Abstract
Because supply chains are complex systems prone to uncertainty, statistical analysis is a useful tool for capturing their dynamics. Using data on acquisition history and data from case study reports, we used regression analysis to predict backorder aging using National Item Identification Numbers (NIINs) as unique identifiers. More than 56,000 NIINs were identified and used in the analysis. Bayesian analysis was then used to further investigate the NIIN component variables. The results indicated that it is statistically feasible to predict whether an individual NIIN has the propensity to become a backordered item. This paper describes the structure of a Bayesian network from a real-world supply chain data set and then determines a posterior probability distribution for backorders using a stochastic simulation based on Markov blankets. Fuzzy clustering was used to produce a funnel diagram that demonstrates that the Acquisition Advice Code, Acquisition Method Suffix Code, Acquisition Method Code, and Controlled Inventory Item Code backorder performance metric of a trigger group dimension may change dramatically with variations in administrative lead time, production lead time, unit price, quantity ordered, and stock. Triggers must be updated regularly and smoothly to keep up with the changing state of the supply chain backorder trigger clusters of market sensitiveness, collaborative process integration, information drivers, and flexibility.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have