Smart and Resilient Inventory Planning in B۲C Supply Chains with Vendor-Managed Replenishment

This study develops and evaluates a three-echelon, IoT-enabled vendor-managed inventory (VMI) model for B۲C supply chains operating under a periodic (s,S) policy. Real-time customer-level inventory visibility, provided by smart devices and sensors, is integrated into a mixed-integer linear program that coordinates replenishment decisions from a single manufacturer to multiple retailers and final customers. To hedge against forecast errors and sensor noise, a robust counterpart based on the Bertsimas-Sim budgeted uncertainty framework is proposed, which preserves tractability while limiting conservatism through an uncertainty budget T. Both deterministic and robust formulations are implemented in Python/Gurobi and tested on a structured set of small, medium, and large synthetic instances with distance-based transportation costs, range-limited retailer-customer serviceability, and capacity constraints. Computational experiments show that robust solutions consistently incur higher total costs-quantifying the price of robustness-while remaining computationally tractable (final MIP gaps ۱% across all instances). In the reported experiments, robust solutions also exhibit higher nominal backorders, a known effect of budgeted-uncertainty protection that reallocates capacity to ensure worst-case feasibility. The results highlight a clear cost-robustness trade-off and offer guidance on when robust planning is justified: robust VMI is preferable in volatile environments demanding reliability guarantees, whereas deterministic planning remains attractive in stable settings with accurate forecasts. The modeling and data pipeline support reproducibility and scaling to larger instances, and the proposed formulation provides a practical foundation for resilient, data-driven inventory planning in connected consumer supply chains.

کلیدواژه ها:

Inventory ، Supply Chain ، Vendor Managed Inventory ، Internet Of Things ، Robust Optimization