Epi-Aware Lifeline Analytics: Inventory-Optimized Blood Demand Forecasting with Integrated Epidemiological Intelligence

Maintaining optimal blood stocks is crucial to avoid both supply gaps and unnecessary waste. This paper presents Epi-Aware Lifeline Analytics, a framework that combines epidemiological intelligence with machine-learning models for dynamic demand forecasting and inventory optimization. Historical transfusion records, disease-surveillance data, and contextual factors are integrated weekly and transformed into predictive features such as lagged incidence and a composite risk score. A stacked ensemble of Random Forest, SARIMAX, and LSTM models, tuned via Bayesian optimization, generates high-accuracy forecasts that drive a Mixed-Integer Linear Programming module to minimize wastage while ensuring service levels. The system is deployed as a microservice-based dashboard with realtime heatmaps, automated alerts, and role-specific interfaces. This end-to-end architecture delivers precise forecasting and improved operational efficiency for healthcare blood-supply management.