Mitigating Disruptions in Transportation and Logistics Through Adaptive Neuro-Fuzzy Inference-Based Supply Chain Resilience

To evaluate and improve supply chain resilience across various industries, including food manufacturing, this paper investigates the application of adaptive neuro-fuzzy inference systems (ANFIS) in assessing supply chain resilience, with a focus on transportation and logistics. ANFIS models were developed and validated using two real-world industrial cases: a local plastic manufacturer and a global food manufacturer, to capture and analyze relationships within supply chains across key attributes, including security, survivability, recovery, technological capabilities, and sustainability. This research highlights the effectiveness of the fuzzy inference system (FIS), including membership function types within the ANFIS framework, in achieving computational efficiency and precision for modeling dynamic supply chain systems. The models were validated using two standard metrics: Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE), ensuring accurate predictions. Results showed that the global food company had lower average errors, indicating a more stable and reliable supply chain with a testing error of 5%. In contrast, the local plastic company exhibited higher errors in some areas, with an average error of 8%, suggesting areas for improvement. Overall, this study underscores the potential of ANFIS as a powerful tool for providing valuable insights into the strengths and weaknesses of supply chains, thereby offering a robust approach to enhancing their resilience. The findings suggest that companies can significantly benefit from adopting this neuro-fuzzy model (i.e., ANFIS), particularly when combined with continuous data refinement and targeted improvements to address specific vulnerabilities in their supply chains.