Scenario-Based Stochastic Optimization of Perishable Goods Supply Chain Networks: Considering Disruption and Utilizing Machine Learning

Perishable goods supply chains require advanced management due to demand fluctuations, time sensitivity, and disruptions. This study presents a data-driven model that optimizes the supply chain network by considering multiple disruption types and horizontal collaboration among distribution centers. Leveraging a rich dataset that includes uncertain parameters such as demand a fuzzy c-means clustering algorithm is employed to generate realistic scenarios, effectively reducing uncertainty and enhancing model accuracy. The primary innovations of this study include taking multiple disruption types and horizontal collaboration between distribution centers into account to improve supply chain resilience and utilizing two combined solution methods to increase computational efficiency. Computational results demonstrate that reliable distribution centers play the primary role in responding to baseline demand, whereas unreliable centers act as complementary centers. Furthermore, horizontal collaboration among the distribution centers reduces shortage costs and increases overall system resilience. This data-driven model can serve as an effective tool and aid supply chain management in making effective, optimal decisions when confronting disruptions and demand fluctuations by presenting an approach aimed at optimizing uncertainty management and enhancing resilience in perishable goods supply chains.