Blockchain, Recommender System, and Artificial Intelligence: Leveraging advanced digital technology for disruption risk mitigation in the medical device closed-loop supply chain

Motivated by the global medical device supply chain (SC) disruption (shortage) chal-lenges triggered by the COVID-19 pandemic, in light of revolutionary trends such as the technological revolution (i.e., Industry 4.0 and 5.0), especially the rapid develop-ment of advanced digital technologies such as artificial intelligence (AI), Blockchain (BLC), Digital Twin (DT)/Simulation, and Big Data Analysis (BDA), simultaneously considering the requirements of sustainability, this thesis aims to build a resilient and sustainable medical device CLSC by leveraging advanced digital tools. This goal is achieved by (1) developing a novel BLC-based intelligent recom-mender system for SC disruption on the forward side of the medical device CLSC and (2) developing an ML-based medical device recall initiator prediction system, on the reverse side. To mitigate the disruption risk on the forward side of SC and ensure the overall CLSC operates effectively, this research first proposes a data-driven supply chain dis-ruption response framework baseline based on intelligent recommender system tech-niques. The blockchain (BLC) technology is integrated into the baseline architecture of the IRS to promise the reliability and stability of the proposed IRS as a resilient de-cision support system. The novel BLC-IRS framework can generate high-quality rec-ommendation results and identify internal and external supplementary resource infor-mation in an agile, safe, and real-time manner, thereby facilitating a higher resilience level in the physical supply chain. The BLC-IRS was implemented on a practical use case, and the information exchange mechanism in the BLC network through a smart contract was prototyped. The BLC-IRS framework was validated as an effective digi-tal SCRes measure by a System Dynamics (SD) simulation model. The simulation results demonstrate that the initial proposed BLC-IRS framework can be effectively implemented as a SC disruption mitigation measure in the SCRes response phase, enabling SC participants to better react to SC disruptions at the physical level. On the reverse side, to reduce the unexpected reverse logistics activities in the medical device sector, this research developed a machine learning-based medical de-vice recall initiator prediction framework with five algorithms to conduct proactive fail-ure detection through a practical industrial case study. In this case study, an accuracy rate of 88.18% is achieved, indicating the potential of the proposed framework in as-sisting manufacturers with asset predictive failure detection, thereby reducing recalls. A comparative analysis of prediction performance between the proposed framework and the most similar research was presented. The comparison results showed that the dataset, clustering method, and key input features chosen by this thesis are valid and efficient. The proposed predictive framework obtains higher accuracy, scalability, and practical values. The primary objective of the proposed predictive framework is to establish a sustainable and resilient supply chain utilising an intelligent decision-making support tool. This research contributes two executable digital SCRes solutions for academia and industry, as most of the previous discussions on utilizing advanced digital tech-nologies for developing SCRes are limited to the conceptual level. The deliveries of the whole research are significant from a comprehensive perspective, as this re-search is a piece of work that: (1) Considered both the forward side and reverse side of the medical device CLSC. (2) Developed SCRes with both proactive and reactive measures. (3) Combined the requirements of resiliency and sustainability.