Proactive computing in industrial maintenance decision making

http://imu.ntua.gr
Proactive Computing in Industrial
Maintenance Decision Making
Alexandros Bousdekis
Supervisor: Professor Gregoris Mentzas

http://imu.ntua.gr
Outline
• Introduction
• The Thesis
• Information System
• Evaluation in Industrial Environments
• Conclusions and Future Work
7/5/2018
2

http://imu.ntua.gr
Outline
• Introduction
• The Thesis
7/5/2018
3

http://imu.ntua.gr
The thesis research area
7/5/2018
4

http://imu.ntua.gr
Industry 4.0
7/5/2018
5

http://imu.ntua.gr
Industry 4.0
7/5/2018
6
Industry 4.0 indicates the flexibility that exists in value-
creating networks which enables machines and plants to
adapt their behaviour to changing orders and operating
conditions through self-optimization and
reconfiguration.
Platform Industrie 4.0. (2015). Platform Industrie 4.0: Umsetzunsstrategie Industrie 4.0. Berlin.

http://imu.ntua.gr
Industrial Maintenance
7/5/2018
7

http://imu.ntua.gr
7/5/2018
ICCS – Alexandros Bousdekis 8
• Maintenance is a key operation function
within manufacturing enterprises.
• New practices put failure prediction at the
backbone of maintenance decision making.
Guillén, A. J. et al (2016): A framework for effective management of condition based maintenance programs in the context of
industrial development of E-Maintenance strategies. Computers in Industry, 82, 170-185.
• However, the stochastic degradation process
leads to high uncertainty in the decision
making process.
• E-maintenance refers to the convergence of
emerging information and communication
technologies with information systems to enable
decision making in a proactive way
– Technology
– Models, methods, algorithms
Muller, A., Marquez, A. C., & Iung, B. (2008). On the concept of e-maintenance: Review and current research. Reliability Engineering
& System Safety, 93(8), 1165-1187.

http://imu.ntua.gr
Proactive Computing
7/5/2018
9

http://imu.ntua.gr
Proactive Computing (1/2)
7/5/2018
10
• Getting physical
– Proactive systems will be intimately
connected to the world around them
• using sensors and actuators to both
monitor and shape their physical
surroundings
• Getting real
– Proactive computers will routinely
respond to external stimuli
• at faster-than-human speeds.
• Getting out
– Interactive computing deliberately
places human beings in the loop
• Shrinking time constants and sheer
numbers demand research into proactive
modes of operation in which humans are
above the loop.
Tennenhouse, D. (2000). Proactive computing. Communications of the ACM, 43(5), 43-50.

http://imu.ntua.gr
Proactive Computing (2/2)
• Event-Driven Architecture (EDA) provides
an architectural computing paradigm that has
the ability to react to changes by processing
events.
7/5/2018
11
Source: Etzion O. (2016). Proactive Computing:
Changing the Future. RTInsights.
• Proactivity is referred to the ability to avoid or
eliminate the impact of undesired future events, or to
exploit future opportunities on the basis of real-time
predictions about a future event.
– “Detect-Predict-Decide-Act”
Engel, Y. & Etzion, O. (2011): Towards proactive event-driven computing. In Proceedings of the 5th ACM international conference on Distributed event-based system
(pp. 125-136). ACM.

http://imu.ntua.gr
Literature Review Outcomes
7/5/2018
12
Industry 4.0
• Documentation of “Platform Industrie 4.0”
• 19 scientific publications
• 7 consulting reports
• Standards
• 13 consulting and software reports
Proactive Computing
• 2 consulting and software reports
Manufacturing companies are slowly adopting
novel technologies.
Proactive computing needs successful applications in
order to further prove its effectiveness.
Maintenance and logistics
management are strongly
interconnected.
Industry 4.0 is at the very
first steps of its evolution.
The dynamic manufacturing environment poses
challenges to decision making.
Most applications currently supported by event
processing platforms are reactive by nature.

http://imu.iccs.gr
Research Questions
13
7/5/2018
RQ4
• How to incorporate context-awareness in
proactive decision making?

http://imu.ntua.gr
Outline
• Introduction
• The Thesis
7/5/2018
14

http://imu.iccs.gr
Next generation of maintenance in Industry 4.0
15
7/5/2018
RQ4
RQ1
• What is the next generation of industrial
maintenance in Industry 4.0?

http://imu.ntua.gr
Limitations of existing maintenance solutionsExistingLimitations
Physical models of equipment not easily extensible to other equipment.
Rarely exploit sensor-generated big data processing infrastructures.
Not incorporate data-driven decision making capabilities.
Focus on a specific aspect of maintenance
(e.g. condition monitoring, diagnostics, etc.).
At a conceptual level with limited practical applications in real industrial
environments.
7/5/2018
16
[J2] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). Review, analysis and synthesis of
prognostic-based decision support methods for condition based maintenance. Journal of Intelligent
Manufacturing, 1-14.
[C2] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). Supporting the Selection of
Prognostic-based Decision Support Methods in Manufacturing. In ICEIS (pp. 487-494).

http://imu.ntua.gr
Towards Proactive Maintenance
7/5/2018
17
Proactive Maintenance is a new
maintenance strategy that is based upon
4 technological pillars: Industry 4.0, IoT,
Big Data and Proactive Computing.
• It provides real-time monitoring,
detections, predictions and proactive
recommendations about maintenance
actions in a data-driven way.
• The aim is to support decision making
in order to eliminate or mitigate the
impact of future failures with the aim
to maximize reliability of operations
and improve business performance.

http://imu.ntua.gr
The Conceptual Architecture for Proactive Maintenance
7/5/2018
18
[C9] Bousdekis, A., & Mentzas, G. (2017). Condition-Based Predictive Maintenance in the Frame of
Industry 4.0. In IFIP International Conference on Advances in Production Management Systems (pp. 399-
406). Springer, Cham.
[C13] Bousdekis, A., Mentzas, G., Hribernik, K., Lewandowski, M., von Stietencron, M., & Thoben, K. D.
(2018). A Unified Architecture for Proactive Maintenance in Manufacturing Enterprises. In Enterprise
Interoperability: I-ESA ’18 Proceedings. Springer International Publishing. (In Press)

http://imu.iccs.gr
Proactive Decision Making in Maintenance Operations
19
7/5/2018
RQ4
RQ2
• How to support proactive decision making
in maintenance operations?

http://imu.ntua.gr
Limitations of existing maintenance decision making algorithmsExistingLimitations
They assume perfect maintenance. Imperfect maintenance actions with various
degrees is not usually considered.
Recommendations for immediate implementation of certain actions.
Not triggered by real-time predictions about future failures.
7/5/2018
20
Not embedded in an event-driven computational environment.
[J1] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). A proactive decision making
framework for condition-based maintenance. Industrial Management & Data Systems, 115(7), 1225-
1250.

http://imu.ntua.gr
Decision Making in the context of Proactivity
7/5/2018
21
 Proactive decision making aims to
enable business analysts to create and
configure decision method instances
for mitigating a future undesired event,
which lays outside the desired states
space.
 Based on the predictions for
undesirable situations derived on the
basis of streaming data, decision
methods instances are enacted online
to generate mitigating action
recommendations and optimal time of
action implementation.

http://imu.ntua.gr
Proactive Decision Making for Maintenance Actions
Proactive Expected Loss Rate Optimization
•Recommendation of the optimal time for a
pre-defined action
•PDF of the event occurrence can be of
arbitrary type.
Proactive Markov Decision Process
•Recommendation of the optimal (perfect or
imperfect) maintenance action and the
optimal time of applying it.
•PDF of the event occurrence belongs to
exponential family.
7/5/2018
22
Proactive event-driven decision methods for maintenance actionsProactive Markov Decision Process
[J4] Bousdekis, A., Papageorgiou N., Magoutas, B., Apostolou, D., & Mentzas, G. (2018). Enabling
Condition-Based Maintenance Decisions with Proactive Event-driven Computing. Computers in Industry,
100, 173-183.
[C4] Bousdekis, A., & Mentzas, G. (2015). A Proactive Decision Support System for Maintenance Cost
Minimisation in Manufacturing Enterprises. In 4th Student Conference of Hellenic Operational Research
Society (HELORS) (pp. 61-66).

http://imu.ntua.gr
Proactive Decision Making for Logistics Actions
Proactive joint replacement and spare parts
inventory decision model
Proactive joint maintenance and spare parts
inventory decision model
7/5/2018
23
Proactive event-driven decision methods for joint maintenance and
spare parts ordering optimization
𝐶 𝑚 𝑡 = 𝑐𝑓 𝑡 ∗ 𝑃 𝜀
0, 𝑡 + 𝑐𝑓 𝑡 + 𝑐𝑝 𝑡 ∗ 𝑃𝑎
𝜀
𝑡, 𝑇 + 𝑐𝑝 𝑡 ∗ 𝑃 𝜀
0, 𝑇
𝐶𝑜 𝑡 = 𝑐𝑠 𝑡 ∗ 𝑃 𝜀
0, 𝑡 + 𝐿 + 𝑐𝑠 𝑡 ∗ 𝑃𝑎
𝜀
𝑡 + 𝐿, 𝑇 + 𝑐ℎ 𝑡 ∗ 𝑃 𝜀
0, 𝑇
Proactive Selection of Maintenance Spare Parts’ Suppliers
[C8] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A Proactive
Event-driven Decision Model for Joint Equipment Predictive Maintenance and Spare Parts Inventory
Optimization. Procedia CIRP, 59, 184-189.
[C12] Bousdekis, A., & Mentzas, G. (2018). A proactive model for joint maintenance and logistics
optimization in the frame of Industrial Internet of Things. In Operational Research in Business and
Economics. Springer International Publishing. (In Press)
[C10] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A Framework
for Integrated Proactive Maintenance Decision Making and Supplier Selection. In IFIP International
Conference on Advances in Production Management Systems (pp. 416-424). Springer, Cham.
• Markowitz Portfolio Theory (MPT)
optimization
• It recommends the optimal portfolio of
suppliers at the recommended future
ordering time.

http://imu.iccs.gr
Continuous Improvement of Proactive Decision Making
24
7/5/2018
RQ4
RQ3
• How to conduct continuous improvement
of proactive decision making?

http://imu.ntua.gr
Limitations of existing feedback mechanismsExistingLimitations
7/5/2018
25
Expert-driven and constant domain knowledge (e.g. cost functions) for the
configuration of decision methods.
Data acquisition not considered for improving maintenance decision making.
Limited research on eliminating user input inaccuracy and sensor noise.
Limited research on addressing sensitivity to input parameters in proactive
decision making, especially to those related to costs.

http://imu.ntua.gr
7/5/2018
26
• Through Sensor-Enabled Feedback (SEF):
– the user is informed online about the estimated cost of action during its
implementation.
– the updated cost function of the specific action is used in the next recommendation in
which this action is involved.

http://imu.ntua.gr
Sensor-Enabled Feedback (SEF)
7/5/2018
27
Real-time Bayesian
changepoint detection
Kalman Filter
Curve Fitting with non-
negativity constraints
(Levenberg–Marquardt)
[C5] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2016). Continuous
Improvement of Proactive Event-driven Decision Making through Sensor-Enabled Feedback (SEF). In
ICEIS (pp. 166-173).
1
2
3
[J3] Bousdekis, A., Papageorgiou N., Magoutas, B., Apostolou, D., & Mentzas, G. (2018). Information
Processing for Generating Recommendations ahead of Time in an IoT-based Environment. International
Journal of Monitoring and Surveillance Technologies Research (IJMSTR), 5(4), 38-62.

http://imu.iccs.gr
Context-awareness in proactive decision making
28
7/5/2018
RQ1
• What is the next generation of industrial
maintenance in Industry 4.0?
RQ2
• How to support proactive decision making
in maintenance operations?
RQ3
• How to conduct continuous improvement
of proactive decision making?
RQ4
proactive decision making?RQ4

http://imu.iccs.gr
Context-awareness
29
7/5/2018
Perera, C., Zaslavsky, A., Christen, P., & Georgakopoulos, D. (2014). Context aware computing for the internet of things: A survey. IEEE communications surveys &
tutorials, 16(1), 414-454.
Schmidt, B., Galar, D., & Wang, L. (2016). Context awareness in predictive maintenance. In Current trends in reliability, availability, maintainability and safety (pp.
197-211). Springer, Cham.
Thaduri, A., Kumar, U., & Verma, A. K. (2017). Computational intelligence framework for context-aware decision making. International Journal of System
Assurance Engineering and Management, 8(4), 2146-2157.
Context: “any information that can be used to characterize the situation of an entity.
An entity is a person, place, or object that is considered relevant to the interaction
between a user and an application, including the user and applications themselves.”
Although it is a well-known concept in pervasive and mobile computing, it has just
started to emerge in industrial maintenance.
Machine Learning is considered a context modelling approach in terms of its
objectives. It is the best approach for intelligent context-aware systems since it is able
to deal with uncertainty in a future context.

http://imu.ntua.gr
Limitations of existing context-aware mechanisms
ExistingLimitations
7/5/2018
30
Limited and conceptual research works for industrial maintenance in the frame
of Industry 4.0.
Focused on reactive applications rather than proactive ones.
Considered in detection and prediction algorithms, but not in decision making
algorithms.
Not considered the prediction of future context for proactive decision making
in industrial maintenance.

http://imu.ntua.gr
The probabilistic context-aware model
• Context-aware Model Initialization
– Bayesian Network
7/5/2018
31
– Bayesian Cost Risk Functions
𝐶 𝑛 𝑡 = 𝐶 𝑛,𝑖 𝑡 ∗ 𝑃 𝐶 𝑛 𝑡 = 𝐶 𝑛,𝑖 𝑡 |𝐶𝐸1 ∩ … ∩ 𝐶𝐸 𝑚
𝑖=𝑘
𝑖=1
• Context-aware Model Reasoning
– X-means clustering in BN effect nodes
𝐸𝐿𝑅 =
𝐶𝑢𝑒 𝑡 ∗ 𝐺 𝜀
𝑡
𝑡 𝑢𝑒
+
𝐶𝑝𝑎 𝑡 ∗ 𝐺 𝜀
𝑡
𝑡 𝑝𝑎
Proactive event-driven decision methods
E.g.
Implementation of recommended action
Sensor-Enabled Feedback (SEF)
[C7] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2016). A probabilistic
model for context-aware proactive decision making. In Information, Intelligence, Systems & Applications
(IISA), 2016 7th International Conference on (pp. 1-6). IEEE.
[C3] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). A real-time
architecture for proactive decision making in manufacturing enterprises. In OTM Confederated
International Conferences" On the Move to Meaningful Internet Systems" (pp. 137-146). Springer, Cham.

http://imu.ntua.gr
Outline
• Introduction
• The Thesis
7/5/2018
32

http://imu.ntua.gr 7/5/2018
33
PANDDA
• PANDDA (ProActive seNsing enterprise
Decision configurator DAshboard) is a
Python web-application developed
the web2py framework.
• The presentation layer occupies the top
level of the architecture and displays
information related to services available
on the web-based PANDDA
configuration.
• The logic layer controls application
functionality by performing detailed
processing.
• The data layer houses a relational
database engine where the information
needed by the main algorithms of
PANDDA is stored and retrieved.[C1] Magoutas, B., Stojanovic, N., Bousdekis, A., Apostolou, D., Mentzas, G., & Stojanovic, L. (2014).
Anticipation-driven Architecture for Proactive Enterprise Decision Making. In CAiSE (pp. 121-128).
PANDDA in Proactive Maintenance
ProActive seNsing enterprise Decision configurator DAshboardThe PANDDA technical architecture

34
The user has created 3 instances for 3
parts of equipment. The failure prediction
triggers the part of equipment which is
referred to.
For the “Gearbox” instance, the user has
inserted 3 alternative (perfect and
imperfect) maintenance actions, their costs
and the cost of failure.
User Configuration
Proactive
Recommendation
Sensor-Enabled
Feedback
User Configuration

35
Proactive Recommendations
• PANDDA is triggered by a prediction event that there is an exponential probability
distribution for the gearbox breakdown with a time-to-failure in 9 days.
• A proactive recommendation (action-time pair) that minimizes the expected loss is
generated.
• The recommendations of each instance are stored in the system.
User Configuration
Proactive
Recommendation
Sensor-Enabled
Feedback
Proactive
Recommendation

36Alexandros Bousdekis
a. Online cost monitoring b. Cost function update
[C11] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). An Information
System for Deciding and Acting ahead of Time in Sensing Enterprises. In Information, Intelligence,
Systems & Applications (IISA), 2017 8th International Conference on (pp. 1-6). IEEE.
Proactive RecommendationsUser Configuration
Proactive
Recommendation
Sensor-Enabled
Feedback
Sensor-Enabled
Feedback

http://imu.ntua.gr
Outline
• Introduction
• The Thesis
7/5/2018
37

http://imu.ntua.gr
The Oil Drilling case (1/3)
• MHWirth is a leading global provider of first-
class drilling solutions and services.
• Global span covering 5 continents with offices
in more than 20 countries and employs 4,300
professionals.
• Its revenue is approx. 1 billion dollars.
7/5/2018
38
• Oil and gas projects are capital-intensive
investments, with severe consequences in financial
and environmental terms in case of breakdown.
• Since a typical production rate for an oil and gas
corresponds to USD 500,000, the reduction of
downtime is of great significance in the oil and gas.

http://imu.ntua.gr
7/5/2018
39
 Sensors measure parameters that are known to affect the oil rig’s gearbox:
 temperature
 vibration
 friction losses
 environmental conditions
 Detect: a real-time detection service detects a
complex pattern that indicates an abnormal
behaviour of the equipment
 Predict: a real-time predictive analytics service
provides a prediction about the gearbox
breakdown.
 Decide: context-aware proactive recommendations
about joint maintenance and inventory actions.
 Act: continuous monitoring and adaptation of the
whole cycle

http://imu.ntua.gr
Maintenance Spare parts ordering
Onshore maintenance in 98.26 hours Order the DDM in 49.12 hours
7/5/2018
40
Context-aware Proactive Recommendations before SEF
Context-aware Proactive Recommendations after SEF
Offshore maintenance in 85.47 hours Order the gearbox in 42.36 hours

http://imu.ntua.gr
The Automotive Lighting Equipment case (1/3)
• A reduction of scrap rate in the automotive
lighting industry by just 1%, results in savings of
the order of 100,000 Euro per year.
• More than 60 different raw plastic materials may
be used for component production, each with its
own properties.
7/5/2018
41
• Hella Saturnus (Ljubljana, Slovenia) is a part
of the Hella Group.
• Core business is the production of lighting
equipment for motor vehicles.
• 95% of sales are exported worldwide.
• The last annual revenue was 257,000 EUR
• Employs approx. 2,800 people.

http://imu.iccs.gr
42
– Detect: a real-time detection service detects
a complex pattern that indicates an abnormal
behaviour of the equipment
– Predict: a real-time predictive analytics
service provides a prediction about the scrap
rate exceeding a threshold
– Decide: proactive recommendations about
joint maintenance and inventory actions.
– Act: continuous monitoring and adaptation of
the whole cycle
 Sensors measure parameters that are known to affect the moulding machine
and therefore, the scrap rate of cover lens:
 the dust levels in the shop floor
 environmental factors, i.e. temperature and humidity

http://imu.ntua.gr
7/5/2018
43
Supplier A Supplier B Supplier C Supplier D
0.14 0.38 0.26 0.22
Clean the moulding machine in 3.54 hours Order the moulds in 1.32 hours

http://imu.ntua.gr
Pilot Evaluation
7/5/2018
44
Primoz Puhar
Head of Test Engineering
Department
Tor Inge Waag
Specialist Engineer
“The system provides accurate
and reliable information, while
it improves and simplifies
decision making by drilling
operators and maintenance
planners.”
“Proactive Maintenance allows
us to mature and to avoid
defect-causing combinations in
order to move towards a ‘zero
defects’ approach.”
[C14] Bousdekis, A., Magoutas, B., Apostolou, D., Mentzas, G., & Puhar, P. (2018). The ProaSense
Platform for Predictive Maintenance in the Automotive Lighting Equipment Industry. In I-ESA ’18
Workshop Proceedings. ISTE-Wiley. (In Press)
[C6] Bousdekis, A., & Mentzas, G. (2016). A Multiple Criteria Approach Using ELECTRE for the Selection
of Maintenance Strategy in Manufacturing Companies. In Proceedings of 5th International Symposium
and 27th National Conference on Operational Research (pp. 117-121).
Questionnaire-based Evaluation Experts Evaluation

http://imu.ntua.gr
Simulation-based Evaluation
• Evaluation in real industrial environment is challenging due
to:
– The large timescales
– The operations lifecycle
– The criticality of equipment
• Therefore, a simulated computational environment was
created in order to deal with cases that did not arise in the
business cases during the evaluation period.
• Extensive simulation experiments for comparative and
sensitivity analyses show the added value of the thesis.
7/5/2018
45

http://imu.ntua.gr
Proactive Decision Making (1/2)
7/5/2018
46

http://imu.ntua.gr
Proactive Decision Making (2/2)
7/5/2018
47
• Proactive decision making leads to significantly lower losses
• by 29% to 77% with respect to reactive (breakdown maintenance) policy
• by 22% to 65% with respect to preventive (time-based maintenance)
policy
• by 7% to 61% with respect to myopic policy.
• Proactive decision making is highly sensitive with respect to its
input parameters and especially to those related to cost.
• The proactive recommendations significantly change
according to the prediction events.
• The earlier a failure is predicted
– the less the expected loss
– the decision maker has more time to be prepared
EvaluationResults

http://imu.ntua.gr
7/5/2018
48
EvaluationResults
• SEF leads to more reliable recommendations.
• The Standard Error of SEF is:
• 91% lower than the expert initial estimate
• 88% lower than processing noisy data
• The increased reliability of proactive recommendations leads to a more
accurate estimate of the maintenance expected losses by 9% to 88%.
• Noise filtering in SEF has a strong effect on cost function estimation.
• When cost function is high, uncertainty in sensor measurements leads to less
accurate results.
• Higher noise levels lead to less accurate estimations.

http://imu.ntua.gr
Context-awareness in Proactive Decision Making
7/5/2018
49
Total Expected Loss for each approach (Euro)
Scenario Reactive Myopic Proactive Context-aware proactive
1 1,491,360 ± 185,150 827,635 ± 93,234 482,355 ± 71,566 376,810 ± 53,392
2 874,362 ± 41,275 596,122 ± 46,988 333,245 ± 37,461 281,245 ± 31,711
3 122,644 ± 12,476 93,532 ± 11,855 50,769 ± 11,450 42,712 ± 8,120
4 30,550 ± 3,122 22,550 ± 3,044 12,915 ± 2,988 9,675 ± 2,336
5 446,500 ± 23,110 315,000 ± 19,750 191,235 ± 16,814 122,651 ± 15,912
Approach Maintenance Action Logistics Action Total Expected Loss
Reactive Onshore maintenance after oil
rig moving
Immediate ordering of DDM 1,492,000 Euro
Myopic Gearbox replacement when
spare part arrives
Immediate ordering of gearbox 825,000 Euro
Proactive Operate at reduced equipment
load in 95.22 hours
Ordering of swivel hook in 84.23
hours
482,355 Euro
Context-aware
proactive
Offshore maintenance in 85.47
hours
Ordering of gearbox or gears in
42.36 hours
376,850 Euro
EvaluationResults
• Context-awareness in proactive decision making contributes to
higher accuracy in proactive decision methods’ input parameters.
• It increases the sensitivity of proactive decision making.
• To this end, the SEF mechanism acquires even higher importance.
• It leads to differences in expected losses with respect to
proactive decision making without context by 12% to 37%.
• It is sensitive to the time window between the time that a
prediction is received and the time of the predicted future failure.

http://imu.ntua.gr
Outline
• Introduction and Motivation
• Towards Proactive Maintenance Management
• The PhD Approach
• Deployment in Industrial Environments
• Evaluation
7/5/2018
50

http://imu.ntua.gr
Conclusions
7/5/2018
51
RQ1
• What is the next generation of industrial maintenance in Industry 4.0?
RQ4
• Framework for Proactive Maintenance
RQ2
• How to support proactive decision making in maintenance operations?
RQ4
• Proactive decision methods for: (i) maintenance actions; (ii) logistics actions;
RQ3
• How to conduct continuous improvement of proactive decision
making?
RQ4
• Sensor-Enabled Feedback (SEF)
RQ4
• How to incorporate context-awareness in proactive decision making?
RQ4
• Probabilistic context-aware model

http://imu.ntua.gr
Future Work
• To consider interdependencies among different parts of
equipment in proactive decision making.
– This would be applicable to complex manufacturing systems and sensor
networks.
– In this way, the failure predictions will be interrelated and proactive decision
making will provide recommendations about sets of maintenance actions.
• To utilize legacy data analytics and FMECA outcomes for the
configuration and enrichment of the proactive decision models.
• To develop a generic prescriptive analytics approach for supporting
proactive decision making.
7/5/2018
52
[C15] Lepenioti, K., Bousdekis, A., Apostolou, D., Mentzas, G. (2018). Prescriptive Analytics: A Survey of
Approaches and Methods. In International Conference on Business Information Systems (BIS). Springer,
Cham. (In Press)

http://imu.ntua.gr
Journal Publications
• [J1] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). A proactive
decision making framework for condition-based maintenance. Industrial Management & Data
Systems, 115(7), 1225-1250. Impact Factor: 2.948
• [J2] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). Review, analysis
and synthesis of prognostic-based decision support methods for condition based
maintenance. Journal of Intelligent Manufacturing, 1-14. Impact Factor: 3.667
 (NTUA Thomaideio Award 2015)
• [J3] Bousdekis, A., Papageorgiou N., Magoutas, B., Apostolou, D., & Mentzas, G. (2018).
Information Processing for Generating Recommendations ahead of Time in an IoT-based
Environment. International Journal of Monitoring and Surveillance Technologies Research
(IJMSTR), 5(4), 38-62.
• [J4] Bousdekis, A., Papageorgiou N., Magoutas, B., Apostolou, D., & Mentzas, G. (2018).
Enabling Condition-Based Maintenance Decisions with Proactive Event-driven Computing.
Computers in Industry, 100, 173-183. Impact Factor: 2.850
7/5/2018
53

http://imu.ntua.gr
Conference Publications (1/2)
• [C1] Magoutas, B., Stojanovic, N., Bousdekis, A., Apostolou, D., Mentzas, G., & Stojanovic, L. (2014).
Anticipation-driven Architecture for Proactive Enterprise Decision Making. In CAiSE (pp. 121-128).
• [C2] Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). Supporting the Selection of
Prognostic-based Decision Support Methods in Manufacturing. In ICEIS (pp. 487-494).
• [C3] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). A real-time
architecture for proactive decision making in manufacturing enterprises. In OTM Confederated International
Conferences" On the Move to Meaningful Internet Systems" (pp. 137-146). Springer, Cham.
• [C4] Bousdekis, A., & Mentzas, G. (2015). A Proactive Decision Support System for Maintenance Cost
Minimisation in Manufacturing Enterprises. In 4th Student Conference of Hellenic Operational Research
Society (HELORS) (pp. 61-66).
 (Best Paper Award)
• [C5] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2016). Continuous
Improvement of Proactive Event-driven Decision Making through Sensor-Enabled Feedback (SEF). In
ICEIS (pp. 166-173).
• [C6] Bousdekis, A., & Mentzas, G. (2016). A Multiple Criteria Approach Using ELECTRE for the Selection
of Maintenance Strategy in Manufacturing Companies. In Proceedings of 5th International Symposium and
27th National Conference on Operational Research (pp. 117-121).
• [C7] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2016). A probabilistic
model for context-aware proactive decision making. In Information, Intelligence, Systems & Applications
7/5/2018
54

http://imu.ntua.gr
Conference Publications (2/2)
• [C8] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A Proactive Event-driven
Decision Model for Joint Equipment Predictive Maintenance and Spare Parts Inventory Optimization. Procedia CIRP,
59, 184-189.
 (Best Paper Award)
 (NTUA Thomaideio Award 2016)
• [C9] Bousdekis, A., & Mentzas, G. (2017). Condition-Based Predictive Maintenance in the Frame of Industry 4.0. In
IFIP International Conference on Advances in Production Management Systems (pp. 399-406). Springer, Cham.
• [C10] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A Framework for
Integrated Proactive Maintenance Decision Making and Supplier Selection. In IFIP International Conference on
Advances in Production Management Systems (pp. 416-424). Springer, Cham.
• [C11] Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). An Information System
for Deciding and Acting ahead of Time in Sensing Enterprises. In Information, Intelligence, Systems & Applications
• [C12] Bousdekis, A., & Mentzas, G. (2018). A proactive model for joint maintenance and logistics optimization in the
frame of Industrial Internet of Things. In Operational Research in Business and Economics. Springer International
Publishing. (In Press)
• [C13] Bousdekis, A., Mentzas, G., Hribernik, K., Lewandowski, M., von Stietencron, M., & Thoben, K. D. (2018). A
Unified Architecture for Proactive Maintenance in Manufacturing Enterprises. In Enterprise Interoperability: I-ESA ’18
Proceedings. Springer International Publishing. (In Press)
• [C14] Bousdekis, A., Magoutas, B., Apostolou, D., Mentzas, G., & Puhar, P. (2018). The ProaSense Platform for
Predictive Maintenance in the Automotive Lighting Equipment Industry. In I-ESA ’18 Workshop Proceedings. ISTE-
Wiley. (In Press)
• [C15] Lepenioti, K., Bousdekis, A., Apostolou, D., Mentzas, G. (2018). Prescriptive Analytics: A Survey of Approaches
and Methods. In International Conference on Business Information Systems (BIS). Springer, Cham. (In Press)
7/5/2018
55

http://imu.ntua.gr
Awards
7/5/2018
56
CIRP Best Paper Award
Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A
Proactive Event-driven Decision Model for Joint Equipment Predictive Maintenance and
Spare Parts Inventory Optimization. Procedia CIRP, 59, 184-189.
HELORS Best Paper Award
Bousdekis, A., & Mentzas, G. (2015). A Proactive Decision Support System for Maintenance
Cost Minimisation in Manufacturing Enterprises. In 4th Student Conference of Hellenic
Operational Research Society (HELORS) (pp. 61-66).
NTUA Thomaideio Award 2015
Bousdekis, A., Magoutas, B., Apostolou, D., & Mentzas, G. (2015). Review, analysis and
synthesis of prognostic-based decision support methods for condition based maintenance.
Journal of Intelligent Manufacturing, 1-14.
NTUA Thomaideio Award 2016
Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A
Proactive Event-driven Decision Model for Joint Equipment Predictive Maintenance and
Spare Parts Inventory Optimization. Procedia CIRP, 59, 184-189.

Proactive computing in industrial maintenance decision making

More Related Content

Similar to Proactive computing in industrial maintenance decision making

Recently uploaded

Proactive computing in industrial maintenance decision making