Abstract
Stockouts in hub-and-spoke distribution networks can disrupt service levels and drive-up logistics costs. This project develops a Mixed-Integer Linear Programming (MILP) model to support inventory recovery decisions by integrating forecasted stockout risk, inventory positions, demand, and operational constraints. The model evaluates hub-to-DC and lateral DC-to-DC transfers, balancing trade-offs between transportation cost, underutilized trucks, early transfer penalties, and service failure costs. A key feature is its configurability at the SKU and customer level, allowing planners to simulate tailored recovery actions. The model consumes externally provided stockout probability forecasts at the distribution center (DC) level to inform recovery prioritization. Results show the model can reduce stockouts to zero under feasible conditions while highlighting the cost trade-offs of recovery options. The model is deployed through an interactive dashboard to support scenario analysis and rapid planning. This framework provides supply chain planners with a data-driven approach to proactively mitigate stockouts and improve network resilience.
