42
On the combination of Sensor Data in Supply Chain Automation Challenge and Research Agenda SENSOR DATA IN SUPPLY CHAIN AUTOMATION 1 Presented by Akkaranan Pongsathornwiwat Assistant Researcher LogEn i4.0, SIIT, Thammasat University

On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

On the combination of Sensor Data in Supply Chain Automation

Challenge and Research Agenda

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 1

Presented byAkkaranan PongsathornwiwatAssistant ResearcherLogEn i4.0, SIIT, Thammasat University

Page 2: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Contents ❑ Automation: From manufacturing To supply chain

❑ Sensor in supply chain automation

❑Managing sensor data in digital supply chain: single or multiple data is useful for productivity and operational efficiency?

❑ On the combination of sensor data: trends and research agenda

❑What’s next in sensor and AI technology

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 2

Page 3: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Automation: From manufacturing To supply chain❑Why is industrial automation important?

❑ Increase operational efficiency

❑ Increase productivity

❑ Increase resource utilizations

❑ Increase customer satisfaction

❑ Increase profitability

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 3

https://images.app.goo.gl/DY8cr1PmYwfbBeEN7

Page 4: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Automation: From manufacturing To supply chain❑ There are THREE important industrial solutions for helping better operationalexcellence.

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 4

https://images.app.goo.gl/7ChNs3yGB63LjPzL7

Machining Solutions Measuring Solutions Factory automation & Robots Solutions

https://images.app.goo.gl/i543YiXcnndWtnYp9https://images.app.goo.gl/beAJ9fhH9MquVj2w8

Page 5: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Automation: From manufacturing To supply chain

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 5

❑ The heart of three industrial solutions is the process of monitoring and control engineering.

❑ Sensor plays important role in detecting and collecting data for analysis in order to providewhat-if scenarios in targetrecognition.

https://images.app.goo.gl/6Bk2ZHtArhoYjrW88

Page 6: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Automation: From manufacturing To supply chain

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 6

❑ Taking advantages from automation in manufacturing to make a supply chain smarter. ❑ Increase visibility

❑ Increase transparency

❑ Increase predictive capability

❑ Increase adaptability

https://images.app.goo.gl/6Bk2ZHtArhoYjrW88

Page 7: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Sensor in supply chain automation❑ The need of supply chain automation and tracking system.

❑ Case study: A big car assembly manufacturer in Thailand

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 7

Page 8: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Sensor in supply chain automation

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 8

https://images.app.goo.gl/NVCq3XuNc4WGV2ZV8

Increase operationalefficiency throughautomation

Reduce repair costs andmaintenance downtimethrough better monitoring

Perform real-time inventory tracking with improved demand planningEnhance customer service

by connecting moreclosely to the customer

Page 9: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

The smart sensor ecosystem

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 9

Source: http://engineering.nyu.edu/gk12/amps-cbri/pdf/Intro%20to%20Sensors.pdf

Page 10: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Analyzing sensor data in supply chain automation❑ IBM’s supply chain teams proposed four stages of supply chain analytics

1) Descriptive analytics => Visualization

2) Predictive analytics => Predict events or future outcomes

3) Prescriptive analytics => what should we do

4) Cognitive analytics => deal with a human nature problem responding to the case that never happens in the past; for example, an organization answer complex questions in natural language — in the way a person or team of people might respond to a question.

Source: https://www.ibm.com/supply-chain/supply-chain-analytics

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 10

Page 11: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

11

Decision and Business IntelligentD

ecis

ion

In

tell

igen

ce

Business Intelligence

Optimization modeling: Planning

Visualization – Is the best happening?

Simulation modelling

Forecasting modeling

Statistical Analysis

Query/Drill down – Where exactly is the problem?

Standard report – What happened?

ERP – Record Transactional Data

Page 12: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

12

Industry 4.0 Maturity Levels

acatech Industrie 4.0 Maturity Index - A Multidimensional Maturity Model

ERP Optimization

Page 13: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

How to combine sensor data for useful cases?❑ A single sensor may not be enough to derive a desired level of target estimation or hypothesis identification in supply chain applications.

❑ Therefore, multiple sensors are required to achieve a complete and accurate description of an environment or process of interest.

❑ The simple question is how to combine sensor data gathered for useful cases.

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 13

Page 14: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

How to combine sensor data for useful cases?❑ A generic framework for multiple data combination

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 14

Sensor Data Unification and Assignment methods

Weight and reliability determination

Rule of Combination techniques

Frikha, A., & Moalla, H. (2015). Analytic hierarchy process for multi-sensor data fusion based on belief function theory. European Journal of Operational Research, 241(1), 133-147.

Page 15: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

How to combine sensor data for useful cases?❑ A generic framework for multiple data combination

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 15

Sensor Data Unification and Assignment methods

Weight and reliability determination

Rule of Combination techniques

Frikha, A., & Moalla, H. (2015). Analytic hierarchy process for multi-sensor data fusion based on belief function theory. European Journal of Operational Research, 241(1), 133-147.

Page 16: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Sensor Data-related Fusion AspectFig. Taxonomy of data fusion methodologies: different data fusion algorithms can be roughly categorized based on one of the four challenging problems of input data that are mainly tackled

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 16

Khaleghi, B., Khamis, A., Karray, F. O., & Razavi, S. N. (2013). Multisensor data fusion: A review of the state-of-the-art. Information fusion, 14(1), 28-44.

Page 17: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Theoretical Foundation❑ Fusion of imperfect data

❑ Probabilistic fusion

❑ Evidential belief reasoning (DSET)

❑ Fusion and fuzzy reasoning

❑ Hybrid fusion approaches

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 17

Khaleghi, B., Khamis, A., Karray, F. O., & Razavi, S. N. (2013). Multisensor data fusion: A review of the state-of-the-art. Information fusion, 14(1), 28-44.

Page 18: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

How to combine sensor data for useful cases?❑ A generic framework for multiple data combination

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 18

Sensor Data Unification and Assignment methods

Weight and reliability determination

Rule of Combination techniques

Frikha, A., & Moalla, H. (2015). Analytic hierarchy process for multi-sensor data fusion based on belief function theory. European Journal of Operational Research, 241(1), 133-147.

Page 19: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Some promising research issues❑Most fusion systems are optimistic in that they assume that all sensors are reliableand pay more attention to uncertainty modeling and fusion methods.

❑ The performance of the fusion system is, however, highly dependent on sensor performance and adaptability to the operating environment as well as ability to estimate the reliability of each sensor readings (pieces of evidence).

❑ Two issues that need to be considered❑ The data derived from multiple sources (signals or humans) is usually imperfect (imprecise,

uncertain, and even conflicting). ❑ The imperfection and unreliability of sensor data are often attributed to technical and noise

(environmental noise, presence of unknown targets, meteorological conditions, etc.) factors.

❑ It needs an index for quantifying sensor performance and weighing readings.

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 19

Page 20: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Some promising research issues❑We can then treat such issues in Multiple Criteria Decision-Making Problem (MCDM).

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 20

A Hierarchical structure for the weight evaluation of sensor BOEs, proposed by Frikha & Moalla (2015).

multiple conflicting decision factors

Page 21: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Ongoing data fusion research ❑ Automated fusion

❑ Belief reliability

❑ Fusion evaluation❑ Evaluating the quality of input data to the fusion system

❑ Assessing the performance of the fusion system

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 21

Page 22: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Analyzing sensor data in supply chain automation❑ IBM’s supply chain teams proposed four stages of supply chain analytics

1) Descriptive analytics => Visualization

2) Predictive analytics => Predict events or future outcomes

3) Prescriptive analytics => what should we do

4) Cognitive analytics => deal with a human nature problem responding to the case that never happens in the past; for example, an organization answer complex questions in natural language — in the way a person or team of people might respond to a question.

Source: https://www.ibm.com/supply-chain/supply-chain-analytics

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 22

Page 23: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Our research interests❑ Cognitive analytics => deal with a human nature problem responding to the case that never happens in the past; for example, an organization answer complex questions in natural language — in the way a person or team of people might respond to a question.

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 23

Page 24: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Problem formulation

❑ Linguistic assessments represent uncertain and vagueness of information due to human’s ability.

❑ Linguistic values are totally different from numeric ones.

❑ A computing with word (CW) methodology is necessary!

24

Page 25: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Data set

25

Page 26: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Data set

26

Page 27: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Data modelling ❑ How to model the uncertain linguistic assessment?

❑We reformulate the Dempster-Shafter theory of evidence for modeling the data.

27

Page 28: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Data representation

28

Page 29: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Data representation

29

Page 30: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Future Directions ❑ It is not effective to define a unique linguistic term set to be used by alldecision-makers (Herrera et al, 2016; Jiang et al., 2017).

30

Page 31: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Future Directions ❑ To extend a further study to application feedback stage in order for continuous improvement (Wu et al., 2012).

❑ So we aim to gather more experts’ opinions in order to build a software-based decision support system (DSS)

31

Page 32: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Further readings 1) Zhang, Y., Liu, Y., Zhang, Z., Chao, H. C., Zhang, J., & Liu, Q. (2017). A weighted

evidence combination approach for target identification in wireless sensor networks. IEEE Access, 5, 21585-21596.

2) Frikha, A., & Moalla, H. (2015). Analytic hierarchy process for multi-sensor data fusion based on belief function theory. European Journal of Operational Research, 241(1), 133-147.

3) Khaleghi, B., Khamis, A., Karray, F. O., & Razavi, S. N. (2013). Multi-sensor data fusion: A review of the state-of-the-art. Information fusion, 14(1), 28-44.

4) Sentz, K., & Ferson, S. (2002). Combination of evidence in Dempster-Shafer theory (Vol. 4015). Albuquerque: Sandia National Laboratories.

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 32

Page 33: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

What’s next in sensor and AI technology

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 33

Page 34: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Cyber space

Physical space

IoT

Smart Machine

Smart Machine

Total Synchronize

34

AI Cloud

Page 35: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

AI Planning and AI Scheduling with Production Simulation

AI Planning

AI Scheduling

ProductionPlanning/Scheduling

GD.findi Production Simulation

35

Page 36: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Apply this course:

Page 37: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification
Page 38: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Handout MaterialsGD.findi:

shorturl.at/jzCH2

Page 39: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Purpose of UAP

• Share the concept of Deep Thinking world-wide.

• Collaborate each other to establish new methodology.

• Share computing resources each other through worldwide Deep Thinking computing network.

Application form:

University Alliance Program

Please E-mail [email protected] to join us!

Page 40: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

LogEn i4.0 CONTACTE-mail: [email protected]

Page 41: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Final remarks!

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 41

Page 42: On the combination of Sensor Data in Supply Chain Automation …logistics.nida.ac.th/wp-content/uploads/2019/11/3Senser... · 2019. 11. 1. · estimation or hypothesis identification

Any questions or suggestions is welcome!

SENSOR DATA IN SUPPLY CHAIN AUTOMATION 42