Modeling and Managing Product Unavailability Risk in Inventory Through a Fuzzy Bayesian Network

In the field of inventory management, where important decisions must be made to maintain a balance between product availability and storage costs, adopting innovative approaches for assessing and managing potential risks is essential. This article focuses on analyzing the risk associated with the unavailability of a specific product in stock. It presents a hybrid methodology that combines Bayesian networks for modeling causal relationships between factors influencing stock levels, and fuzzy logic to calculate conditional probabilities of intermediate network nodes, accounting for uncertainty in decision-making in this domain. The developed model’s validation is confirmed through the verification of three axioms to ensure its reliability and accuracy. After validation, eight different scenarios were anticipated to assess the factors influencing the reduction of product quantity in stock using a sensitivity analysis. The result of this model was further applied in an event tree analysis to explore the impact of various management strategies on revenue loss.