Task-Specific Decision Support Systems in

Multi-Level Production Systems based on the

digital shadow
Daniel Pause1, Philipp Brauner², Marco Faber³, Markus Fischer1, Philipp Hünnekes4,
Henning Petruck³, Alexander Mertens³, Verena Nitsch³, Günther Schuh4, Volker Stich1, Martina Ziefle²
1Institute for Industrial Management, RWTH Aachen University, Aachen, Germany

{daniel.pause, markus.fischer, volker.stich}@fir.rwth-aachen.de

²Human-Computer Interaction Center (HCIC), RWTH Aachen University, Aachen, Germany
{brauner, ziefle}@comm.rwth-aachen.de
³Institute of Industrial Engineering and Ergonomics, RWTH Aachen University, Aachen, Germany
{m.faber, h.petruck, a.mertens, v.nitsch}@iaw.rwth-aachen.de
4Laboratory for Machine Tools and Production Engineering (WZL), RWTH Aachen University, Aachen, Germany

{p.huennekes, g.schuh}@wzl.rwth-aachen.de

Abstract — Due to the increasing spread of Information and needs custom parts that have to be constructed and produced.
Communication Technologies (ICT) suitable for shop floors, the Therefore, companies need new routing plans with new target
production environment can more easily be digitally connected times. Due to the increasing demand in individual products
to the various decision making levels of a production system. and specifications, the number of variants increases, thereby
This connectivity as well as an increasing availability of high- leading to higher complexity. Modular kits facilitate
resolution feedback data, can be used for decision support for production planning, but different routings and process steps
all levels of the company and supply chain. To enable data must be created for all variants. In order to meet customer
driven decision support, different data sources were structured requirements, this paper presents an approach which tries to
and linked. The data was combined in task-specific digital
create a high degree of transparency for the order and
shadows, selecting clustering and aggregation rules to gain
production process in a holistic approach. The increased
information. Visual interfaces for task-specific decision support
systems (DSS) were developed and evaluated positively by transparency allows a better evaluation of production
domain experts. The complexity of decision making on different capacities and their flexibility. To gain transparency, data
levels was successfully reduced as an effect of the processed from the production has to be collected, processed and
amounts of data. These interfaces support decision making, but presented to support the decision-making processes for a
can additionally be improved if DSS are extended with smart variety of tasks along the order processing.
agents as proposed in the Internet of Production.
II. INTRODUCTION
Keywords—Decision Support System, Production Networks, For enabling cross-domain decision-making processes in
Production Management, Internet of Production, Digital Shadow manufacturing companies the Internet of Production (IoP) was
developed as reference architecture in Aachen, Germany. The
I. MOTIVATION
IoP is structured along three cycles that are oriented towards
Complexity in production networks increases the life cycle of products: Development Cycle, Production
continuously due to customer wishes for flexibility, Cycle and User Cycle. Over the cycles, the layers build up the
customized products and shorter innovation cycles. This IoP as a hierarchical structure. The expert systems (e.g. ERP,
transforms static supply chains into dynamic production PLM, CAD) are the lowest level which enable the domain-
networks. Due to increasing use of communication specific experts the access to information. The higher layer
technology, the transparency along the supply chain raises. access the data stored in expert systems through a
The current status and the estimated arrival time of the orders “Middleware+”. In the Smart Data layer the data will be
are known at any time. aggregated and consolidated in the Digital Shadow. The
Digital Shadow is a description of the data formats, the data
To meet the requirements for individualized products at
sources and the data granularity needed for further task-
low selling prices the automation and the planning of the
specific evaluation of the data. This turns the digital shadow
production have to be optimized. Each individualized product
in a real-time representation of the production system.

Fig. 1. Company planning levels supported by decision support systems

l-))) 
 There may be different Digital Shadows serving the methodology. The decision support system consists of four
various data driven tasks of a company like order processes, key components: knowledge management, conceptual
the production planning or the assembly of a product. modelling, experimental design and decision modeling. The
Advanced Analytics and Learning Algorithms use these decision modeling component uses qualitative and
Digital Shadows to transform production data into usable quantitative models to analyze the supply chain configuration.
information and perform fully automated control loops, semi-
automated processes, and decision support systems for human B. Intra-company decision support
decision makers and working persons on the shop floor. In The scope of shop floor operators has evolved
order to offer an effective, efficient and satisfying decision- considerably from “low knowledge level and single tasks” to
making basis, the relevant data must be identified, processed, “high knowledge level and extensive tasks” requiring
and presented in decision dashboards [1]. The requirements of effective DSS [1, 11]. Using manual, autonomous, or human-
different users should be met and cognitive under- and robot collaboration workstations the time of replacement for a
overload should be avoided. Additionally interpretation and machining tool can, for instance, be supported by a DSS.
decision errors should be prevented. To achieve all of this, Benkedjouh et al. [12] implemented the monitoring of tool
principles of ergonomics, user-centered and participatory wear as part of a prognostics and health management
design, as well as contextual design must be taken into account comprising the step of decision support. In [13] a DSS for a
for the design of the decision dashboards [2–6]. milling plant was developed. By simulating the system
behavior for multiple alternatives of interventions in case of
With this reference architecture, the decision support for an alarm, recommendations for appropriate actions can be
various decisions can be implemented based on different derived and are presented to the operator. A rule-based DSS
expert systems. Fig. 1 shows the different planning levels of a for the manual assembly of final trim and wheels in an
company that are all affected by the above changes. To automotive production with the target of making decisions
support human decisions in a company the goal of this paper like when, where, and how components should be installed is
is to design a role-based DSS based on the IoP as reference presented in [14]. Another important decision task in
architecture. production is the scheduling that was focused in [15] by
III. STATE OF THE ART FOR DECISION SUPPORT IN PRODUCTION implementing a schedule methodology for machine tool
availability. With the motivation of energy-optimizing, Plitsos
NETWORKS
et al. [16] implemented a multi-objective DSS for production
There are already numerous studies in the literature [6]- scheduling, which is able to decrease the energy consumption
[27] dealing with decision support in production networks. In in a plant.
the following, studies in the context of inter- and intra-
company decision support are presented. The degree of decision support in process planning
depends on the chosen approach and degree of automation,
A. Inter-company decision support ranging from manual, variant, semi-generative to generative
The existing work in the context of inter-company [17]. While manual process planning traditionally relies on
decision support mainly uses methods of mathematical human knowledge and experience [18], the expert planner’s
optimization as well as simulative approaches. Bounif and decisions are usually supported by comparison catalogues of
Bourahla [7] developed a framework for decision support in relative costs as well as registers of the companies’ machines
supply chain networks and combined optimization methods and their capabilities [19]. Decisions in variant process
and simulation techniques. The four main components are planning are supported by the identification of similar
supply chain model, decision module, optimization module, components in a part family based on classification systems
and simulation module. The supply chain model describes the [20]. This classification and identification can be supported by
network architecture and provides information for the approaches such as presented in [21] using Fuzzy Logic and
decision module. The Decision Module proposes strategic, Artificial Neural Networks for identifying suitable existing
tactical, and operational decisions based on information and process plans. For (semi-)generative process planning,
decision rules [7]. The solutions are then evaluated in the numerous approaches for automation and decision support
optimization and simulation module [7]. exist [18, 22].
Pires et al. [8] presented a simulation-based optimization Existing approaches generally support the generation of
approach for integrated material inventory, production and process plan information for consideration and possible
transport planning and control [8]. The model consists of the further editing by the expert planner. They are generally based
supply chain reality or test bench and a simulation-based on automated manufacturing constraints identification,
optimization block (SBO). The supply chain reality or test identification of possible alternatives for process plan
bench describes the environment in which the planning is elements in a predetermined database, as well as alternative
applied. From this, the SBO is supplied with real-time data selection based on optimization criteria such as balancing the
and the simulation scenarios are updated. resource workload [23–25]. To account for effects of the
dynamic shop floor environment Joo et al. [26] and Schmidt
The hybrid simulation model of Tan et al. [9] combines a [27] suggest dynamic approaches generating suggestions for
system dynamics approach with agent-based simulation and individual operations in real time based on the current
thus enables simulation and integrated consideration of the conditions.
entire supply chain. The comprehensive view of the supply
chain is carried out using a System Dynamic approach, while In reality, process planning is increasingly supported by
the detailed view of operational processes is carried out using computer-aided process planning (CAPP) systems to meet,
an agent-based model[9]. among other things, the requirements regarding planning
speed, quality and standardization [28]. Since truly generative
Chandra and Grabis [10] developed a decision support systems have not been realized yet, commercial CAPP
system as part of their supply chain configuration systems either support the variant or semi-generative planning


 approach utilizing either standardized or altered master The decision points derived from the SCOR reference
process plans based on parts families [17]. The described structure are divided into strategic, tactical, and operational
CAPP systems provide valuable process support for planning decisions. For the development of a decision support based on
components in the pre-defined part families. However, these digital shadows, operational decisions will be examined in
systems do not meet the requirements to generate process more detail. Analyzing the “source” processes of the SCOR
plans for components with feature combinations beyond the reference structure, the relevant decision points including
boundaries of the originally defined part families or their personnel deployment, supplier selection, order pooling, order
composite parts [29]. Furthermore, these systems rely on the triggering, and supply quantity are verified. The “deliver”
results of earlier planning operations [17] and therefore processes are used to derive decision points in the distribution
extrapolate potential errors and inconsistencies. area. In order to improve the quality of decisions, companies
need a comprehensive and realistic image of corporate events.
While in the described work the emphasis is on the To describe, for example, an order in the purchasing or
preparation of decisions by means of mathematical and delivery department both tangible aspects (the good, transport
simulative approaches, in the given work the necessary vehicles, etc.) and intangible aspects (ID-numbers,
conditions for a decision support are to be developed. For this geolocation, etc.) have to be considered.
purpose, data has to be made accessible and filtered for
relevant data. In contrast to the examined work the given work During the procurement and delivery process deviations
will focus not only on the development of the DSS but also on from the schedule may occur and are being detected in real-
the provision and selection of the used data. time (e.g. wrong geolocation of the transport vehicle). There
should be an unique order- or delivery-ID to each order to
IV. DIGITAL SHADOWS FOR THE DIFFERENT PLANNING LEVELS enable the traceability of an order during the procurement and
IN A PRODUCTION NETWORK delivery process. Furthermore, product ID is recorded to
The methodology for deriving the dashboard which are differentiate between different products of an order. Based on
described in the fifth section is show in Fig. 2. The method the integration of tracking technologies (e.g. GPS) a live
consists of three steps. The first step comprises the analysis of tracking of the geolocation of an order is possible to ensure
the decision structure, from which the data requirements are the routing of the order between two points. This enables to
derived (Section IV). The third step (Section V) comprises the determine the current location of the order. Deviations from
design of the dashboards in iterative loops with possible users the planned route are converted to delays in the delivery.
for the user-oriented presentation of the data. Featured by the use of sensor technologies and a real-time
routing, the status of an order between different steps in the
A. Inter-company decisions and digital shadow procurement plan can be obtained and decisions regarding
The Supply Chain Operations Reference (SCOR) model disposition of order can be made on data supported basis.
[30] is used for identifying relevant decision points for inter-
B. Intra-company decisions and digital shadow
company planning activities. SCOR is the most widely
accepted framework for evaluating and comparing supply Customer wishes for flexibility, customized products, and
chain activities and performance. It comprises different shorter innovation cycles have to be supported by flexible and
schematic process levels. In case of inter-company scope of fast inter-company decisions such as efficient process
this section “source” and “deliver” processes are considered planning [31]. Despite the development towards Simultaneous
in greater detail, since they represent the external boundaries Engineering (SE), new product designs or product variants
of the enterprise and communicate with external actors in the finalized by the engineering design department need to be
supply chain. translated into reliable macro level [29] process plans quickly,
in order to facilitate order planning.

Fig. 2. Methodology of developing the decision support dashboards for the inter-company decisions


 To support the decision on a suitable process sequence for person is supported in deciding when to take a break. By
machining a component, past production orders can be keeping track of the work already performed and utilizing the
clustered according to their similarity regarding a combination task forecast, the system suggests the time of a break wisely
of several features, such as diameters, lengths, and number of between two tasks, so that the completion of one task is not
drillings, grooves etc. New components, for which a process delayed and optimizes the distribution of the breaks. In
plan is to be generated, can then be easily assigned to these addition to the above data, the physical and mental load caused
clusters. Information on process sequences, successfully used by the assembly of the specific products has to be considered
in the past to process the required feature combinations, can for the calculation of the break times.
be used by the process planner. Similarly, to allocate specific
resources to a planned process sequence, information on V. DESIGN OF A DASHBOARD FOR DECISION SUPPORT
machine sequences, successfully used in the past in given To support human decisions and improve their quality,
process sequences, can support the process planner’s decision several dashboards from supply chain level (top level) to shop
as well. floor level (bottom level) were developed. For the dashboard
Processing times of production orders on specific on supply chain level the order number, material number as
resources are subject to fluctuations, due to various well as a brief description of the material are known for each
influencing factors, such as tool conditions and other order and extracted from the database of the company’s ERP
parameters. However, by providing statistical information on system. The material is assigned to a supplier that is identified
the distribution of real processing times for the required with the Supplier ID. In addition to this information the
feature combinations on a given resource, the process decision maker can see the order quantity, the dimensions of
planner’s decision on determining a standard time for new the material ordered and the trigger point of the order.
components can be effectively supported. Using the promised delivery date, the expected delivery
In contrast to fully automated production systems, the date and the planned assembly date, the DSS calculates the
operation times of hybrid production systems like human- delivery date difference and the difference to the assembly
robot collaboration workstations or manual workstations are date. The decision maker is able to identify an order that
much less predictable and can therefore only be considered as requires an action to avoid a delay in production due to a too
a reference point of operation times in the planning process. small difference from the assembly date. Furthermore, the
In reality, the operation times are influenced by individual system displays the orders for which the risk is higher, but an
capabilities of the working person, form on the day, action is not immediately required. This makes the decision
distractions during work, etc., which might lead to shifts in the maker aware of these orders.
production plan. Consequently, the completion times and The difference between delivery dates is used to calculate
possibly even completion orders differ from the planning the adherence to delivery and the average delivery delay.
phase. By adapting to these uncertainties a prediction of the Based on this information, the human decision maker sees the
real operating times taking the current working person’s supplier’s performance and can forward this information to
performance into account can be made [32]. These predictions the supplier relationship management.
can be used to forecast the production schedule and therefore
be utilized in an optimization of the production plan by The process planning view gives a concise overview over
workstations to keep the desired order of product completion. products to be planned, as well as process information on past
In case of a possible workstation shift, the shift option is production orders matching the required feature combination.
presented to the supervisor, who can then decide and initiate To support the process sequence determination, the three most
the necessary actions. Necessary data for these decisions are common process sequences of production orders matching the
real-time operation times of all workstations of the production feature combination of a selected part are listed in descending
system, norm times of the production processes and the order of occurrence in past production orders. After reviewing
production plan. these options, the process planner can select one of the
suggested process sequences for adoption to the suggested
At the workstation itself, the same data is used to display process plan. Based on the decision for a process sequence,
deviations from the targeted completion time providing past production orders are filtered and the three most
transparency of the working person’s efficiency to achieve the commonly allocated machine sequences, matching the feature
target time. In order to avoid stress, the robot also adapts to combination of the selected part, are displayed for review.
the working speed and initiates the next collaborative activity, Again, a suitable machine sequence can be selected by the
when the worker has completed the current work step. In process planner and subsequently will be added to the
production, the ergonomic design of work is of crucial suggested process plan. To support decision making regarding
importance. For that reason, taking healthy body postures and the standard times of the processes in the suggested process
breaks according to ergonomic guidelines is supported by the plan, the process planner is provided with the average
system. In case of critical body postures, the system displays processing time of the applicable past production orders on the
a warning and provides a suggestion on how the posture can allocated resource. This average processing time is presented
be improved. This assistance is based on the Ovako Working as the default value on a slider, which can be modified by the
Posture Analysis System (OWAS) [33], which takes joint process planner within the range of processing times of past
angles to calculate the goodness of postures. The data required production orders. The selected processing time is adopted as
are captured by markerless motion tracking using a 3D depth- standard time to the suggested process plan. The suggested
image camera. process plan on macro level is now fully populated and can be
In regard to the physical and mental work load, several modified by the expert planner or transmitted to the ERP
small breaks are recommended rather than one single long system to facilitate order planning.
break for sustained activities [34]. Thereby, the risk of The view on the shop floor is divided in two parts. The
accidents can be minimized, and the maximum stress of the first view, which displays an overview of all workstations or
working person can be reduced. For this reason, the working machines in the factory, is intended for the supervisor. In


addition to the order list, provided with ID and estimated production process, and the design of multiple user-centered
production time, there is a status display for each workstation and role-adaptive decision dashboards that facilitate
in terms of a traffic light showing the current productivity monitoring and controlling the various aspects of production
status aggregated from the quality of the manufactured goods process in manufacturing companies.
and the deviation from the standard time. Individual
information about the quality, occupancy time, and deviation With the developed dashboards, the system provides a
from the standard time is also depicted. If an optimization of straightforward user-interface and support on different levels
the production order is possible by shifting tasks between of the company. The major improvement is the provision of
workstations, the information with details about the task and aggregated data for the decision maker. Before he had to
involved workstations pops up in the decision support area on collect and evaluate the necessary data from different sources
the right. For further information, a logging module at the manually. The automatic generation of proposals based on
bottom informs about relevant events such as finished historical decisions or agent-based decision making would be
products, new orders, or workstation changes. an improvement for the user.

The second part of the shop floor view is intended for For developing this demonstrator, we initially focused on
usage by the working person. On the left side of the view, the the interconnection of the different data sources into a single
model. In the future, special attention will have to be paid to
same information as in the supervisor view is shown. The
middle part of the view contains information about the product the usability and user experience of such dashboards: While
type and assembly instructions. The DSS for improving accessing, presenting, and utilizing large amount of digital
physical ergonomics is displayed on the right side. The body data at different levels of a production system promises
postures are displayed in the categories red, yellow, and green enormous advantages, these vast amounts of data must be
in a detailed itemization for back, arms and legs. A red level usable for the human decision maker. Following, for example,
indicator displays the need for a pause. The higher the level, Norman’s “Seven stages of action”, successful interaction
the sooner a pause should be taken. with systems and the human-in-the-loop requires that the
presented data is correctly perceived, interpreted, and
VI. VALIDATION evaluated by operators [35].
To evaluate the developed decision support system for the ACKNOWLEDGMENT
different dashboards domain experts were asked to evaluate
the functions and the usability. The domain experts can be The authors would like to thank the German Research
divided by their backgrounds. One group has experience in the Foundation DFG for the kind support within the Cluster of
development of business applications and the other group in Excellence “Integrative Production Technology for High-
working with various business applications across multiple Wage Countries” (EXC 128) and the Cluster of Excellence
company levels. The domain experts were asked for their “Internet of Production (ID 390621612). [36].
feedback while getting explained and using the dashboards.
REFERENCES
Both groups pointed out the innovation factor and the high [1] M. Holm, “The future shop-floor operators, demands,
usability of the DSS dashboards. The dashboard for
requirements and interpretations,” Journal of
supporting decisions in the supply chain received positive
feedback for the usability and fast overview of all open Manufacturing Systems, vol. 47, pp. 35–42, 2018.
supplier deliveries. An integration into automatic production [2] André Calero Valdez, Philipp Brauner, Anne Kathrin
planning via agents was suggested as a further field of Schaar, Andreas Holzinger, and Martina Ziefle,
research to further reduce the workload on the decision “Reducing Complexity with Simplicity - Usability
makers. The support for creating routing for new components Methods for Industry 4.0,” in Proceedings 19th
was praised as a real time saver because the data was presented Triennial Congress of the IEA, 2015.
in a comprehensible interface without searching in many [3] H. Beyer and K. Holtzblatt, Contextual design:
systems. They proposed an evaluation of the quality of the Defining customer-centered systems. San Francisco,
routings as an improvement for further steps. The dashboards Calif: Morgan Kaufmann, 1998.
for controlling the shop floor were evaluated positively. The [4] C. Ware, Information visualization: Perception for
dashboard with the overview of all workstations has many design, 3rd ed. Amsterdam, Boston: Elsevier/MK,
familiar elements which both evaluating expert groups know 2013.
from different business applications. Moreover, the field with [5] P. Brauner and M. Ziefle, “Human Factors in
suggested changes of orders was evaluated as a positive Production Systems,” in Lecture Notes in Production
innovation. As an improvement the domain experts would see Engineering, Advances in Production Technology, C.
the automatic generation of decisions based on historic data of Brecher, Ed., Cham: Springer International Publishing,
the shop floor controller. The dashboard for the workstation 2015, pp. 187–199.
operator was praised for the innovative support for the [6] P. Brauner, R. Philipsen, A. Calero Valdez, and M.
physical ergonomics. Here an improvement would be a
Ziefle, “What happens when decision support systems
suggestion of pause time between two batches or orders.
fail?: The importance of usability on performance in
VII. CONCLUSION erroneous systems,” Behaviour & Information
Technology, vol. 24, pp. 1–18, 2019.
In this article the motivation, design, and implementation
of interconnected smart decision dashboards for the Industrial [7] M. E. Bounif and M. Bourahla, “Decision Support
Internet and the Internet of Production was presented. A key Technique for Supply Chain Management,” CIT, vol.
feature is the seamless integration of production data from 21, no. 4, p. 255, 2013.
multiple tiers of a production network (supply chain, [8] M. C. Pires, E. M. Frazzon, A. M. Carreirão Danielli,
production, shop floor) into a single holistic data model, the M. Kück, and M. Freitag, “Towards a simulation-based
aggregation of this data into a Digital Shadow of the optimization approach to integrate supply chain


 planning and control,” Procedia CIRP, vol. 72, pp. [22] Y. Yusof and K. Latif, “Survey on computer-aided
520–525, 2018. process planning,” Int J Adv Manuf Technol, vol. 75,
[9] W. J. Tan, W. Cai, and A. N. Zhang, “Integration no. 1-4, pp. 77–89, 2014.
design of supply chain hybrid simulation,” in [23] M. Sadaiah, D. R. Yadav, P. V. Mohanram, and P.
Proceedings of the 2017 Winter Simulation Radhakrishnan, “A Generative Computer-Aided
Conference, W. K. Chan et al., Eds., Piscataway, NJ: Process Planning System for Prismatic Components,”
IEEE, 2017, pp. 493–504. Int J Adv Manuf Technol, vol. 20, no. 10, pp. 709–719,
[10] C. Chandra and J. Grabis, Supply Chain Configuration: 2002.
Concepts, Solutions, and Applications, 2nd ed. New [24] H. C. Lee, W. C. Jhee, and H.-S. Park, “Generative
York: Springer, 2016. CAPP through projective feature recognition,”
[11] J. Bützler, S. Kuz, H. Petruck, M. Faber, and C. M. Computers & Industrial Engineering, vol. 53, no. 2, pp.
Schlick, “Function Allocation Between Humans and 241–246, 2007.
Systems in Self-optimizing Production Networks,” [25] Y. Nonaka et al., “Generating alternative process plans
Procedia Manufacturing, vol. 3, pp. 371–378, 2015. for complex parts,” CIRP Annals, vol. 62, no. 1, pp.
[12] T. Benkedjouh, K. Medjaher, N. Zerhouni, and S. 453–458, 2013.
Rechak, “Health assessment and life prediction of [26] J. Joo, S. Park, and H. Cho, “Adaptive and dynamic
cutting tools based on support vector regression,” J process planning using neural networks,” International
Intell Manuf, vol. 26, no. 2, pp. 213–223, 2015. Journal of Production Research, vol. 39, no. 13, pp.
[13] E. Elghoneimy and W. A. Gruver, “Agent-Based 2923–2946, 2001.
Decision Support and Simulation for Wood Products [27] J. Schmidt, “Integrierte Arbeitsplanung und
Manufacturing,” IEEE Trans. Syst., Man, Cybern. C, Fertigungssteuerung auf Basis von
vol. 42, no. 6, pp. 1656–1668, 2012. Zustandsinformationen,” Dissertation, Leibniz
[14] V. A. Prabhu, N. Muhandri, B. Song, and A. Tiwari, Universität Hannover; TEWISS - Technik und Wissen.
“Decision support system enabled by depth imaging [28] G. Schuh, J.-P. Prote, and P. Hünnekes, “Process
sensor data for intelligent automation of moving Planning Assessment Framework,” in Advances in
assemblies,” Proceedings of the Institution of Production Management Systems, I. Moon, G. M. Lee,
Mechanical Engineers, Part B: Journal of Engineering J. Park, D. Kiritsis, and G. von Cieminski, Eds., Cham:
Manufacture, vol. 232, no. 1, pp. 51–66, 2018. Springer International Publishing, 2018, pp. 473–481.
[15] D. Mourtzis et al., “Cloud-Based Adaptive Shop-Floor [29] A. Azab and H. ElMaraghy, “Sequential process
Scheduling Considering Machine Tool Availability,” in planning: A hybrid optimal macro-level approach,”
Proceedings of the ASME 2015, November 13-19, Journal of Manufacturing Systems, vol. 26, no. 3-4, pp.
2015, Houston, Texas, USA, Houston, Texas, USA, 147–160, 2007.
2016, V015T19A017. [30] Supply Chain Council, Supply Chain Operations
[16] S. Plitsos, P. P. Repoussis, I. Mourtos, and C. D. Reference Model: Revision 11.0. United States of
Tarantilis, “Energy-aware decision support for America.
production scheduling,” Decision Support Systems, vol. [31] G. Schuh, J.-P. Prote, M. Luckert, and P. Hünnekes,
93, pp. 88–97, 2017. “Knowledge Discovery Approach for Automated
[17] C. Reuter, T. Nuyken, S. Schmitz, and S. Dany, Process Planning,” Procedia CIRP, vol. 63, pp. 539–
“Iterative Improvement of Process Planning Within 544, 2017.
Individual and Small Batch Production,” in Advances [32] H. Petruck and A. Mertens, “Predicting Human Cycle
in Production Management Systems, S. Umeda et al., Times in Robot Assisted Assembly,” in Proceedings of
Eds., Cham: Springer International Publishing, 2015, the AHFE 2017, July 17-21, 2017, Los Angeles,
pp. 283–290. California, USA, S. Trzcielinski, Ed., Cham, s.l.:
[18] X. Xu, L. Wang, and S. T. Newman, “Computer-aided Springer International Publishing, 2018, pp. 25–36.
process planning – A critical review of recent [33] O. Karhu, P. Kansi, and I. Kuorinka, “Correcting
developments and future trends,” International Journal working postures in industry: A practical method for
of Computer Integrated Manufacturing, vol. 24, no. 1, analysis,” Applied Ergonomics, vol. 8, no. 4, pp. 199–
pp. 1–31, 2011. 201, 1977.
[19] W. Eversheim, Organisation in der Produktionstechnik [34] P. Tucker, S. Folkard, and I. Macdonald, “Rest breaks
3. Berlin, Heidelberg: Springer Berlin Heidelberg, and accident risk,” The Lancet, vol. 361, no. 9358, p.
2002. 680, 2003.
[20] H.-P. Wiendahl, Betriebsorganisation für Ingenieure, [35] D. Norman, The design of everyday things: Revised
8th ed. München: Hanser, 2014. and expanded edition. New York, NY: Basic Books,
[21] P.-T. Chang and C.-H. Chang, “An integrated artificial 2013.
intelligent computer-aided process planning system,” [36] C. Schlick et al., “Cognition-Enhanced, Self-
International Journal of Computer Integrated optimizing Production Networks,” in Integrative
Manufacturing, vol. 13, no. 6, pp. 483–497, 2000. production technology: Theory and applications, C.
Brecher and D. Özdemir, Eds., Cham, Switzerland:
Springer, 2017, pp. 645–743.