Optimization of Pull-Based Replenishment Strategy for Buffer in Tobacco Warehouse Logistics System

To optimize the reliability of the tobacco warehouse logistics system, a pull-based replenishment strategy is designed for the buffer at the packing station. The logistics system is characterized by a large scale, coupled logistics processes, and uncertain material transportation times, which make the identification of optimal strategy parameters challenging. To address this, a framework for optimizing the parameters is proposed. Initially, a simulation model of the tobacco warehouse is developed. Regression models linking the strategy parameters to system performance indicators are constructed through simulation experiments and stacking ensemble learning, forming a surrogate mathematical model. Subsequently, the gray wolf optimizer (GWO) is employed to identify the optimal strategy parameters with the objective of minimizing the load on the rail-guided vehicle (RGV) system. The final results demonstrate that the optimal replenishment strategy reduced the load on the RGV system by 3.68%. The optimized system is better at handling unexpected RGV failures.