Design, Formal Analysis, and Historical-Sales-Based Evaluation of Combined Hybrid Inventory Control Under Uncertainty

Background: Inventory controllers combine triggering mechanisms, ordering rules and forecasting models, but integration does not guarantee economic improvement. This study formalizes a mechanism–policy–model architecture for coordinated two-source replenishment. Methods: Formal analysis establishes a bounded non-overordering property and restricted optimization results. An out-of-sample simulation uses 52 weeks of observed sales for 811 products, split equally between calibration and evaluation; supplier prices, capacities and lead times are scenario assumptions. A separate Monte Carlo study examines independent, intermittent and correlated demand with fixed or random supplier lead times. Results: Relative to a tuned periodic base-stock policy, the full hybrid costs 16.28% more under the historical-sales scenario assumptions and achieves a lower immediate fill rate. In the Monte Carlo study, performance rankings vary by scenario: the full hybrid lowers cost under intermittent demand but is more expensive under normally distributed demand. Conclusions: The architecture provides an auditable design and evaluation framework rather than a universally superior inventory policy. The formal results do not establish global stochastic optimality. Complete industrial validation requires observed procurement and inventory records.