A Supervised Machine Learning Approach to Data-driven Simulation of Resilient Supplier Selection in Digital Manufacturing

This article focuses on an entirely different approach to analyzing the risk profiles of supplier performance under uncertainty by utilizing the data analytics capabilities in digital manufacturing. Digital manufacturing peculiarly challenges supplier selection with dynamic order allocations, and opens new opportunities to exploit digital data to improve sourcing decisions. The authors develop a hybrid technique, combining simulation and machine learning and examine its applications for data-driven decision-making support in resilient supplier selection.

The increase in data availability and the emergence of new digital technologies, such as machine learning, cloud computing, the internet of things (IoT), and blockchain, have enabled managers and governments to use intelligent decision-making principles when manage uncertainties.

The Big Data phenomenon forced the development of new techniques in fast analytics and data science as part of business intelligence. Al‐tay et al. point out supply chain agility and supply chain resilience are dynamic capabilities that have significant effects on supply chain performance.

Digital technologies do not only enable data-driven decision support tools, but also stimulate the development of new production forms, such as smart manufacturing and Industry 4.0. These new forms of digital manufacturing are characterized by higher flexibility, make-to-order environments, and customer-driven SC dynamic structuring, all of which require dynamic supplier selection analysis.

At the same time, digital manufacturing is expected to face increased risk of disruption due to increasing complexity and globalization. As such, there is also a need for new modeling approaches with which to analyze resilient supplier selection in novel organizational networks. Furthermore, Big Data can be essential in supplier risk management as it can enable detailed understanding of supplier performance in regards to the identification of opportunities for better sourcing.

Digital manufacturing peculiarly challenges supplier selection with dynamic order allocations and provides new opportunities to exploit digital data to improve sourcing decisions. In this study, supervised machine learning (SML) is further investigated with regards to its application for supplier selection in digital manufacturing in relation to resilience.

The simulation model is performed with AnyLogic software and represents a make-to-order manufacturing system which has up to four raw material suppliers.

Source: International Journal of Information Management, vol. 49, pp. 86-97. PDF link below image.


Simulation and Machine learning.jpg

Integration between simulation and machine learning models.

Postes connexes