A random forest based neutrosophic fuzzy pentagonal framework for automobile spare parts inventory optimization

Automobile spare-parts inventory management is challenged by uncertain demand, fluctuating costs, and incomplete information, making conventional inventory models less effective for determining optimal replenishment decisions. This study proposes a Machine learning-based Neutrosophic–GMIR inventory optimization framework to address uncertainty in spare-parts planning. Pentagonal fuzzy numbers are employed to represent uncertain inventory parameters, followed by GMIR defuzzification and neutrosophic transformation to incorporate truth, indeterminacy, and falsity into inventory parameter estimation. A nonlinear inventory cost model is solved using the Newton–Raphson method to determine the optimal replenishment quantity, while a random forest regressor predicts the optimal order quantity based on the neutrosophic inventory parameters. Numerical results demonstrate that the proposed model reduces the optimal order quantity from 278 to 245 units (11.9%) and decreases the total inventory cost by 12.8% compared with the GMIR-based model, indicating improved inventory efficiency under uncertainty. By reducing excess inventory, improving resource utilization, and supporting data-driven replenishment decisions, the proposed framework contributes to sustainable development goal (SDG) 9: Industry, Innovation and Infrastructure and SDG 12: Responsible Consumption and Production. The proposed approach provides an effective decision-support framework for sustainable automobile spare-parts inventory management under uncertain operating conditions.