 3: System layers and their demands.
As shown in the figure above, the new demands on logis-
tic processes can be assigned to the system layers such
as the decision system, information system and execution
system and its task layers such as organisation and man-
agement, informatics methods, I&C technologies and
material flow and logistics. As a result, the demands can
be categorised in organisational demands, technological
demands and process-related demands.
4.1 Organisational demands
The creation of some new organisational general condi-
tions is essential to be able to establish autonomous logis-
tic processes. First of all, a definition of the term autono-
mous logistic processes and dissociation from conven-
tional logistic processes are necessary to get a consistent
understanding of this paradigm. Based on this definition,
the limits of conventional and autonomous control must be
investigated to be able to identify possible applications.
Characteristic for autonomous logistic processes is an
increased distributed information demand. So the avail-
ability of adequate information at the correct place in time
is one of the main organisational demands of autonomous
logistic processes.
The main target for establishing autonomous logistic proc-
esses is an increased efficiency of the logistics system.
Therefore, it is necessary to develop an evaluation system
that considers the changes in order processing due to
autonomy. It will enable the user to evaluate autonomous
processes against conventionally managed processes.
Furthermore it is necessary to define limitations of man-
agement and autonomy, to integrate different autonomy
views (resource, object, part) within the system structure.
So dynamical targets can be defined and new characteris-
tics of autonomy can be considered.
Autonomous control of logistic processes contains the
ability to react to special and unanticipated events. Con-
cerning the scheduling of appropriate activities adequate
management strategies are required to avoid or minimise
the risk. Another organisational requirement is the devel-
opment of a local quality management, where several
parts control their own quality management or that of other
items.
4.2 Technological demands
Regarding the information system layer there are some
new technological demands that must be considered with
respect to the paradigm shift. New or changed technologi-
cal demands result from a relocation of planning and con-
trol functions to parts or resources, especially regarding
the data management (consistency of data, high amount
of memory, etc.), data handling (coverage of information
overload, standardised interfaces etc.) as well as the abil-
ity of the system items to communicate and to cooperate.
The parts’ mobility makes new demands on data commu-
nication and localisation. Technologies like Bluetooth or
WLAN were suitable for mobile data communication, radio
communication based or satellite based positioning sys-
tems for parts’ localisation. Because of the employment of
mobile data communication technology, it is important that
the data security guidelines are observed. The hardware
demands will also change. The importance of transponder
technology will increase, especially concerning the men-
tioned amount of memory. An autonomous planning and
control of production systems at a lower level like subsys-
tem or part level contains relevant demands on used or
new software systems in the form of new functions.
4.3 Process-related demands
In addition to the organisational and technological de-
mands, certain process-related demands on the material
flow system and logistics system must be fulfilled to en-
able autonomous logistics processes.
A development of strategies is needed to use the process
immanent intelligence of subsystems and system ele-
ments, to reach autonomous decisions to achieve own or
predetermined aims. Therefore the intelligent items of the
system are able to execute independent problem solving
algorithms. By developing necessary software tools and
by evaluating the system in simulations it should be possi-
ble to model the selected system with its corresponding
processes. Currently existing planning and control sys-
tems have to be adapted to the new demands. New plan-
ning and control methods and their functions have to be
developed. By the implementation of autonomous logistics
processes the robustness of the logistics system must be
warranted to achieve undisturbed production. A further
requirement on the logistics system is the divisibility of
orders. This is necessary to enable autonomous control of
single processes in general. Maybe the autonomous items
have to influence each other. A fundamental requirement
to ensure the use of autonomous processes is the war-
ranty of logical (as described above) and physical reactiv-
ity of the involved systems. These include e.g. materials
handling, application technique and all other productive
units.
5 COLLABORATIVE RESEARCH CENTRE (CRC 637)
FOR AUTONOMY IN LOGISTICS
5.1 Idea of CRC 637
Since the beginning of the year 2004, the new German
Collaborative Research Centre “Autonomous Cooperating
Logistics Processes: A Paradigm Shift and its Limitations”
(CRC 637) has been established at the University of Bre-
men [13].
The objective of the CRC 637 is the systematic and broad
investigation and application of “autonomy” as a new con-
trol paradigm for logistics processes. For this, appropriate
concepts and models as well as methods and tools are
being researched and developed in twelve scientific sub-
projects. Thereby, the general idea of autonomy in logis-
tics includes concepts of management theory, computer
science as well as technical autonomy concepts from the
divers engineering domains. The CRC 637 uses a system
concept by Ropohl [14], who divides a system into an
execution system, an information system and a decision
system. The outcomes are three task layers of the CRC
637, depicted in figure 4.
Logistics System Task Layers within the CRC 637
Information System
Informatics methods
and I&C Technologies
Execution System
Human
Technical System
Mass
Energy
Information
Material flow and
Logistics
Decision System
Organisation and
Management
Figure 4: Holistic perspective on logistics and resulting
task layers within the CRC 637.
Resulting from figure 4, the objects of research are
• the autonomous physical material flow,
• the realisation of autonomy in the information system,
• the management of autonomous logistics processes.
To guarantee this holistic view on logistics and to cover
the three task layers depicted in figure 4, the CRC 637
encompasses researchers from the scientific disciplines
manufacturing engineering, business studies, computer
science, electrical engineering and mathematics.
For a structured investigation and application of the para-
digm “autonomy”, the CRC 637 is divided into three pro-
ject domains.
Project domain A “Modelling Foundations for Autonomous
Logistics Processes” will investigate the theoretical basis
of autonomy as well as modelling techniques for autono-
mous logistics processes. Thus, project domain A pro-
vides a general foundation for the other research activities
of the CRC 637.
Project domain B “Methods and Tools for Autonomous
Logistics Processes” uses the theoretic concepts and
modelling techniques from domain A or other related re-
search areas to develop methods and tools for efficient
and dynamic autonomous logistics processes. During the
first phase of the CRC 637, the applicability of already
established methods for autonomous issues will be inves-
tigated und applied. In the second phase, domain B will
focus on finding synergy effects between the divers con-
cepts developed in Project Domain A. The aim is the reali-
sation and implementation of methods and tools and their
transfer into practical projects.
In project domain C “Applications for Autonomous Logis-
tics Processes”, the results from domain A and B will be
used to solve specific practical problems. Insofar, project
domain C is based on the other two project domains and
will start in the second phase of the CRC 637. The subpro-
jects in this domain will search for potential logistics appli-
cation fields for autonomous logistics processes.
These project domains form the structure of the CRC 637,
which contains twelve subprojects (figure 5). Each subpro-
ject considers a certain problem from the viewpoint of the
particular scientific discipline, so that the holistic view of
autonomy in logistics is guaranteed.
Central Application Platform and Demonstrator
Planned for the 2nd and
3rd phase of the CRC
B1 Reactive Planning and Control
(Manuf.Engin. / Elec.Engin.)
B2 Adaptive Business Processes - Modelling and
Methodology
(Manuf.Engin.)
B3 Mobile Communication Networks and Models
(Elec.Engin.)
B4 Knowledge Management
(C.Sc.)
B5 Risk Management
(C.Sc. / Manuf.Engin.)
B6 Sensor Systems
(Elec.Engin.)
B7 Autonomous Adaptation of Vehicle Schedules
(Econ.)
A1 Fundamental Studies
(Manuf.Engin.)
A2 Sustainable Management
(Econ.)
A3 Monitoring of Autonomous Systems
(Econ.)
A4 Rule-based Graph Transformation
(C.Sc.)
A5 Dynamics of Autonomous Systems
(Manuf.Engin. / Math.)
C: Applications for
Autonomous Logistic
Processes
B: Methods and Tools for
Autonomous Logistic Processes
A: Modelling Foundations for
Autonomous Logistic Processes
Collaborative Research Centre 637
Autonomous Cooperating Logistic Processes: A Paradigm Shift and its Limitations
Figure 5: Structure of the CRC 637.
5.2 Subprojects from manufacturing engineering
Five subprojects will next be described which are related
to manufacturing engineering and technology.
The subproject A1 “Process-Oriented Fundamental Stud-
ies of Autonomous Logistic Processes” is motivated by the
main questions:
• Where do conventionally managed and autonomous
processes differ?
• What changes will autonomy cause in order process-
ing?
• Which methods are suited?
• How are autonomous processes measured and evalu-
ated?
As the project title points out, one aim of this project is to
define fundamental knowledge concerning autonomous
logistics processes, especially concerning the production
processes and the production planning and control proce-
dure. The second aim of the project is to develop an
evaluation system in order to measure the achievement of
logistics targets. Therefore, the project currently consists
of two parts: The fundamental studies and the develop-
ment of the evaluation system.
The subproject A5 “Modelling and Analysis of Dynamics of
Autonomous Logistics Processes” looks for autonomy in
self-organising systems e.g. biological systems and wants
to apply such natural autonomy concepts to logistics proc-
esses. The central questions to be answered are:
• How to design the local autonomy to reach a desired
global behaviour?
• How to design the global structure to enable local
autonomy?
To answer these questions, this subproject is arranged in
a dual way. The faculties of Mathematics and Manufactur-
ing Engineering cooperate to analyse the system‘s dynam-
ics as well as the system‘s performance. The mathemati-
cians focus more on the dynamics, while the production
engineers focus more on the performance of autonomous
logistics processes. Based on differential-equation sys-
tems, the mathematical side uses analytical models and
analyses the dynamic behaviour. The logistics side uses
discrete-event simulation and analyses both dynamics and
performance by statistical methods and time series analy-
sis. The dual way provides the possibility to evaluate and
improve the different models by comparing each other.
The subproject B1 “Reactive Planning and Control to Sup-
port Autonomous Objects in Multi-Modal Transportation
Processes” is motivated by constraint like varying envi-
ronmental conditions, inadequate knowledge and atomisa-
tion of load units, which appear in multi-modal transporta-
tion networks.
machine’s capacity, for example, by increasing its per-
formance and to counteract the bottleneck of the process
step turning.
Conclusion: In contrast to conventional control, where
generally the machine breakdown is recognised manually
by a worker, who takes necessary steps especially an
adaptation of the centrally planned PPC, autonomous
control is characterised by independent and immediate
reaction to unanticipated events. In this case, the machine
recognises its breakdown and takes necessary actions.
The basis for all further actions is the adjusted PPC of the
work step turning, which generates a set of information
flows to several system elements. Every system element
(resources, parts etc.) has got its own PPC or necessary
PPC-functions and so the ability to react fast and flexible
to changing general conditions. By means of this simple
scenario it becomes apparent that a focused, fast and
flexible reaction to unexpected disturbance in the form of
individual, self initialised adaptations of processes is pos-
sible by dislocating the planning and control functions as
well as embedding of partial intelligence to separate sub-
systems or system elements. In this case, positive effects
of autonomous logistic processes compared with conven-
tional controlled logistic processes could be less down
time because of minor response time, a decrease of per-
sonnel costs as one result of an increased level of auto-
mation and improved machine’s efficiency due to decen-
tralised, autonomous capacity planning. Those and other
effects will be analysed in the near future in more detail
with the help of simulation and process studies.
7 SUMMARY
The idea of autonomous logistic processes is presented in
this paper. Autonomy can occur as intelligent items, parts,
resources or processes. The main characteristic is that
those intelligent things are able to decide on their future
“life”. The described vision combines the current logistics
developments with new technologies. The RFID technol-
ogy enables logistic processes to become autonomous in
the future. To use this new technology requires the fulfil-
ment of several demands such as defining what kind of
information is needed to save on a possible RFID-tag of
an item. Those questions and others will be answered in a
new German Collaborative Research Centre (CRC 637)
established at the University of Bremen since January
2004. The main targets as well as the subprojects related
to manufacturing engineering and technology are pre-
sented. In order to demonstrate the achieved results of the
CRC 637, an application platform will be provided.
First approaches are shown on the example of a produc-
tion process, in which a resource disturbance demands
new planning and control strategies in order to fulfil the
order procedure with its given logistics targets. The aim is
to show the differences in information flows between cen-
tralised PPC systems and the idea of decentralised and
autonomous acting items and processes.
Future work will focus on the definition of autonomy in
logistic processes, its major characteristics as well as on
the development of new methods for autonomous logistic
processes. One key question to be answered concen-
trates on the limitations of autonomy: How much auton-
omy is useful and how much of conventional management
is still necessary? The new CRC 637 with its interdiscipli-
nary view will find answers to those questions in the fu-
ture.
8 ACKNOWLEDGMENTS
This research is funded by the German Research Founda-
tion (DFG) as the Collaborative Research Centre 637
“Autonomous Cooperating Logistic Processes: A Para-
digm Shift and its Limitations” (SFB 637).
9 REFERENCES
[1] Reichl, H.; Wolf, J., Okt. 2001, Things that think, "TU-
Berlin - Forschung aktuell", No. 49/Jahrgang 18, Ber-
lin, pp. 1887-1898.
[2] Kuhn, A.; Hellingrath, H., 2002, Supply Chain Mana-
gement – Optimierte Zusammenarbeit in der Wert-
schöpfungskette, Springer-Verlag, Berlin, Heidel-
berg.
[3] Westkämper, E.; Jendoubi, L., 2003, Smart Factories
– Manufacturing Environments and Systems of the
Future. In: Bley, H. (Hrsg.), Proceedings of the 36th
CIRP International Seminar on Manufacturing Sys-
tems, pp. 13-16.
[4] Finkenzeller, K., 2002, RFID-Handbuch - Grundlagen
und praktische Anwendungen induktiver Funkanla-
gen, Transponder und kontaktloser Chipkarten, 3.
Aufl., Carl Hanser Verlag, München.
[5] Zahariadis, Th., 2003, Evolution of the Wireless PAN
and LAN standards, in: Schumny, H. (Eds.), Com-
puter Standards & Interfaces, Volume 26, Issue 3,
pp. 175-185.
[6] Gebresenbet, G., Ljungberg, D., 2001, Coordination
and Route Optimization of Agricultural Goods Trans-
port to Attenuate Environmental Impact, in: Journal
of Agricultural Engineering Research, Academic
Press, Volume 80, Issue 4, pp. 329-342.
[7] Cheung, K. H., Misic, J., 2002, On virtual private
networks security design issues, in: Akyildiz, I.,
Rudin, H. (Eds.), Computer Networks, Volume 38,
Issue 2, pp. 165-179.
[8] Michalski, R. S.; Bratko, I.; Kubat, M., 1998, Machine
Learning and Data Mining, Methods and Applica-
tions, John Wiley, West Sussex, UK.
[9] Langenheinrich, M.; Friedemann, M., 2003, Digitali-
sierung des Alltags – Was ist Pervasive Computing?
In: Das Parlament - Aus Politik und Zeitgeschichte,
Deutscher Bundestag, No. 42, Berlin, pp. 6-12.
[10] Fleisch, E.; Kickuth, M.; Dierks, M., 2003, Ubiquitous
Computing: Auswirkungen auf die Industrie, in: In-
dustrie Management 19(2003)6, pp. 29-31.
[11] Scholz-Reiter, B., 1998, Chancen und Möglichkeiten
der reaktiven Planung und Steuerung von intermoda-
len Stückguttransporten, in: Fluhr, M. (Ed.), Innovati-
ve Lösungen für den Verkehr von morgen, In Time,
Berlin, pp. 32-48.
[12] Daganzo, C. F., 2003, A theory of supply chains,
Springer Verlag, Berlin, Heidelberg.
[13] Freitag, M., Herzog, O., Scholz-Reiter, B., 2004,
Selbststeuerung logistischer Prozesse – Ein Para-
digmenwechsel und seine Grenzen, in: Industrie Ma-
nagement 20(2004)1, pp. 23-27.
[14] Ropohl, G., 1979, Eine Systemtheorie der Technik –
Grundlegung der Allgemeinen Theorie, Carl Hanser
Verlag, München.
doc_594874527.pdf
Due to the dynamic and structural complexity of today's logistics systems and networks, central planning and control of logistic processes becomes increasingly difficult. Thus, decentralised and autonomous control of logistic processes is required.
Autonomous Logistic Processes
– New Demands and First Approaches –
B. Scholz-Reiter (2), K. Windt, M. Freitag
University of Bremen, Bremen, Germany
Abstract
Due to the dynamic and structural complexity of today’s logistics systems and networks, central planning and
control of logistic processes becomes increasingly difficult. Thus, decentralised and autonomous control of
logistic processes is required. Based on recent IC technologies such as RFID and wireless communication
networks, intelligent items, which can communicate and coordinate each other, are possible. These techno-
logical developments require novel concepts and strategies to implement autonomy in logistic processes.
This paper sketches the vision of autonomy in logistics, describes the new demands on autonomous logistic
processes and introduces a new German Collaborative Research Centre, which investigates autonomy as a
new control paradigm for logistic processes.
Keywords:
Production, Control, Autonomy
1 INTRODUCTION
There is an ongoing paradigm shift from centralised con-
trol of ‘non-intelligent’ items in hierarchical structures to-
wards decentralised control of ‘intelligent’ items in heterar-
chical structures in logistic processes. Those intelligent
items could either be raw materials, components or prod-
ucts as well as transit equipment (e.g. pallets, packages)
or transportation systems (e.g. conveyors, trucks). Reichl
describes such items as things that think [1]. The main
characteristic of an intelligent item is its capability to con-
trol itself, which means that these items act autonomous in
their planning and production processes.
Autonomy in general means the capability of a system,
process or an item to design its input-, throughput- and
output-profiles as an anticipative or reactive answer to
changing constraints of environmental parameters. One
specific criterion of autonomous processes or items is to
render a decision by itself on the basis of parameters,
which can lead to different but in principal predetermined
process or order fulfilment steps. The dynamic develop-
ment of information and communication technologies, e.g.
the RFID (Radio Frequency Identification) technology,
makes intelligent processes (and therefore intelligent
items or autonomy) possible.
Since January 2004, a German Collaborative Research
Centre (CRC 637) funded by the German Research Foun-
dation (DFG) has been established at the University of
Bremen. It is named “Autonomous Cooperating Logistic
Processes – A Paradigm Shift and its Limitations”. The
interdisciplinary project with the participating faculties of
engineering technology, economics, informatics, mathe-
matics and electrical engineering concentrates on model-
ling of autonomous cooperating logistic processes, de-
signing methods and adequate tools as well as on evalua-
tion for practical use.
This paper will explain the meaning of autonomy in logistic
processes on different examples and will give an overview
of possible logistics applications in the near future. One
key problem of autonomy is the detection of borderlines
from conventional management to autonomy.
The next chapter will give a description of the actual situa-
tion of logistic processes. The dynamics in logistics struc-
tures are characterised and the new technologies, impor-
tant for autonomous logistic processes, are addressed.
Next, a vision of autonomous processes is formulated in
response to existing and emerging problems.
The new demands on logistic processes are derived and
structured in technological, organizational and process
related demands followed by an overview of the organisa-
tional and topical structure of the collaborative research
centre 637.
The ideas and research objectives of the different sub
projects of the CRC will be presented in chapter 5. The
paper ends with first approaches demonstrated on an
example of a production process and with a closing sum-
mary.
2 CURRENT SITUATION IN LOGISTICS
2.1 Increasing Dynamics and Complexity
The more and more rapidly changing conditions in present
markets have an extensive impact on logistic processes.
The worldwide presence and market development of a
growing number of companies implicate the development
of global and complex intra- and inter-corporate logistics
networks [1].
The shift from seller to buyer markets pushed by the emer-
gence of the internet economy and the increasing impor-
tance of customer orientation and individualisation involve
a simultaneous atomization of payloads and increase of
shipment frequency as well as overall transport volume
[2].
There are naturally and economically limited infrastructure
expansion possibilities. In such an environment, today’s
concepts of planning and controlling of logistic processes
are starting to fail. The complexity and dynamics of widely
ramified and distributed value chains complicate the sup-
ply of all relevant information to a central entity.
This development requires new methods of planning and
control in production, transportation and procurement,
which means that instead of having central control of com-
plexity and dynamics, decentralized autonomous logistic
processes are implemented. One assumed characteristic
of autonomous processes or items is the adaptation of its
behaviour in a more effective manner on changing con-
straints. Due to this development, the achievement of
Published in: Monostori, L. (ed.): Proc. 37th CIRP International Seminar on Manufacturing Systems. Hungarian Academy of Science,
Budapest, Hungaria, 2004, pp. 357-362
This example demonstrates the combination of logistics
structures and new technologies to carry out autonomous
logistic processes and to comply with demands of today’s
markets.
4 NEW DEMANDS ON LOGISTIC PROCESSES
Establishing autonomous logistic processes requires a set
of general conditions on the logistics system, which must
be fulfilled. Figure 3 gives an overview of these demands
according to the holistic perspective on logistics.
Task
Layers
Logistic
System
Organisational demands
• Definition of autonomous logistics processes
• Definition of limits of conventional and autonomous control
• Availability of adequate information at the correct place in time
• Ability to measure and evaluate autonomous logistics processes
• Design and structuring of dynamic distributed targets
• Management strategies to consider external and internal risks
• Methods to establish efficient distributed Quality Management Systems
Technological demands
• System items’ ability to communicate and cooperate
• Distributed data management and date handling
• Mobile data communication technologies
• Data security guidelines regarding establishing mobile data communication
• Localisation’s ability
• Mobile hardware components (transponder etc.)
• Software requirements (new PPC-/ERP-functions)
Process-related demands
• Development of autonomous decision algorithms
• Development of strategies to use the process immanent intelligence
• Ability to model autonomous logistics processes
• Adaptation/development of PCS- and logistics-functions
• Robustness (resource, object, part)
• Divisibility of orders / mergence of intelligent items (e.g. assembly stage)
• Logical and physical reactivity
New demands on Logistics Processes
Organisation
and Management
Decision
System
Informatics
Methods and
I&C Technologies
Information
System
Material Flow
and Logistics
Execution
System
Task
Layers
Logistic
System
Task
Layers
Logistic
System
Organisational demands
• Definition of autonomous logistics processes
• Definition of limits of conventional and autonomous control
• Availability of adequate information at the correct place in time
• Ability to measure and evaluate autonomous logistics processes
• Design and structuring of dynamic distributed targets
• Management strategies to consider external and internal risks
• Methods to establish efficient distributed Quality Management Systems
Technological demands
• System items’ ability to communicate and cooperate
• Distributed data management and date handling
• Mobile data communication technologies
• Data security guidelines regarding establishing mobile data communication
• Localisation’s ability
• Mobile hardware components (transponder etc.)
• Software requirements (new PPC-/ERP-functions)
Process-related demands
• Development of autonomous decision algorithms
• Development of strategies to use the process immanent intelligence
• Ability to model autonomous logistics processes
• Adaptation/development of PCS- and logistics-functions
• Robustness (resource, object, part)
• Divisibility of orders / mergence of intelligent items (e.g. assembly stage)
• Logical and physical reactivity
New demands on Logistics Processes
Organisation
and Management
Decision
System
Organisation
and Management
Decision
System
Informatics
Methods and
I&C Technologies
Information
System
Informatics
Methods and
I&C Technologies
Information
System
Material Flow
and Logistics
Execution
System
Material Flow
and Logistics
Execution
System
Figure 3: System layers and their demands.
As shown in the figure above, the new demands on logis-
tic processes can be assigned to the system layers such
as the decision system, information system and execution
system and its task layers such as organisation and man-
agement, informatics methods, I&C technologies and
material flow and logistics. As a result, the demands can
be categorised in organisational demands, technological
demands and process-related demands.
4.1 Organisational demands
The creation of some new organisational general condi-
tions is essential to be able to establish autonomous logis-
tic processes. First of all, a definition of the term autono-
mous logistic processes and dissociation from conven-
tional logistic processes are necessary to get a consistent
understanding of this paradigm. Based on this definition,
the limits of conventional and autonomous control must be
investigated to be able to identify possible applications.
Characteristic for autonomous logistic processes is an
increased distributed information demand. So the avail-
ability of adequate information at the correct place in time
is one of the main organisational demands of autonomous
logistic processes.
The main target for establishing autonomous logistic proc-
esses is an increased efficiency of the logistics system.
Therefore, it is necessary to develop an evaluation system
that considers the changes in order processing due to
autonomy. It will enable the user to evaluate autonomous
processes against conventionally managed processes.
Furthermore it is necessary to define limitations of man-
agement and autonomy, to integrate different autonomy
views (resource, object, part) within the system structure.
So dynamical targets can be defined and new characteris-
tics of autonomy can be considered.
Autonomous control of logistic processes contains the
ability to react to special and unanticipated events. Con-
cerning the scheduling of appropriate activities adequate
management strategies are required to avoid or minimise
the risk. Another organisational requirement is the devel-
opment of a local quality management, where several
parts control their own quality management or that of other
items.
4.2 Technological demands
Regarding the information system layer there are some
new technological demands that must be considered with
respect to the paradigm shift. New or changed technologi-
cal demands result from a relocation of planning and con-
trol functions to parts or resources, especially regarding
the data management (consistency of data, high amount
of memory, etc.), data handling (coverage of information
overload, standardised interfaces etc.) as well as the abil-
ity of the system items to communicate and to cooperate.
The parts’ mobility makes new demands on data commu-
nication and localisation. Technologies like Bluetooth or
WLAN were suitable for mobile data communication, radio
communication based or satellite based positioning sys-
tems for parts’ localisation. Because of the employment of
mobile data communication technology, it is important that
the data security guidelines are observed. The hardware
demands will also change. The importance of transponder
technology will increase, especially concerning the men-
tioned amount of memory. An autonomous planning and
control of production systems at a lower level like subsys-
tem or part level contains relevant demands on used or
new software systems in the form of new functions.
4.3 Process-related demands
In addition to the organisational and technological de-
mands, certain process-related demands on the material
flow system and logistics system must be fulfilled to en-
able autonomous logistics processes.
A development of strategies is needed to use the process
immanent intelligence of subsystems and system ele-
ments, to reach autonomous decisions to achieve own or
predetermined aims. Therefore the intelligent items of the
system are able to execute independent problem solving
algorithms. By developing necessary software tools and
by evaluating the system in simulations it should be possi-
ble to model the selected system with its corresponding
processes. Currently existing planning and control sys-
tems have to be adapted to the new demands. New plan-
ning and control methods and their functions have to be
developed. By the implementation of autonomous logistics
processes the robustness of the logistics system must be
warranted to achieve undisturbed production. A further
requirement on the logistics system is the divisibility of
orders. This is necessary to enable autonomous control of
single processes in general. Maybe the autonomous items
have to influence each other. A fundamental requirement
to ensure the use of autonomous processes is the war-
ranty of logical (as described above) and physical reactiv-
ity of the involved systems. These include e.g. materials
handling, application technique and all other productive
units.
5 COLLABORATIVE RESEARCH CENTRE (CRC 637)
FOR AUTONOMY IN LOGISTICS
5.1 Idea of CRC 637
Since the beginning of the year 2004, the new German
Collaborative Research Centre “Autonomous Cooperating
Logistics Processes: A Paradigm Shift and its Limitations”
(CRC 637) has been established at the University of Bre-
men [13].
The objective of the CRC 637 is the systematic and broad
investigation and application of “autonomy” as a new con-
trol paradigm for logistics processes. For this, appropriate
concepts and models as well as methods and tools are
being researched and developed in twelve scientific sub-
projects. Thereby, the general idea of autonomy in logis-
tics includes concepts of management theory, computer
science as well as technical autonomy concepts from the
divers engineering domains. The CRC 637 uses a system
concept by Ropohl [14], who divides a system into an
execution system, an information system and a decision
system. The outcomes are three task layers of the CRC
637, depicted in figure 4.
Logistics System Task Layers within the CRC 637
Information System
Informatics methods
and I&C Technologies
Execution System
Human
Technical System
Mass
Energy
Information
Material flow and
Logistics
Decision System
Organisation and
Management
Figure 4: Holistic perspective on logistics and resulting
task layers within the CRC 637.
Resulting from figure 4, the objects of research are
• the autonomous physical material flow,
• the realisation of autonomy in the information system,
• the management of autonomous logistics processes.
To guarantee this holistic view on logistics and to cover
the three task layers depicted in figure 4, the CRC 637
encompasses researchers from the scientific disciplines
manufacturing engineering, business studies, computer
science, electrical engineering and mathematics.
For a structured investigation and application of the para-
digm “autonomy”, the CRC 637 is divided into three pro-
ject domains.
Project domain A “Modelling Foundations for Autonomous
Logistics Processes” will investigate the theoretical basis
of autonomy as well as modelling techniques for autono-
mous logistics processes. Thus, project domain A pro-
vides a general foundation for the other research activities
of the CRC 637.
Project domain B “Methods and Tools for Autonomous
Logistics Processes” uses the theoretic concepts and
modelling techniques from domain A or other related re-
search areas to develop methods and tools for efficient
and dynamic autonomous logistics processes. During the
first phase of the CRC 637, the applicability of already
established methods for autonomous issues will be inves-
tigated und applied. In the second phase, domain B will
focus on finding synergy effects between the divers con-
cepts developed in Project Domain A. The aim is the reali-
sation and implementation of methods and tools and their
transfer into practical projects.
In project domain C “Applications for Autonomous Logis-
tics Processes”, the results from domain A and B will be
used to solve specific practical problems. Insofar, project
domain C is based on the other two project domains and
will start in the second phase of the CRC 637. The subpro-
jects in this domain will search for potential logistics appli-
cation fields for autonomous logistics processes.
These project domains form the structure of the CRC 637,
which contains twelve subprojects (figure 5). Each subpro-
ject considers a certain problem from the viewpoint of the
particular scientific discipline, so that the holistic view of
autonomy in logistics is guaranteed.
Central Application Platform and Demonstrator
Planned for the 2nd and
3rd phase of the CRC
B1 Reactive Planning and Control
(Manuf.Engin. / Elec.Engin.)
B2 Adaptive Business Processes - Modelling and
Methodology
(Manuf.Engin.)
B3 Mobile Communication Networks and Models
(Elec.Engin.)
B4 Knowledge Management
(C.Sc.)
B5 Risk Management
(C.Sc. / Manuf.Engin.)
B6 Sensor Systems
(Elec.Engin.)
B7 Autonomous Adaptation of Vehicle Schedules
(Econ.)
A1 Fundamental Studies
(Manuf.Engin.)
A2 Sustainable Management
(Econ.)
A3 Monitoring of Autonomous Systems
(Econ.)
A4 Rule-based Graph Transformation
(C.Sc.)
A5 Dynamics of Autonomous Systems
(Manuf.Engin. / Math.)
C: Applications for
Autonomous Logistic
Processes
B: Methods and Tools for
Autonomous Logistic Processes
A: Modelling Foundations for
Autonomous Logistic Processes
Collaborative Research Centre 637
Autonomous Cooperating Logistic Processes: A Paradigm Shift and its Limitations
Figure 5: Structure of the CRC 637.
5.2 Subprojects from manufacturing engineering
Five subprojects will next be described which are related
to manufacturing engineering and technology.
The subproject A1 “Process-Oriented Fundamental Stud-
ies of Autonomous Logistic Processes” is motivated by the
main questions:
• Where do conventionally managed and autonomous
processes differ?
• What changes will autonomy cause in order process-
ing?
• Which methods are suited?
• How are autonomous processes measured and evalu-
ated?
As the project title points out, one aim of this project is to
define fundamental knowledge concerning autonomous
logistics processes, especially concerning the production
processes and the production planning and control proce-
dure. The second aim of the project is to develop an
evaluation system in order to measure the achievement of
logistics targets. Therefore, the project currently consists
of two parts: The fundamental studies and the develop-
ment of the evaluation system.
The subproject A5 “Modelling and Analysis of Dynamics of
Autonomous Logistics Processes” looks for autonomy in
self-organising systems e.g. biological systems and wants
to apply such natural autonomy concepts to logistics proc-
esses. The central questions to be answered are:
• How to design the local autonomy to reach a desired
global behaviour?
• How to design the global structure to enable local
autonomy?
To answer these questions, this subproject is arranged in
a dual way. The faculties of Mathematics and Manufactur-
ing Engineering cooperate to analyse the system‘s dynam-
ics as well as the system‘s performance. The mathemati-
cians focus more on the dynamics, while the production
engineers focus more on the performance of autonomous
logistics processes. Based on differential-equation sys-
tems, the mathematical side uses analytical models and
analyses the dynamic behaviour. The logistics side uses
discrete-event simulation and analyses both dynamics and
performance by statistical methods and time series analy-
sis. The dual way provides the possibility to evaluate and
improve the different models by comparing each other.
The subproject B1 “Reactive Planning and Control to Sup-
port Autonomous Objects in Multi-Modal Transportation
Processes” is motivated by constraint like varying envi-
ronmental conditions, inadequate knowledge and atomisa-
tion of load units, which appear in multi-modal transporta-
tion networks.
machine’s capacity, for example, by increasing its per-
formance and to counteract the bottleneck of the process
step turning.
Conclusion: In contrast to conventional control, where
generally the machine breakdown is recognised manually
by a worker, who takes necessary steps especially an
adaptation of the centrally planned PPC, autonomous
control is characterised by independent and immediate
reaction to unanticipated events. In this case, the machine
recognises its breakdown and takes necessary actions.
The basis for all further actions is the adjusted PPC of the
work step turning, which generates a set of information
flows to several system elements. Every system element
(resources, parts etc.) has got its own PPC or necessary
PPC-functions and so the ability to react fast and flexible
to changing general conditions. By means of this simple
scenario it becomes apparent that a focused, fast and
flexible reaction to unexpected disturbance in the form of
individual, self initialised adaptations of processes is pos-
sible by dislocating the planning and control functions as
well as embedding of partial intelligence to separate sub-
systems or system elements. In this case, positive effects
of autonomous logistic processes compared with conven-
tional controlled logistic processes could be less down
time because of minor response time, a decrease of per-
sonnel costs as one result of an increased level of auto-
mation and improved machine’s efficiency due to decen-
tralised, autonomous capacity planning. Those and other
effects will be analysed in the near future in more detail
with the help of simulation and process studies.
7 SUMMARY
The idea of autonomous logistic processes is presented in
this paper. Autonomy can occur as intelligent items, parts,
resources or processes. The main characteristic is that
those intelligent things are able to decide on their future
“life”. The described vision combines the current logistics
developments with new technologies. The RFID technol-
ogy enables logistic processes to become autonomous in
the future. To use this new technology requires the fulfil-
ment of several demands such as defining what kind of
information is needed to save on a possible RFID-tag of
an item. Those questions and others will be answered in a
new German Collaborative Research Centre (CRC 637)
established at the University of Bremen since January
2004. The main targets as well as the subprojects related
to manufacturing engineering and technology are pre-
sented. In order to demonstrate the achieved results of the
CRC 637, an application platform will be provided.
First approaches are shown on the example of a produc-
tion process, in which a resource disturbance demands
new planning and control strategies in order to fulfil the
order procedure with its given logistics targets. The aim is
to show the differences in information flows between cen-
tralised PPC systems and the idea of decentralised and
autonomous acting items and processes.
Future work will focus on the definition of autonomy in
logistic processes, its major characteristics as well as on
the development of new methods for autonomous logistic
processes. One key question to be answered concen-
trates on the limitations of autonomy: How much auton-
omy is useful and how much of conventional management
is still necessary? The new CRC 637 with its interdiscipli-
nary view will find answers to those questions in the fu-
ture.
8 ACKNOWLEDGMENTS
This research is funded by the German Research Founda-
tion (DFG) as the Collaborative Research Centre 637
“Autonomous Cooperating Logistic Processes: A Para-
digm Shift and its Limitations” (SFB 637).
9 REFERENCES
[1] Reichl, H.; Wolf, J., Okt. 2001, Things that think, "TU-
Berlin - Forschung aktuell", No. 49/Jahrgang 18, Ber-
lin, pp. 1887-1898.
[2] Kuhn, A.; Hellingrath, H., 2002, Supply Chain Mana-
gement – Optimierte Zusammenarbeit in der Wert-
schöpfungskette, Springer-Verlag, Berlin, Heidel-
berg.
[3] Westkämper, E.; Jendoubi, L., 2003, Smart Factories
– Manufacturing Environments and Systems of the
Future. In: Bley, H. (Hrsg.), Proceedings of the 36th
CIRP International Seminar on Manufacturing Sys-
tems, pp. 13-16.
[4] Finkenzeller, K., 2002, RFID-Handbuch - Grundlagen
und praktische Anwendungen induktiver Funkanla-
gen, Transponder und kontaktloser Chipkarten, 3.
Aufl., Carl Hanser Verlag, München.
[5] Zahariadis, Th., 2003, Evolution of the Wireless PAN
and LAN standards, in: Schumny, H. (Eds.), Com-
puter Standards & Interfaces, Volume 26, Issue 3,
pp. 175-185.
[6] Gebresenbet, G., Ljungberg, D., 2001, Coordination
and Route Optimization of Agricultural Goods Trans-
port to Attenuate Environmental Impact, in: Journal
of Agricultural Engineering Research, Academic
Press, Volume 80, Issue 4, pp. 329-342.
[7] Cheung, K. H., Misic, J., 2002, On virtual private
networks security design issues, in: Akyildiz, I.,
Rudin, H. (Eds.), Computer Networks, Volume 38,
Issue 2, pp. 165-179.
[8] Michalski, R. S.; Bratko, I.; Kubat, M., 1998, Machine
Learning and Data Mining, Methods and Applica-
tions, John Wiley, West Sussex, UK.
[9] Langenheinrich, M.; Friedemann, M., 2003, Digitali-
sierung des Alltags – Was ist Pervasive Computing?
In: Das Parlament - Aus Politik und Zeitgeschichte,
Deutscher Bundestag, No. 42, Berlin, pp. 6-12.
[10] Fleisch, E.; Kickuth, M.; Dierks, M., 2003, Ubiquitous
Computing: Auswirkungen auf die Industrie, in: In-
dustrie Management 19(2003)6, pp. 29-31.
[11] Scholz-Reiter, B., 1998, Chancen und Möglichkeiten
der reaktiven Planung und Steuerung von intermoda-
len Stückguttransporten, in: Fluhr, M. (Ed.), Innovati-
ve Lösungen für den Verkehr von morgen, In Time,
Berlin, pp. 32-48.
[12] Daganzo, C. F., 2003, A theory of supply chains,
Springer Verlag, Berlin, Heidelberg.
[13] Freitag, M., Herzog, O., Scholz-Reiter, B., 2004,
Selbststeuerung logistischer Prozesse – Ein Para-
digmenwechsel und seine Grenzen, in: Industrie Ma-
nagement 20(2004)1, pp. 23-27.
[14] Ropohl, G., 1979, Eine Systemtheorie der Technik –
Grundlegung der Allgemeinen Theorie, Carl Hanser
Verlag, München.
doc_594874527.pdf